CN117809637A - Control device and control method for intelligent desktop - Google Patents
Control device and control method for intelligent desktop Download PDFInfo
- Publication number
- CN117809637A CN117809637A CN202410121082.9A CN202410121082A CN117809637A CN 117809637 A CN117809637 A CN 117809637A CN 202410121082 A CN202410121082 A CN 202410121082A CN 117809637 A CN117809637 A CN 117809637A
- Authority
- CN
- China
- Prior art keywords
- user
- voice
- signal
- sitting posture
- interrupt
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 62
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 64
- 230000005236 sound signal Effects 0.000 claims abstract description 53
- 230000001960 triggered effect Effects 0.000 claims abstract description 37
- 238000001914 filtration Methods 0.000 claims abstract 2
- 230000004044 response Effects 0.000 claims description 76
- 230000008859 change Effects 0.000 claims description 57
- 238000012937 correction Methods 0.000 claims description 26
- 230000005484 gravity Effects 0.000 claims description 19
- 238000013519 translation Methods 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 6
- 230000007613 environmental effect Effects 0.000 abstract description 7
- 238000010586 diagram Methods 0.000 description 22
- 230000006870 function Effects 0.000 description 12
- 210000003128 head Anatomy 0.000 description 11
- 230000006872 improvement Effects 0.000 description 10
- 238000005516 engineering process Methods 0.000 description 9
- 238000004364 calculation method Methods 0.000 description 8
- 210000001217 buttock Anatomy 0.000 description 7
- 238000012545 processing Methods 0.000 description 7
- 230000001755 vocal effect Effects 0.000 description 7
- 239000011159 matrix material Substances 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 4
- 238000004590 computer program Methods 0.000 description 3
- 238000010438 heat treatment Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000003491 array Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 230000004807 localization Effects 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 241000238558 Eucarida Species 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000004886 head movement Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/20—Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
- G10L2015/223—Execution procedure of a spoken command
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Circuit For Audible Band Transducer (AREA)
Abstract
Description
技术领域Technical field
本发明涉及家具控制器,特别涉及一种智能桌面的控制装置及控制方法。The present invention relates to furniture controllers, and in particular to a control device and a control method for an intelligent desktop.
背景技术Background technique
智能桌面是集成了多种功能的智能家具,例如可以在智能桌面上集成无线充电、氛围灯等功能,由于功能增加,需要提供更加丰富的控制方式,如申请人在先申请的专利(公开号为CN209450059U)提供了通过触摸按键对智能桌面功能进行控制的方案。智能桌面可以作为单独家具出现,例如茶几等,也可以是作为其他智能家具的配件,例如在智能沙发椅的扶手上部署智能桌面,也可以在化妆柜上部署智能桌面等,可见后续智能桌面可以作为配件被添加到更多的智能家具中,以提升智能家具的集成度。然而,通过按键方式进行智能桌面的控制仍然不够智能,语音控制是本领域的技术改进方向。A smart desktop is smart furniture that integrates multiple functions. For example, wireless charging, ambient lighting and other functions can be integrated on the smart desktop. Due to the increase in functions, it is necessary to provide a richer control method, such as the patent previously applied by the applicant (publication number CN209450059U) provides a solution for controlling smart desktop functions through touch buttons. Smart desktops can appear as separate furniture, such as coffee tables, or as accessories for other smart furniture, such as deploying smart desktops on the armrests of smart sofas and chairs, or deploying smart desktops on makeup cabinets. It can be seen that subsequent smart desktops can be used as Accessories are added to more smart furniture to improve the integration of smart furniture. However, controlling the smart desktop through keystrokes is still not smart enough, and voice control is the direction of technical improvement in this field.
语音控制中,一般需要考虑功耗和语音识别准确性的问题。其中,由于语音识别需要持续采集声音并将采集到的声音信号通过ADC模块后转换为数字信号,通过处理器执行语音识别算法对这些数字信号进行分析,以分析采集到的声音是否包含触发词或称之为唤醒词,例如,触发词可以是“hey,siri”。当分析出采集的声音包含触发词时,就会唤醒语音助手程序,例如,语音助手程序可以根据从语音中识别出的指令,如调高灯光亮度,控制对应的灯具调高亮度。In voice control, power consumption and voice recognition accuracy generally need to be considered. Among them, since speech recognition needs to continuously collect sounds and convert the collected sound signals into digital signals through the ADC module, the processor executes the speech recognition algorithm to analyze these digital signals to analyze whether the collected sounds contain trigger words or Call it a wake-up word. For example, a trigger word could be "hey, siri." When it is analyzed that the collected sound contains a trigger word, the voice assistant program will be awakened. For example, the voice assistant program can control the corresponding lamp to increase the brightness according to the instructions recognized from the voice, such as increasing the brightness of the lights.
可见,持续监控智能家具所处环境的声音的方案,造成功耗较高的问题。对此,公开号为CN113470641B的中国专利披露了一种解决方案,即增加一个协处理器作为声音检测器,在主处理器保持在低功率状态时,该协处理器仍然可以持续监测语音信号并进行仅有该协处理器完成语音信号的分析,判断是否存在触发词,只有在判断存在触发词时,才唤醒主处理器作进一步确认。该方案增加功耗比主处理器低的协处理器来代替原先主处理器对声音信号的持续监控行为,可以降低整体设备维持语音识别功能的功耗。It can be seen that the solution of continuously monitoring the sound of the environment in which the smart furniture is located causes the problem of high power consumption. In this regard, the Chinese patent with publication number CN113470641B discloses a solution, namely, adding a coprocessor as a sound detector. When the main processor remains in a low-power state, the coprocessor can still continuously monitor the voice signal and perform analysis of the voice signal by the coprocessor alone to determine whether there is a trigger word. Only when it is determined that there is a trigger word, the main processor is awakened for further confirmation. This solution adds a coprocessor with lower power consumption than the main processor to replace the original main processor's continuous monitoring of the sound signal, which can reduce the power consumption of the overall device to maintain the voice recognition function.
但是在尝试将上述降低语音识别功耗的方案应用到智能桌面等智能家具时,发明人发现上述现有技术仍然存在需要改进之处:However, when trying to apply the above-mentioned solution for reducing speech recognition power consumption to smart furniture such as smart desktops, the inventor found that the above-mentioned existing technology still needs improvement:
1、需要额外在智能桌面的控制装置中增加一块具有数据处理能力的声音芯片作为协处理器,才能实现持续对智能桌面周围声音进行监控。1. It is necessary to add an additional sound chip with data processing capabilities as a co-processor to the control device of the smart desktop in order to continuously monitor the sounds around the smart desktop.
2、虽然协处理器的功耗低于主处理器,但是协处理器为了对采集到的每笔模拟语音信号转换成数字信号进行分析判断是否存在触发词,就必须要保持其ADC模块持续处于启动状态,增大处理器的资源开销。2. Although the power consumption of the coprocessor is lower than that of the main processor, in order to convert each collected analog voice signal into a digital signal and analyze and determine whether there is a trigger word, the coprocessor must keep its ADC module in the started state, which increases the resource overhead of the processor.
3、环境嘈杂时,会容易采集到环境噪声,而对于每一笔环境噪声,协处理器都被频繁触发去执行算法进行分析,增大处理器的资源开销。3. When the environment is noisy, it is easy to collect environmental noise. For each environmental noise, the coprocessor is frequently triggered to execute the algorithm for analysis, which increases the resource overhead of the processor.
发明内容Summary of the invention
本发明的目的在于提供一种智能桌面的控制装置及控制方法,在无需增加协处理器情况下监控智能桌面周围的声音,并只有在检测到有人说话的声音才启动处理器的ADC模块,无需持续处于启动状态,同时处理器仅对采集到的人声信号执行语音识别算法,不会被环境噪声频繁触发,从而降低处理器的资源开销。The purpose of the present invention is to provide a control device and a control method for a smart desktop, which can monitor the sounds around the smart desktop without adding a co-processor, and only start the ADC module of the processor when the sound of someone speaking is detected, without adding a co-processor. It is continuously in the startup state. At the same time, the processor only executes the speech recognition algorithm on the collected human voice signals and will not be frequently triggered by environmental noise, thereby reducing the resource overhead of the processor.
为实现上述发明目的,提供一种智能桌面的控制装置,所述控制装置包括:In order to achieve the above object of the invention, a control device for a smart desktop is provided, and the control device includes:
麦克风,用于采集音频信号;Microphone, used to collect audio signals;
放大电路,用于放大麦克风采集到的音频信号;Amplification circuit, used to amplify the audio signal collected by the microphone;
带通滤波电路,用于过滤非人声频段的音频信号;所述人声频段范围β为:300Hz≤β≤5000Hz;A band-pass filter circuit is used to filter audio signals in non-human voice frequency bands; the human voice frequency band range β is: 300Hz≤β≤5000Hz;
音频输出电路,用于将人声频段的音频信号传递到与执行语音识别算法的芯片的ADC引脚;Audio output circuit, used to transmit the audio signal in the human voice band to the ADC pin of the chip that executes the speech recognition algorithm;
中断触发电路,用于响应接收到的人声频段的音频信号向执行语音识别算法的芯片的中断引脚发送中断触发信号;An interrupt trigger circuit, configured to send an interrupt trigger signal to the interrupt pin of the chip executing the speech recognition algorithm in response to the received audio signal in the human voice frequency band;
执行语音识别算法的芯片用于执行以下程序:The chip that executes the speech recognition algorithm is used to perform the following procedures:
检测到来自目标中断引脚发起的中断时,按照预设参数启动目标ADC模块;所述目标中断引脚为与中断触发电路连接的中断引脚,所述目标ADC模块为与音频输出电路连接的ADC引脚对应的ADC模块。When an interrupt initiated from the target interrupt pin is detected, the target ADC module is started according to the preset parameters; the target interrupt pin is an interrupt pin connected to the interrupt trigger circuit, and the target ADC module is connected to the audio output circuit The ADC module corresponding to the ADC pin.
作为进一步的改进,所述音频输出电路包括电压跟随器,所述电压跟随器的输入端与带通滤波电路的输出端连接,所述电压跟随器的输出端与执行语音识别算法的芯片的ADC引脚连接。As a further improvement, the audio output circuit includes a voltage follower, the input end of the voltage follower is connected to the output end of the bandpass filter circuit, and the output end of the voltage follower is connected to the ADC of the chip that executes the speech recognition algorithm. pin connections.
作为进一步的改进,所述中断触发电路包括同相放大器、整流单元、常闭继电器回路、下降沿信号电路;As a further improvement, the interrupt trigger circuit includes a non-inverting amplifier, a rectifier unit, a normally closed relay circuit, and a falling edge signal circuit;
同相放大器的输入端与带通滤波电路的输出端连接,用于放大人声频段的音频信号;The input end of the non-inverting amplifier is connected to the output end of the bandpass filter circuit, and is used to amplify the audio signal in the human voice frequency band;
整流单元的输入端与同相放大器的输出端连接,用于对人声频段的音频信号进行整流并将整流后的人声频段的音频信号输出给下降沿信号电路;The input end of the rectifier unit is connected to the output end of the non-inverting amplifier, and is used to rectify the audio signal in the vocal frequency band and output the rectified audio signal in the vocal frequency band to the falling edge signal circuit;
下降沿信号电路响应于输入端接收到的人声频段的音频信号,向执行语音识别算法的芯片的中断引脚发送下降沿信号作为中断触发信号;The falling edge signal circuit responds to the audio signal in the human voice band received at the input end and sends a falling edge signal as an interrupt trigger signal to the interrupt pin of the chip executing the speech recognition algorithm;
常闭继电器回路的驱动单元响应执行语音识别算法的芯片的输出引脚输出的高电平信号,关断下降沿信号电路输入端与整流单元输出端之间的连接,以停止下降沿信号电路对人声频段的音频信号的接收。The drive unit of the normally closed relay circuit responds to the high-level signal output by the output pin of the chip executing the speech recognition algorithm, and turns off the connection between the input end of the falling edge signal circuit and the output end of the rectification unit to stop the pairing of the falling edge signal circuit. Reception of audio signals in the human voice band.
另一方面,本发明还提供了一种智能桌面的控制方法,所述方法应用于执行语音识别算法的芯片,所述执行语音识别算法的芯片与上述的智能桌面的控制装置连接,所述方法包括:On the other hand, the present invention also provides a control method for a smart desktop, the method is applied to a chip that executes a speech recognition algorithm, the chip that executes the speech recognition algorithm is connected to the control device of the smart desktop, the method comprises:
检测到来自目标中断引脚发起的中断时,按照预设参数启动目标ADC模块;所述目标中断引脚为与中断触发电路连接的中断引脚,所述目标ADC模块为与音频输出电路连接的ADC引脚对应的ADC模块。When an interrupt initiated from the target interrupt pin is detected, the target ADC module is started according to the preset parameters; the target interrupt pin is an interrupt pin connected to the interrupt trigger circuit, and the target ADC module is connected to the audio output circuit The ADC module corresponding to the ADC pin.
作为进一步的改进,所述方法还包括:As a further improvement, the method also includes:
检测到来自目标中断引脚发起的中断时,根据预设时长开始倒计时;所述倒计时的剩余时长为目标ADC模块处于启动状态的剩余时长;When an interrupt initiated from the target interrupt pin is detected, a countdown starts according to the preset duration; the remaining duration of the countdown is the remaining duration of the target ADC module in the startup state;
倒计时未结束时,持续通过目标输出引脚输出高电平信号;所述目标输出引脚为与常闭继电器回路的驱动单元连接的输出引脚;When the countdown is not over, a high level signal is continuously output through a target output pin; the target output pin is an output pin connected to a driving unit of a normally closed relay circuit;
倒计时结束时,控制目标输出引脚停止输出高电平信号。When the countdown ends, the control target output pin stops outputting a high level signal.
作为进一步的改进,所述方法还包括:As a further improvement, the method also includes:
在倒计时过程中,检测到ADC模块有信号输入则重置倒计时。During the countdown process, if a signal input to the ADC module is detected, the countdown will be reset.
作为进一步的改进,所述智能桌面的控制装置还包括执行语音识别算法的芯片连接的麦克风阵列,所述智能桌面安装在沙发椅的扶手上,所述方法还包括:As a further improvement, the control device of the smart desktop further includes a microphone array connected to a chip that executes a speech recognition algorithm, the smart desktop is mounted on an armrest of a sofa chair, and the method further includes:
检测到来自目标中断引脚发起的中断时,按照预设参数启动与麦克风阵列连接的ADC模块;When an interrupt initiated from the target interrupt pin is detected, the ADC module connected to the microphone array is started according to the preset parameters;
根据麦克风阵列采集到的音频信号对沙发椅进行控制。The sofa chair is controlled based on the audio signals collected by the microphone array.
作为进一步的改进,所述方法包括:As a further improvement, the method comprises:
检测到沙发椅从待机状态切换至使用状态时,将预设的初始范围作为响应音区;When it is detected that the sofa chair switches from standby state to use state, the preset initial range is used as the response sound zone;
当麦克风阵列采集到响应音区的第一笔语音信号时,将当前的用户坐姿标记为初始坐姿并计算第一笔语音信号的声源坐标;When the microphone array collects the first voice signal responding to the sound zone, the current user sitting posture is marked as the initial sitting posture and the sound source coordinates of the first voice signal are calculated;
基于第一音区修正算法计算第一笔语音信号的声源坐标对应的空间范围,将所述空间范围更新为初始坐姿对应的响应音区;Calculate the spatial range corresponding to the sound source coordinates of the first speech signal based on the first sound zone correction algorithm, and update the spatial range to the response sound zone corresponding to the initial sitting posture;
检测到用户坐姿发生第一类改变时,基于第二音区修正算法计算最新用户坐姿对应的响应音区。When a first-type change in the user's sitting posture is detected, the response range corresponding to the latest user's sitting posture is calculated based on the second range correction algorithm.
作为进一步的改进,所述沙发椅包括坐垫和靠背、部署在坐垫上的第一压力传感器阵列以及部署在靠背上的第二压力传感器阵列,所述初始坐姿为仅有坐垫平面上的第一压力传感器阵列被用户触发时的用户坐姿;As a further improvement, the sofa chair includes a seat cushion and a backrest, a first pressure sensor array deployed on the seat cushion, and a second pressure sensor array deployed on the backrest. The initial sitting posture is only the first pressure on the seat cushion plane. The user's sitting position when the sensor array is triggered by the user;
所述基于第一音区修正算法计算第一笔语音信号的声源坐标对应的空间范围,具体包括:The calculation of the spatial range corresponding to the sound source coordinates of the first speech signal based on the first sound zone correction algorithm specifically includes:
根据以下的公式一计算出第一笔语音信号的声源坐标对应的空间范围(x,y,z):The spatial range (x, y, z) corresponding to the sound source coordinates of the first speech signal is calculated according to the following formula 1:
其中,为初始坐姿状态下,坐垫平面上的第一压力传感器阵列中被用户触发的压力传感器组成的几何图形的重心点P0的坐标,r为重心点P0到第一笔语音信号的声源坐标的距离,θ为预设的人体摆动角度,0°≤θ≤45°。in, is the coordinate of the center of gravity point P0 of the geometric figure composed of the pressure sensors triggered by the user in the first pressure sensor array on the seat cushion plane in the initial sitting position, r is the distance from the center of gravity point P0 to the sound source coordinates of the first voice signal , θ is the preset human body swing angle, 0°≤θ≤45°.
作为进一步的改进,用户坐姿发生第一类改变是指,用户坐姿改变前后仅有坐垫平面上的第一压力传感器阵列被用户触发;As a further improvement, the first type of change in the user's sitting posture means that only the first pressure sensor array on the seat cushion plane is triggered by the user before and after the user's sitting posture changes;
所述基于第二音区修正算法计算最新用户坐姿对应的响应音区,具体包括:The calculation of the response zone corresponding to the latest user's sitting posture based on the second zone correction algorithm specifically includes:
获取最新坐姿状态下,坐垫平面上的第一压力传感器阵列中被用户触发的压力传感器组成的几何图形的重心点P1的坐标;Obtain the coordinates of the center of gravity point P1 of the geometric figure composed of the pressure sensors triggered by the user in the first pressure sensor array on the seat cushion plane in the latest sitting posture;
用P1减去P0,得到平移偏移量;Subtract P0 from P1 to get the translation offset ;
将当前响应音区的所有坐标点加上平移偏移量,得到最新坐姿对应的响应音区的所有坐标点。Add the translation offset to all the coordinate points of the current response sound zone to obtain all the coordinate points of the response sound zone corresponding to the latest sitting posture.
有益效果:Beneficial effects:
本发明提供的一种智能桌面的控制装置及控制方法,能够在无需增加协处理器情况下监控智能桌面周围的声音,并只有在检测到有人说话的声音才启动处理器的ADC模块,无需持续处于启动状态,同时处理器仅对采集到的人声信号执行语音识别算法,不会被环境噪声频繁触发,从而降低处理器的资源开销。The invention provides a control device and control method for a smart desktop, which can monitor the sounds around the smart desktop without adding a co-processor, and only starts the ADC module of the processor when the sound of someone speaking is detected, without the need for continuous In the startup state, at the same time, the processor only executes the speech recognition algorithm on the collected human voice signals and will not be frequently triggered by environmental noise, thereby reducing the resource overhead of the processor.
附图说明BRIEF DESCRIPTION OF THE DRAWINGS
下面结合附图和实施例对本发明进一步地说明;The present invention will be further described below in conjunction with the accompanying drawings and examples;
图1为一个实施例中提供的沙发椅的结构示意图。FIG. 1 is a schematic structural diagram of a sofa chair provided in one embodiment.
图2为实施例1中用户坐姿发生第一类改变的示意图。Figure 2 is a schematic diagram of the first type of change in the user's sitting posture in Embodiment 1.
图3为实施例1中用户坐姿发生第一类改变之前,用户与沙发椅接触区域的示意图。FIG. 3 is a schematic diagram of the contact area between the user and the sofa chair before the user's sitting posture undergoes the first type of change in Embodiment 1. FIG.
图4为实施例1中用户坐姿发生第一类改变之后,用户与沙发椅接触区域的示意图。FIG. 4 is a schematic diagram of the contact area between the user and the sofa chair after the user's sitting posture undergoes the first type of change in Embodiment 1. FIG.
图5为实施例1中初始坐姿对应的响应音区的示意图。FIG. 5 is a schematic diagram of a response sound zone corresponding to an initial sitting posture in Example 1.
图6为实施例2中用户坐姿发生第二类改变的示意图。Figure 6 is a schematic diagram of the second type of change in the user's sitting posture in Embodiment 2.
图7为实施例2中用户坐姿发生第二类改变之前,用户与沙发椅接触区域的示意图。Figure 7 is a schematic diagram of the contact area between the user and the sofa chair before the second change in the user's sitting posture occurs in Embodiment 2.
图8为实施例2中用户坐姿发生第二类改变之后,用户与沙发椅接触区域的示意图。Figure 8 is a schematic diagram of the contact area between the user and the sofa chair after the second change in the user's sitting posture in Embodiment 2.
图9为实施例3中用户坐姿发生第三类改变的示意图。Figure 9 is a schematic diagram of the third type of change in the user's sitting posture in Embodiment 3.
图10为实施例3中用户坐姿发生第三类改变之前,用户与沙发椅接触区域的示意图。Figure 10 is a schematic diagram of the contact area between the user and the sofa chair before the third change in the user's sitting posture occurs in Embodiment 3.
图11为实施例3中用户坐姿发生第三类改变之后,用户与沙发椅接触区域的示意图。Figure 11 is a schematic diagram of the contact area between the user and the sofa chair after the third change in the user's sitting posture in Embodiment 3.
图12为实施例4中用户坐姿发生第三类改变的示意图。Figure 12 is a schematic diagram of the third type of change in the user's sitting posture in Embodiment 4.
图13为实施例4中用户坐姿发生第三类改变之前,用户与沙发椅接触区域的示意图。Figure 13 is a schematic diagram of the contact area between the user and the sofa chair before the third change in the user's sitting posture occurs in Embodiment 4.
图14为实施例4中用户坐姿发生第三类改变之后,用户与沙发椅接触区域的示意图。Figure 14 is a schematic diagram of the contact area between the user and the sofa chair after the third change in the user's sitting posture in Embodiment 4.
图15为实施例5中用户坐姿发生第四类改变的示意图。Figure 15 is a schematic diagram of the fourth type of change in the user's sitting posture in Embodiment 5.
图16为一个实施例中提供的智能桌面的控制方法的流程示意图之一。FIG. 16 is a schematic flowchart of a method for controlling a smart desktop provided in one embodiment.
图17为一个实施例中提供的智能桌面的控制方法的流程示意图之二。Figure 17 is a second schematic flowchart of a smart desktop control method provided in one embodiment.
图18为一个实施例中提供的智能桌面的控制装置的电路框图。FIG. 18 is a circuit block diagram of a control device for a smart desktop provided in one embodiment.
图19为一个实施例中提供的智能桌面的控制装置的电路图。Figure 19 is a circuit diagram of a control device of a smart desktop provided in one embodiment.
图20为一个实施例中中断触发电路的电路图。Figure 20 is a circuit diagram of an interrupt trigger circuit in one embodiment.
具体实施方式Detailed ways
参照图18,本实施例中提供的一种智能桌面的控制装置,所述控制装置包括:18, a control device for a smart desktop is provided in this embodiment, and the control device includes:
麦克风11,用于采集音频信号;Microphone 11, used for collecting audio signals;
放大电路12,用于放大麦克风11采集到的音频信号;An amplifier circuit 12, used to amplify the audio signal collected by the microphone 11;
带通滤波电路13,用于过滤非人声频段的音频信号;所述人声频段范围β为:300Hz≤β≤5000Hz;The band-pass filter circuit 13 is used to filter audio signals in non-vocal frequency bands; the vocal frequency band range β is: 300Hz≤β≤5000Hz;
音频输出电路14,用于将人声频段的音频信号传递到与执行语音识别算法的芯片16的ADC引脚;The audio output circuit 14 is used to transmit the audio signal of the human voice frequency band to the ADC pin of the chip 16 that executes the speech recognition algorithm;
中断触发电路15,用于响应接收到的人声频段的音频信号向执行语音识别算法的芯片16的中断引脚发送中断触发信号;The interrupt trigger circuit 15 is used to send an interrupt trigger signal to the interrupt pin of the chip 16 executing the speech recognition algorithm in response to the received audio signal in the human voice frequency band;
执行语音识别算法的芯片16用于执行以下程序:The chip 16 executing the speech recognition algorithm is used to execute the following procedures:
检测到来自目标中断引脚发起的中断时,按照预设参数启动目标ADC模块;所述目标中断引脚为与中断触发电路15连接的中断引脚,所述目标ADC模块为与音频输出电路14连接的ADC引脚对应的ADC模块。When an interrupt initiated from the target interrupt pin is detected, the target ADC module is started according to the preset parameters; the target interrupt pin is an interrupt pin connected to the interrupt trigger circuit 15, and the target ADC module is connected to the audio output circuit 14 The connected ADC pin corresponds to the ADC module.
本实施例中提供的一种智能桌面的控制装置,能够在无需增加协处理器情况下监控智能桌面周围的声音,并只有在检测到有人说话的声音才启动处理器的ADC模块,无需持续处于启动状态,同时处理器仅对采集到的人声信号执行语音识别算法,不会被环境噪声频繁触发,从而降低处理器的资源开销。The control device for a smart desktop provided in this embodiment can monitor the sounds around the smart desktop without adding a co-processor, and only starts the ADC module of the processor when the sound of someone speaking is detected, without the need to continuously In the startup state, at the same time, the processor only executes the speech recognition algorithm on the collected human voice signals and will not be frequently triggered by environmental noise, thereby reducing the resource overhead of the processor.
如图19所示,所述音频输出电路14包括电压跟随器,所述电压跟随器的输入端与带通滤波电路13的输出端连接,所述电压跟随器的输出端与执行语音识别算法的芯片16的ADC引脚连接。As shown in Figure 19, the audio output circuit 14 includes a voltage follower. The input terminal of the voltage follower is connected to the output terminal of the band-pass filter circuit 13. The output terminal of the voltage follower is connected to the output terminal of the speech recognition algorithm. The ADC pin connections of chip 16.
如图20所示,所述中断触发电路15包括同相放大器151、整流单元152、常闭继电器回路153、下降沿信号电路154;As shown in Figure 20, the interrupt trigger circuit 15 includes a non-inverting amplifier 151, a rectifier unit 152, a normally closed relay circuit 153, and a falling edge signal circuit 154;
同相放大器151的输入端与带通滤波电路13的输出端连接,用于放大人声频段的音频信号;The input end of the non-inverting amplifier 151 is connected to the output end of the bandpass filter circuit 13, and is used to amplify the audio signal in the human voice frequency band;
整流单元152的输入端与同相放大器151的输出端连接,用于对人声频段的音频信号进行整流并将整流后的人声频段的音频信号输出给下降沿信号电路154;The input end of the rectifier unit 152 is connected to the output end of the in-phase amplifier 151, and is used to rectify the audio signal in the human voice frequency band and output the rectified audio signal in the human voice frequency band to the falling edge signal circuit 154;
下降沿信号电路154响应于输入端接收到的人声频段的音频信号,向执行语音识别算法的芯片16的中断引脚发送下降沿信号作为中断触发信号;The falling edge signal circuit 154 responds to the audio signal in the human voice band received at the input end and sends a falling edge signal as an interrupt trigger signal to the interrupt pin of the chip 16 that executes the speech recognition algorithm;
常闭继电器回路153的驱动单元响应执行语音识别算法的芯片16的输出引脚输出的高电平信号,关断下降沿信号电路154输入端与整流单元152输出端之间的连接,以停止下降沿信号电路154对人声频段的音频信号的接收。The drive unit of the normally closed relay circuit 153 responds to the high-level signal output by the output pin of the chip 16 that executes the speech recognition algorithm, and turns off the connection between the input terminal of the falling edge signal circuit 154 and the output terminal of the rectifier unit 152 to stop the decline. The signal circuit 154 receives the audio signal in the vocal frequency band.
需要说明的是,本实施例中,将带通滤波电流输出的人声频道的音频信号分为两路,一路通过音频输出电路14,用于将人声频道的音频信号传递的执行语音识别算法的芯片16的ADC引脚,以控制智能桌面。而另一路通过中断触发电路15使得执行语音识别算法的芯片16能够启动目标ADC模块从而对音频输出电路14,传递的音频信号进行语音处理,以控制智能桌面。It should be noted that in this embodiment, the audio signal of the human voice channel output by the band-pass filter current is divided into two channels, and one channel passes through the audio output circuit 14 for transmitting the audio signal of the human voice channel to the execution speech recognition algorithm. The chip has 16 ADC pins to control the smart desktop. The other channel interrupts the trigger circuit 15 so that the chip 16 executing the speech recognition algorithm can start the target ADC module to perform speech processing on the audio signal transmitted by the audio output circuit 14 to control the smart desktop.
另一方面,在一个实施例中还提供了一种智能桌面的控制方法,所述方法应用于执行语音识别算法的芯片16,所述执行语音识别算法的芯片16与上述的智能桌面的控制装置连接,所述方法包括:On the other hand, in one embodiment, a method for controlling a smart desktop is also provided. The method is applied to a chip 16 that executes a speech recognition algorithm. The chip 16 that executes a speech recognition algorithm is connected to the above-mentioned control device of the smart desktop. Connection, the method includes:
检测到来自目标中断引脚发起的中断时,按照预设参数启动目标ADC模块;所述目标中断引脚为与中断触发电路15连接的中断引脚,所述目标ADC模块为与音频输出电路14连接的ADC引脚对应的ADC模块。When an interrupt initiated from the target interrupt pin is detected, the target ADC module is started according to the preset parameters; the target interrupt pin is an interrupt pin connected to the interrupt trigger circuit 15, and the target ADC module is connected to the audio output circuit 14 The connected ADC pin corresponds to the ADC module.
由于只有人声频段的音频信号可以通过带通滤波电路13,所以本实施例中,芯片16的ADC模块只有在检测到人声信号才会启动,而无需一直启动,由于ADC模块不启动,就没有音频信号需要芯片16处理,节省芯片16的开销。另一方面,一旦检测到人声信号,中断触发电路15就会触发芯片16启动对应的ADC模块,不会错过对人声信号的监控,并且只会对人声信号进行响应,可以排除环境噪声造成的无启动。Since only the audio signal in the human voice frequency band can pass through the bandpass filter circuit 13, in this embodiment, the ADC module of the chip 16 will only be started when a human voice signal is detected, and does not need to be started all the time. Since the ADC module is not started, it will There is no audio signal that needs to be processed by the chip 16, thus saving the cost of the chip 16. On the other hand, once the human voice signal is detected, the interrupt trigger circuit 15 will trigger the chip 16 to start the corresponding ADC module, so that the monitoring of the human voice signal will not be missed, and only the human voice signal will be responded to, which can eliminate environmental noise. Caused by no startup.
在一个实施例中,所述方法还包括:In one embodiment, the method further includes:
步骤S401,检测到来自目标中断引脚发起的中断时,根据预设时长开始倒计时;所述倒计时的剩余时长为目标ADC模块处于启动状态的剩余时长;Step S401: When an interrupt initiated from the target interrupt pin is detected, a countdown is started according to the preset duration; the remaining duration of the countdown is the remaining duration of the target ADC module in the startup state;
步骤S402,倒计时未结束时,持续通过目标输出引脚输出高电平信号;所述目标输出引脚为与常闭继电器回路153的驱动单元连接的输出引脚;Step S402, when the countdown has not ended, continue to output a high-level signal through the target output pin; the target output pin is an output pin connected to the drive unit of the normally closed relay circuit 153;
步骤S403,倒计时结束时,控制目标输出引脚停止输出高电平信号。Step S403: When the countdown ends, the target output pin is controlled to stop outputting a high-level signal.
如图20所示,当持续通过目标输出引脚输出高电平信号时,会使得常闭继电器k断开,从而麦克风11MIC采集到的人声信号不会再触发芯片16产生中断,从而不会影响芯片16对语音信号的处理过程。因此,本实施例提供了在芯片16处理人声信号过程中,对该电路中人声信号会导致频繁中断问题的解决方案。预设时长为3分钟时,即ADC模块处于启动状态的剩余时长为3分钟,倒计时还剩1分钟,则ADC模块处于启动状态的剩余时长为1分钟,如果倒计时结束,就需要停止目标ADC模块的启动状态,不再响应ADC引脚输入的音频信号。由于停止了ADC模块对音频信号的响应,所以就需要恢复中断触发电路15对中断引脚的触发能力,即控制目标输出引脚停止输出高电平信号,此时Vin不再为高电平,则电磁铁N断电,常闭继电器恢复到连通状态,此时整流单元152和下降沿信号电路154重新连通。一旦麦克风11MIC再次检测到人声频段的音频信号就可以再次触发中断。As shown in FIG20, when a high-level signal is continuously output through the target output pin, the normally closed relay k will be disconnected, so that the human voice signal collected by the microphone 11MIC will no longer trigger the chip 16 to generate an interrupt, thereby not affecting the chip 16's processing of the voice signal. Therefore, this embodiment provides a solution to the problem that the human voice signal in the circuit causes frequent interruptions during the chip 16 processing of the human voice signal. When the preset duration is 3 minutes, that is, the remaining duration of the ADC module in the startup state is 3 minutes, and there is 1 minute left in the countdown, then the remaining duration of the ADC module in the startup state is 1 minute. If the countdown ends, it is necessary to stop the startup state of the target ADC module and no longer respond to the audio signal input by the ADC pin. Since the ADC module stops responding to the audio signal, it is necessary to restore the triggering ability of the interrupt trigger circuit 15 to the interrupt pin, that is, control the target output pin to stop outputting a high-level signal. At this time, Vin is no longer a high level, then the electromagnet N is powered off, and the normally closed relay is restored to the connected state. At this time, the rectifier unit 152 and the falling edge signal circuit 154 are reconnected. Once the microphone 11MIC detects an audio signal in the human voice frequency band again, the interrupt can be triggered again.
具体的,所述方法还包括:Specifically, the method also includes:
在倒计时过程中,检测到ADC模块有信号输入则重置倒计时。During the countdown process, if a signal input to the ADC module is detected, the countdown will be reset.
可以理解的是,只要在预设时长,例如3分钟之内,ADC模块接收到人声频段的语音信号就应该重置倒计时,保持持续通过ADC模块监听音频信号。因此,重置倒计时可以保持连续监听用户的音频信号,在用户说话内容较长时,保持持续的监听效果,提升用户体验。It is understandable that as long as the ADC module receives a voice signal in the human voice band within a preset time period, such as 3 minutes, it should reset the countdown and continue to monitor the audio signal through the ADC module. Therefore, resetting the countdown can continue to monitor the user's audio signal. When the user speaks for a long time, the continuous monitoring effect is maintained and the user experience is improved.
另一方面,所述智能桌面的控制装置还包括执行语音识别算法的芯片16连接的麦克风阵列,所述智能桌面安装在沙发椅的扶手上,所述方法还包括:On the other hand, the control device of the smart desktop also includes a microphone array connected to the chip 16 that executes the speech recognition algorithm. The smart desktop is installed on the armrest of the sofa chair. The method also includes:
检测到来自目标中断引脚发起的中断时,按照预设参数启动与麦克风阵列连接的ADC模块;When an interrupt is detected from a target interrupt pin, the ADC module connected to the microphone array is started according to preset parameters;
根据麦克风阵列采集到的音频信号对沙发椅进行控制。The sofa chair is controlled according to the audio signal collected by the microphone array.
具体的,以下将通过实施例1至5描述对沙发椅的控制方法。Specifically, the method of controlling the sofa chair will be described below through Embodiments 1 to 5.
在对实施例详细描述之前,为了理解本申请提出的沙发椅的控制方法对现有技术作出的贡献,有必要对相关背景技术简单说明如下:Before describing the embodiments in detail, in order to understand the contribution of the sofa chair control method proposed in this application to the prior art, it is necessary to briefly describe the relevant background technology as follows:
智能家具提供智能控制旨在提高用户操作的便利性,智能家具集成语音识别后,用户仅需在使用智能家具过程中通过语音说出触发指令,智能家具的控制器通过采集用户的语音信号并进行语音识别,便可根据语音识别结果对智能家具进行控制,而不会影响用户当前对智能家具功能的使用过程。即用户不再需要寻找遥控器或实体按钮来操作功能,提高用户操作的便利性。例如,用户闭眼躺在智能沙发椅上之后,想要调整沙发椅的按摩功能、加热功能时,可以在不睁开眼也不用动的情况下,通过语音说出触发指令,直接控制沙发椅执行相应的加热功能等。Smart furniture provides intelligent control to improve the convenience of user operations. After smart furniture integrates voice recognition, users only need to speak triggering instructions through voice during use of smart furniture. The controller of smart furniture collects the user's voice signal and performs operations. With voice recognition, smart furniture can be controlled based on the voice recognition results without affecting the user's current use of smart furniture functions. That is, users no longer need to look for remote controls or physical buttons to operate functions, improving user convenience. For example, after the user closes his eyes and lies on the smart sofa chair, and wants to adjust the massage function and heating function of the sofa chair, he can directly control the sofa chair by speaking the trigger command without opening his eyes or moving. Perform corresponding heating functions, etc.
沙发椅等智能家具放置在客厅等作为家庭公共空间的地方,一次只有一个使用用户,在用户使用该智能家具并使用语音识别进行功能控制时,容易受到处于同一个空间下的其他人的说话声或者电视声音的干扰。例如,紧挨着沙发椅的椅子上用户说话声或电视播放声音与用户声音混合,导致声源之间的干扰,使得无法有效地分辨和提取目标声源,降低语音信号的清晰度和质量。经检索,相关技术中可以利用麦克风阵列的声源定位并结合波束成形(Beamforming)技术来增强指定音区的语音信号,实现有效分辨和提取目标声源,从而提升语音识别的准确性。如专利文献1披露了可以采用麦克风阵列技术来实现仅采集指定音区(主音区)的声音用于语音控制,并且屏蔽剩余音区的方案,该方案可以提升语音识别的准确性。并且,专利文献1进一步还披露了由于用户移动而导致的用户声源跨音区的解决方案,即用户移动到其他音区之后,认为用户声源从之前的主音区跨入另一个音区,所以需要更新用户所处的最新位置为主音区。Smart furniture such as sofas and chairs are placed in a public space in the home such as the living room. Only one user can use it at a time. When the user uses the smart furniture and uses voice recognition to control functions, he or she is easily affected by the voices of other people in the same space. Or interference from TV sound. For example, the sound of a user speaking on a chair next to a sofa or the sound of a TV playing is mixed with the user's voice, causing interference between sound sources, making it impossible to effectively distinguish and extract the target sound source, and reducing the clarity and quality of the speech signal. After searching, it was found that in related technologies, sound source positioning of microphone arrays and beamforming technology can be used to enhance the speech signal in the specified sound area, so as to effectively distinguish and extract the target sound source, thus improving the accuracy of speech recognition. For example, Patent Document 1 discloses a solution that can use microphone array technology to collect only the sounds of a designated sound area (main sound area) for voice control and shield the remaining sound areas. This solution can improve the accuracy of speech recognition. Moreover, Patent Document 1 further discloses a solution to the problem that the user's sound source crosses the sound zone due to the user's movement. That is, after the user moves to another sound zone, it is considered that the user's sound source crosses from the previous main sound zone to another sound zone. Therefore, it is necessary to update the user's latest location as the main sound zone.
如果将专利文献1的方案应用到沙发椅等智能家具场景时,可以将沙发椅边框(包括沙发椅的扶手、靠背的外侧面等)包围的区域作为主音区,通过只对主音区的声音进行采集和响应,可以提高声音信号质量和语音识别准确性,并且用户使用沙发椅(即躺在沙发椅或坐在沙发椅上的状态)时,由于用户不移动,也不会面临专利文献1中的用户声源跨音区问题。If the solution in Patent Document 1 is applied to a smart furniture scene such as a sofa chair, the area surrounded by the sofa chair frame (including the armrests of the sofa chair, the outer side of the backrest, etc.) can be used as the main sound area, and only the sound in the main sound area can be processed Collection and response can improve the sound signal quality and speech recognition accuracy, and when the user uses the sofa chair (that is, lying on the sofa chair or sitting on the sofa chair), since the user does not move, he will not face the problem in Patent Document 1 The user's sound source crosses the vocal range problem.
但是,在沙发椅的语音控制场景中,却存在非用户声源跨音区的问题,即沙发椅一般布置在客厅并与其他的椅子紧挨在一起,所以在客厅活动的其他人或者坐在紧挨着沙发椅的椅子上的其他人的头部很容易挨着沙发椅主音区的边界,甚至进入到沙发椅的主音区范围内,该现象定义为非用户声源跨音区问题。However, in the voice control scenario of sofa chairs, there is a problem of non-user sound sources crossing the sound zone. That is, sofa chairs are generally arranged in the living room and closely together with other chairs, so other people moving in the living room or sitting The heads of other people sitting on the chairs next to the sofa chair can easily touch the boundary of the main sound range of the sofa chair, or even enter the main sound range of the sofa chair. This phenomenon is defined as a non-user sound source cross-region problem.
由于沙发椅等智能家具的使用场景是公共空间的多人场景,并且存在非用户声源跨音区问题,会导致从沙发椅的主音区采集的目标语音除了包括用户声音之外,还混合有紧挨着沙发椅的椅子上用户说话声,降低语音信号的清晰度和质量以及语音识别准确性。Since the use scenario of smart furniture such as sofas and chairs is a multi-person scenario in a public space, and there is a problem of non-user sound sources crossing sound zones, the target speech collected from the main sound zone of the sofa chair will not only include the user's voice, but also be mixed with The sound of users talking in chairs next to the sofa chair reduces the clarity and quality of the speech signal and the accuracy of speech recognition.
专利文献1,中国专利,公开号,CN111986678B,专利名称,一种多路语音识别的语音采集方法、装置,公开日,2023-12-29。Patent document 1, Chinese patent, publication number, CN111986678B, patent name, a speech collection method and device for multi-channel speech recognition, publication date, 2023-12-29.
本发明提供的一种智能家具控制器及控制方法, 能够根据用户的坐姿动态调整智能家具的响应音区的范围与用户的嘴部位置相匹配,得到以声源为中心的范围更小的响应音区,降低非用户声源进入响应音区的几率,从而获得更高质量的语音信号,提升语音识别准确性。以下通过实施例1至5进行详细说明。The present invention provides a smart furniture controller and control method, which can dynamically adjust the range of the response sound zone of the smart furniture to match the user's mouth position according to the user's sitting posture, obtain a smaller response sound zone centered on the sound source, reduce the probability of non-user sound sources entering the response sound zone, thereby obtaining a higher quality voice signal and improving the accuracy of voice recognition. The following is a detailed description through Examples 1 to 5.
参照图1,本申请提供的智能桌面的控制方法适用于图1中的沙发椅,该沙发椅包括坐垫和靠背、部署在坐垫上的第一压力传感器阵列以及部署在靠背上的第二压力传感器阵列。如图1所示,坐垫所在平面为麦克风阵列标定的直角坐标系的xoy平面,即坐垫所在平面的z分量为0,靠背所在平面为yoz平面,即靠背所在平面的x分量为0,坐标系原点O在上述两个平面的交点位置,靠背能够绕y轴转动以调节靠背的角度。Referring to Figure 1, the control method of the smart desktop provided by this application is applicable to the sofa chair in Figure 1. The sofa chair includes a seat cushion and a backrest, a first pressure sensor array deployed on the seat cushion, and a second pressure sensor deployed on the backrest. array. As shown in Figure 1, the plane where the seat cushion is located is the xoy plane of the rectangular coordinate system calibrated by the microphone array, that is, the z component of the plane where the seat cushion is located is 0, and the plane where the backrest is located is the yoz plane, that is, the x component of the plane where the backrest is located is 0, and the coordinate system The origin O is at the intersection of the above two planes, and the backrest can rotate around the y-axis to adjust the angle of the backrest.
可以理解的是,通过麦克风阵列,可以选择空间中任意一个特定区域的声音进行接收,同时屏蔽其他区域。这主要是通过波束成形(Beamforming)技术和声源定位技术实现的。具体实现步骤如下:It is understandable that through the microphone array, the sound in any specific area in the space can be selected for reception, while other areas are blocked. This is mainly achieved through beamforming technology and sound source positioning technology. The specific implementation steps are as follows:
声源定位:首先,需要确定目标声源的位置。这通常通过分析麦克风阵列捕捉到的声音信号来完成。利用到达不同麦克风的声音信号之间的时间差(Time Difference ofArrival, TDOA),可以估算出声源在空间中的位置。Sound source localization: First, the location of the target sound source needs to be determined. This is usually done by analyzing the sound signals captured by the microphone array. The location of the sound source in space can be estimated by using the time difference (TDOA) between the sound signals arriving at different microphones.
波束成形(Beamforming):一旦确定了声源的位置,就可以使用波束成形技术来“指向”这个特定区域。波束成形是一种信号处理技术,它通过调整麦克风阵列中每个麦克风的信号,使得来自特定方向的信号得到增强,而来自其他方向的信号被抑制。这种方法可以创建一个高度定向的感应区域,即“波束”。Beamforming: Once the location of a sound source is determined, beamforming technology can be used to "point" that specific area. Beamforming is a signal processing technique that adjusts the signal from each microphone in a microphone array so that signals from specific directions are enhanced and signals from other directions are suppressed. This method creates a highly directional sensing area, or "beam."
数字信号处理:波束成形通过数字信号处理算法实现。这些算法会计算如何调整麦克风阵列中每个麦克风的信号,以便聚焦于目标区域。这通常涉及信号的时间延迟、加权和相位调整。Digital signal processing: Beamforming is achieved through digital signal processing algorithms. These algorithms calculate how to adjust the signal from each microphone in the microphone array to focus on the target area. This typically involves time delay, weighting, and phase adjustment of the signal.
动态调整:在实际应用中,如果目标声源在空间中移动,系统可以动态调整波束的指向,以便持续聚焦在移动的声源上。Dynamic adjustment: In practical applications, if the target sound source moves in space, the system can dynamically adjust the direction of the beam to continuously focus on the moving sound source.
以下结合附图和实施例介绍本申请对于现有技术作出的技术贡献。The technical contribution made by this application to the prior art will be introduced below with reference to the drawings and embodiments.
实施例1:Embodiment 1:
如图16所示,本实施例提供了一种智能桌面的控制方法,所述方法包括:As shown in FIG. 16 , this embodiment provides a method for controlling a smart desktop, the method comprising:
步骤S101,检测到沙发椅从待机状态切换至使用状态时,将预设的初始范围作为响应音区。Step S101: When it is detected that the sofa chair switches from the standby state to the use state, the preset initial range is used as the response sound zone.
具体的,待机状态是指没有用户坐在沙发椅上的状态,使用状态是指有用户坐在沙发椅上的状态,状态的判断可以通过第一压力传感器阵列是否被触发来实现,例如第一压力传感器阵列中的一个或多个压力传感器存在读数,则认为处于使用状态,否则认为处于待机状态。Specifically, the standby state refers to a state when no user is sitting on the sofa chair, and the usage state refers to a state when a user is sitting on the sofa chair. The state can be judged by whether the first pressure sensor array is triggered, for example, the first If there is a reading on one or more pressure sensors in the pressure sensor array, it is considered to be in use, otherwise it is considered to be in standby state.
如图1所示,初始范围是指靠背平面沿正x方向移动到坐垫x值较大的一侧时扫过的空间范围,靠背的高度设置为1m,在正常成年人用户挺直腰坐在沙发椅上时,用户的嘴巴会位于初始范围内。当用户坐在沙发椅上需要弯腰低头时,其嘴巴也会位于初始范围内。因此,将预设的初始范围作为响应音区,可以确保仅采集该初始范围内的语音信号作为语音识别模块的输入,以提升语音信号质量,提高语音识别准确性。As shown in Figure 1, the initial range refers to the spatial range swept when the backrest plane moves along the positive x direction to the side of the cushion with a larger x value. The height of the backrest is set to 1m. When a normal adult user sits upright on a sofa chair, the user's mouth will be within the initial range. When the user needs to bend down and lower his head when sitting on a sofa chair, his mouth will also be within the initial range. Therefore, using the preset initial range as the response sound area can ensure that only the voice signal within the initial range is collected as the input of the voice recognition module, so as to improve the quality of the voice signal and the accuracy of voice recognition.
步骤S102,当麦克风阵列采集到响应音区的第一笔语音信号时,将当前的用户坐姿标记为初始坐姿并计算第一笔语音信号的声源坐标。Step S102: When the microphone array collects the first voice signal in the response sound zone, the current sitting posture of the user is marked as the initial sitting posture and the sound source coordinates of the first voice signal are calculated.
举例而已,用户在14:23坐在沙发椅上,所以在14:23检测到沙发椅从待机状态切换至使用状态,此时的响应音区为预设的初始范围。在用户坐下2分钟后,即14:25的时候,用户通过嘴巴在响应音区内发出语音“小智小智,启动加热功能”,即第一笔语音信号,由于麦克风阵列仅对响应音区内的语音进行采集并发送给语音识别模块,因此可以提高语音信号质量和语音识别准确性。通过声源定位技术,可以计算出第一笔语音信号的声源,即嘴巴,在麦克风阵列标定的直角坐标系中的坐标,即声源坐标。For example, the user is sitting on the sofa chair at 14:23, so it is detected that the sofa chair switches from the standby state to the use state at 14:23. The response sound zone at this time is the preset initial range. Two minutes after the user sat down, that is, at 14:25, the user uttered the voice "Xiaozhi Xiaozhi, start the heating function" through his mouth in the response sound area, which is the first voice signal. Since the microphone array only responds to the response sound area, The speech in the area is collected and sent to the speech recognition module, so the speech signal quality and speech recognition accuracy can be improved. Through sound source localization technology, the sound source of the first speech signal, that is, the mouth, can be calculated. The coordinates in the rectangular coordinate system calibrated by the microphone array, that is, the sound source coordinates.
步骤S103,基于第一音区修正算法计算第一笔语音信号的声源坐标对应的空间范围,将所述空间范围更新为初始坐姿对应的响应音区。Step S103: Calculate the spatial range corresponding to the sound source coordinates of the first speech signal based on the first sound zone correction algorithm, and update the spatial range to the response sound zone corresponding to the initial sitting posture.
第一音区修正算法具体在以下的实施例1和实施例2中给出。通过采用第一音区修正算法可以调整响应音区的范围,能够根据用户的坐姿动态调整沙发椅的响应音区的范围与用户的嘴部位置相匹配,得到以声源为中心的范围更小的响应音区,降低非用户声源进入响应音区的几率,从而获得更高质量的语音信号,提升语音识别准确性。The first vocal range correction algorithm is specifically given in the following Embodiment 1 and Embodiment 2. By using the first sound zone correction algorithm, the range of the response sound zone can be adjusted, and the range of the response sound zone of the sofa chair can be dynamically adjusted according to the user's sitting posture to match the user's mouth position, resulting in a smaller range centered on the sound source. The response sound area reduces the probability of non-user sound sources entering the response sound area, thereby obtaining higher quality speech signals and improving speech recognition accuracy.
步骤S104,检测到用户坐姿发生第一类改变时,基于第二音区修正算法计算最新用户坐姿对应的响应音区。Step S104: When it is detected that the first type of change in the user's sitting posture occurs, the response sound zone corresponding to the latest user's sitting posture is calculated based on the second sound zone correction algorithm.
可以理解的是,在已经响应音区从初始范围更新为与初始坐姿匹配的空间范围之后,后续一旦检测到用户坐姿发生改变就会立即动态地修正响应音区的范围,以保证用户在动态使用沙发椅过程中可以持续保持语音信号质量,从而提高语音识别准确性。It can be understood that after the response sound zone has been updated from the initial range to the spatial range that matches the initial sitting posture, once a change in the user's sitting posture is detected, the range of the response sound zone will be immediately and dynamically corrected to ensure that the user can continue to maintain the quality of the voice signal during the dynamic use of the sofa chair, thereby improving the accuracy of voice recognition.
如图2所示,用户不接触靠背坐在沙发椅上的姿势定义为第一姿态,此时,仅有坐垫平面上的第一压力传感器阵列被用户触发时的用户坐姿。本实施例中,初始坐姿为第一姿态。As shown in Figure 2, the posture of the user sitting on the sofa chair without touching the backrest is defined as the first posture. At this time, only the user's sitting posture when the first pressure sensor array on the seat cushion plane is triggered by the user. In this embodiment, the initial sitting posture is the first posture.
本实施例中,所述基于第一音区修正算法计算第一笔语音信号的声源坐标对应的空间范围,具体包括:In this embodiment, the calculation of the spatial range corresponding to the sound source coordinates of the first speech signal based on the first sound zone correction algorithm specifically includes:
根据以下的公式一计算出第一笔语音信号的声源坐标对应的空间范围(x,y,z):Calculate the spatial range (x, y, z) corresponding to the sound source coordinates of the first speech signal according to the following formula 1:
其中,为初始坐姿状态下,坐垫平面上的第一压力传感器阵列中被用户触发的压力传感器组成的几何图形的重心点P0的坐标,r为重心点P0到第一笔语音信号的声源坐标的距离,θ为预设的人体摆动角度,0°≤θ≤45°。in, is the coordinate of the center of gravity point P0 of the geometric figure composed of the pressure sensors triggered by the user in the first pressure sensor array on the seat cushion plane in the initial sitting position, r is the distance from the center of gravity point P0 to the sound source coordinates of the first voice signal , θ is the preset human body swing angle, 0°≤θ≤45°.
如图3所示,初始坐姿时,沙发椅上靠近x轴,被虚线方框围起来的4个压力传感器被用户接触并施加压力,将被接触的4个压力传感器用图3中的实线连接成一个四边形,其重心点为P0。需要说明的是,各压力传感器在麦克风阵列标定的坐标系中的坐标都是已经在出厂前标定好的并写入到存储设备中,可以随时读取得到。如图5所示,第一笔语音信号的声源坐标为Q,m所表示的虚线为坐垫平面,即xoy平面,S1用于在图5中表示被用户触发的压力传感器组成的几何图形的重心点的坐标。可知,Q到S1的距离就是用户臀部与坐垫接触位置到嘴巴的距离,一般地,当用户坐姿不变坐在沙发上时,用户可能会做出弯腰、低头或转头等动作,这些动作会导致用户嘴巴,即发出声音的声源位置发生改变,因此通过公式一在将响应音区从初始范围缩小的同时,还留有符合人体上半身活动余量,无需频繁修改响应音区。如图5所示,通过公式一将响应音区的范围从初始范围对应的六面体缩小到一个类似圆锥体的范围,提升了语音信号质量的同时,无需频繁修改响应音区。As shown in Figure 3, in the initial sitting position, the four pressure sensors surrounded by the dotted squares on the sofa chair close to the x-axis are touched and pressure is applied by the user, and the four touched pressure sensors are connected by the solid lines in Figure 3 to form a quadrilateral, whose center of gravity is P0. It should be noted that the coordinates of each pressure sensor in the coordinate system calibrated by the microphone array have been calibrated before leaving the factory and written into the storage device, and can be read at any time. As shown in Figure 5, the sound source coordinates of the first voice signal are Q, the dotted line represented by m is the cushion plane, that is, the xoy plane, and S1 is used to represent the coordinates of the center of gravity of the geometric figure composed of the pressure sensors triggered by the user in Figure 5. It can be seen that the distance from Q to S1 is the distance from the user's buttocks to the seat cushion to the mouth. Generally, when the user sits on the sofa without changing his sitting posture, he may bend, lower his head, or turn his head. These actions will cause the user's mouth, that is, the position of the sound source of the sound to change. Therefore, Formula 1 is used to reduce the response sound area from the initial range while leaving a margin that is consistent with the upper body movement of the human body, and there is no need to frequently modify the response sound area. As shown in Figure 5, Formula 1 reduces the range of the response sound area from the hexahedron corresponding to the initial range to a range similar to a cone, which improves the quality of the voice signal without frequently modifying the response sound area.
本实施例中,图3中这4个压力传感器有读数的状态用于表示初始坐姿。例如沙发椅上的压力传感器表示用户坐姿时,可以利用压力传感器是否有读数来构造用户坐姿的数据结构,已知第一压力传感器阵列(M1)有12个压力传感器,第二压力传感器阵列(M2)有16个压力传感器,对各压力传感器编号,得到由两个二进制数组表示的用户坐姿数据,M1=[000000011110],M2=[0000000000000000],该数据表示当前用户坐姿下,第一压力传感器阵列中编号为2、3、4、5的压力传感器有读取,第二压力传感器阵列未被触发。如图4所示其用户坐姿数据为M1=[000110011000],M2=[0000000000000000],该数据表示当前用户坐姿下,第一压力传感器阵列中编号为4、5、8、9的压力传感器有读取,第二压力传感器阵列未被触发。显然图3和图4都属于第一姿态,在都属于第一姿态情形,发生的用户坐姿的改变称之为第一类改变。In this embodiment, the state of readings from the four pressure sensors in Figure 3 is used to represent the initial sitting posture. For example, when the pressure sensor on the sofa chair indicates the user's sitting posture, you can use whether the pressure sensor has a reading to construct the data structure of the user's sitting posture. It is known that the first pressure sensor array (M1) has 12 pressure sensors, and the second pressure sensor array (M2 ) There are 16 pressure sensors. Number each pressure sensor to obtain user sitting posture data represented by two binary arrays, M1=[000000011110], M2=[0000000000000000]. This data represents the first pressure sensor array under the current user sitting posture. The pressure sensors numbered 2, 3, 4, and 5 are read, but the second pressure sensor array is not triggered. As shown in Figure 4, the user's sitting posture data is M1=[000110011000], M2=[0000000000000000]. This data indicates that under the current user's sitting posture, the pressure sensors numbered 4, 5, 8, and 9 in the first pressure sensor array have readings. Take, the second pressure sensor array is not triggered. Obviously, both Figures 3 and 4 belong to the first posture. In both cases, the change in the user's sitting posture is called the first type of change.
具体的,用户坐姿发生第一类改变是指,用户坐姿改变前后仅有坐垫平面上的第一压力传感器阵列被用户触发。Specifically, the first type of change in the user's sitting posture means that only the first pressure sensor array on the seat cushion plane is triggered by the user before and after the user's sitting posture changes.
所述基于第二音区修正算法计算最新用户坐姿对应的响应音区,具体包括:The calculation of the response zone corresponding to the latest user's sitting posture based on the second zone correction algorithm specifically includes:
步骤S201,获取最新坐姿状态下,坐垫平面上的第一压力传感器阵列中被用户触发的压力传感器组成的几何图形的重心点P1的坐标。Step S201: Obtain the coordinates of the center of gravity point P1 of the geometric figure composed of the pressure sensors triggered by the user in the first pressure sensor array on the seat cushion plane in the latest sitting posture.
图3为用户坐姿发生第一类改变前的示意图,图4为用户坐姿发生第一类改变后的示意图,可知,用户坐姿发生第一类改变前后都维持第一姿态,但是用户的臀部在坐垫平面上挪动了位置,通过用户坐姿数据可以判断出用户坐姿发生了改变。已知各传感器的坐标,根据求几何图形重心的方法可以容易求出P0和p1,此处不再赘述。Figure 3 is a schematic diagram before the first type of change in the user's sitting posture. Figure 4 is a schematic diagram after the first type of change in the user's sitting posture. It can be seen that the user maintains the first posture before and after the first type of change in the user's sitting posture, but the user's buttocks are in the seat cushion. The position has moved on the plane, and it can be determined from the user's sitting posture data that the user's sitting posture has changed. Knowing the coordinates of each sensor, P0 and p1 can be easily obtained based on the method of finding the center of gravity of geometric figures, which will not be described again here.
步骤S202,用P1减去P0,得到平移偏移量;Step S202, subtract P0 from P1 to obtain the translation offset. ;
步骤S203,将当前响应音区的所有坐标点加上平移偏移量,得到最新坐姿对应的响应音区的所有坐标点。Step S203: Add the translation offset to all the coordinate points of the current response sound area to obtain all the coordinate points of the response sound area corresponding to the latest sitting posture.
可以理解的是,在步骤S201至步骤S203中,无需在用户坐姿发生第一类改变后重新使用公式一求出最新坐姿对应的响应音区的所有坐标点,而是直接利用平移偏移量与发生第一类改变前的响应音区(即步骤S203中的当前响应音区)进行简单加和运算即可,降低了更新响应音区的运算量,节省了计算机资源。It can be understood that in steps S201 to S203, there is no need to reuse Formula 1 to find all the coordinate points of the response sound zone corresponding to the latest sitting posture after the first type of change in the user's sitting posture. Instead, the translation offset and The response sound area before the first type of change occurs (ie, the current response sound area in step S203) can be simply summed, which reduces the calculation amount of updating the response sound area and saves computer resources.
进一步的,所述方法还包括:Further, the method also includes:
利用第三音区修正算法对公式一进行修正;Use the third range correction algorithm to correct Formula 1;
修正后的公式一为:The revised formula 1 is:
其中,d为通过实验预先测定的坐垫下沉深度;Among them, d is the seat cushion sinking depth measured in advance through experiments;
通过以下步骤确定坐垫下沉深度d:Determine the seat cushion sinking depth d through the following steps:
获取坐垫上的第一压力传感器阵列测得的压力值;Obtain the pressure value measured by the first pressure sensor array on the seat cushion;
根据所述压力值查表确定坐垫下沉深度d。Look up the table according to the pressure value to determine the seat cushion sinking depth d.
如图5所示,m表示的虚线为待机状态下的坐垫平面,即x0y平面,而n表示的虚线为使用状态下的坐垫平面,可以看到,n平面的z分量要小于m平面的z分量,意味着用户坐下时使得坐垫平面发生下沉,而公式一中是以坐垫平面不下沉来进行理想状态计算,所以考虑到下沉因素,需要采用修正后的公式一。而下沉的深度d可以通过查表确定,下沉的深度不仅与用户体重有关,还与沙发材质有关,所以厂商在出厂前会预先对不同型号的沙发都通过实验预先测定了不同压力值下的下沉深度,用于修正公式一,从而提升划分响应音区的精度。As shown in Figure 5, the dotted line represented by m is the seat cushion plane in the standby state, that is, the x0y plane, and the dotted line represented by n is the seat cushion plane in the use state. It can be seen that the z component of the n plane is smaller than the z component of the m plane. The weight means that when the user sits down, the seat cushion plane will sink. In Formula 1, the ideal state calculation is based on the seat cushion plane not sinking. Therefore, taking the sinking factor into consideration, the modified Formula 1 needs to be used. The depth of sinking d can be determined by looking up the table. The depth of sinking is not only related to the user's weight, but also to the material of the sofa. Therefore, manufacturers will pre-measure different types of sofas under different pressure values through experiments before leaving the factory. The sinking depth is used to modify formula 1, thereby improving the accuracy of dividing the response zone.
在一个实施例中,所述方法还包括:In one embodiment, the method further comprises:
当第一压力传感器阵列中压力值最大的压力传感器的压力值与其他所有被触发的压力传感器的差值均大于预设压力差值时,以压力最大的压力传感器的坐标点作为该用户坐姿下,坐垫平面上的第一压力传感器阵列中被用户触发的压力传感器组成的几何图形的重心点的坐标。When the difference between the pressure value of the pressure sensor with the largest pressure value in the first pressure sensor array and all other triggered pressure sensors is greater than the preset pressure difference value, the coordinate point of the pressure sensor with the largest pressure value is used as the user's sitting posture. , the coordinates of the center of gravity point of the geometric figure composed of the pressure sensors triggered by the user in the first pressure sensor array on the cushion plane.
可以理解的是,如图5所示,假设在第一姿态下,坐垫有4个压力传感器被触发,并且图5中S1表示其中一个压力传感器,其值最大并且与其他所有别触发的压力传感器的差值均大于预设压力差值,则直接以S1作为重心点。其原因是,当用户头部位于身体正中间时,由于臀部构造是两边突出中间凹陷,所以一般都是两瓣臀部共同承受差不多大小的体重,所以坐垫上的压力传感器一般会有两组传感器的读数是差不多的,此时头部一般位于两瓣臀部中间,需要根据几何图像重心来确定人体摆动的球心,即重心点坐标。而当人体倾斜身体使得头部偏向于某一侧时,就会导致人体的重量主要有某一侧臀部支撑,意味此时头部正好位于某一瓣臀部的正上方,所以可以直接将与其他所有别触发的压力传感器的差值均大于预设压力差值的压力传感器S1的坐标作为人体摆动的球心,即将S1直接作为重心点坐标带入到公式一中,获得更加精确的响应音区划分效果。It can be understood that, as shown in Figure 5, assuming that in the first posture, the cushion has 4 pressure sensors that are triggered, and S1 in Figure 5 represents one of the pressure sensors, which has the largest value and is triggered separately from all other pressure sensors. The differences are all greater than the preset pressure difference, then S1 is directly used as the center of gravity. The reason is that when the user's head is in the middle of the body, because the buttocks structure is protruding on both sides and concave in the middle, the two buttocks generally bear a similar amount of weight, so the pressure sensor on the seat cushion usually has two sets of sensors. The readings are almost the same. At this time, the head is generally located in the middle of the two buttocks. It is necessary to determine the spherical center of the human body swing based on the center of gravity of the geometric image, that is, the coordinates of the center of gravity point. When the human body tilts the body so that the head is tilted to a certain side, the weight of the human body will be mainly supported by a certain hip, which means that the head is just above a certain hip at this time, so it can be directly combined with other hips. The coordinates of the pressure sensor S1 whose difference values of all individually triggered pressure sensors are greater than the preset pressure difference value are used as the center of the sphere of human body swing. That is, S1 is directly brought into Formula 1 as the coordinates of the center of gravity point to obtain a more accurate response sound zone. Division effect.
实施例2Example 2
在本实施例中,用户接触靠背坐在沙发椅上的姿势定义为第二姿态,此时,坐垫平面上的第一压力传感器阵列和靠背上的第二压力传感器阵列二者同时被用户触发时的用户坐姿。本实施例中,初始坐姿为第二姿态。In this embodiment, the posture of the user touching the backrest and sitting on the sofa chair is defined as the second posture. At this time, when both the first pressure sensor array on the seat cushion plane and the second pressure sensor array on the backrest are triggered by the user at the same time user sitting posture. In this embodiment, the initial sitting posture is the second posture.
所述基于第一音区修正算法计算第一笔语音信号的声源坐标对应的空间范围,具体包括:The calculation of the spatial range corresponding to the sound source coordinates of the first speech signal based on the first sound zone correction algorithm specifically includes:
根据以下的公式二计算出第一笔语音信号的声源坐标对应的空间范围(x,y,z):The spatial range (x, y, z) corresponding to the sound source coordinates of the first speech signal is calculated according to the following formula 2:
其中,为初始坐姿状态下,第一笔语音信号的声源坐标,R为预设的颈部摆动长度,8cm≤R≤12cm,A,B,C是与靠背所在平面垂直的法向量的三个分量,D 是将靠背所在平面的坐标带入后求得的常数。in, is the sound source coordinate of the first voice signal in the initial sitting position, R is the preset neck swing length, 8cm≤R≤12cm, A, B, and C are the three components of the normal vector perpendicular to the plane of the backrest. , D is a constant obtained by bringing in the coordinates of the plane where the backrest is located.
不同于实施例1中,以臀部为人体摆动的中心点,通过公式一求出供人体在不改变用户坐姿情况下允许用户弯腰、低头活动的类圆锥体范围作为响应音区。本实施例中,当用户背包接触靠背并且背部不离开靠背的情况下,设置音区应该主要考虑头部(活动器官主要是颈部)的活动范围,而非实施例1中考虑的腰部的活动范围。Different from Embodiment 1, where the buttocks are used as the center point of the human body swing, the cone-like range of the human body that allows the user to bend over and lower the head without changing the user's sitting posture is calculated as the response sound zone through Formula 1. In this embodiment, when the user's backpack contacts the backrest and the back does not leave the backrest, setting the sound zone should mainly consider the range of movement of the head (the movable organ is mainly the neck), rather than the movement of the waist considered in Embodiment 1. scope.
因此,本实施例中,如公式二求出的响应音区的范围是以嘴巴为球心,以颈部摆动长度(人颈部一般长8-12cm)为半径,先求出球体,然后通过A,B,C是与靠背所在平面垂直的法向量的三个分量,限定了靠背平面的平面方程,同时限定只将球体与靠背正面(即朝向用户的一面)相交的位置作为响应音区,因为在靠背背面的范围不在沙发内,是需要排除的范围。可以理解的是,在不调节靠背倾斜角度情况下,靠背平面为yoz平面,法向量为(1,0,0),即A=1,B=0,C=0。并且将靠背平面上一个坐标带入该平面方程就可以求出D。Therefore, in this embodiment, the range of the response sound area calculated by Formula 2 is based on the mouth as the center of the sphere and the swing length of the neck (the human neck is generally 8-12cm long) as the radius. The sphere is first calculated, and then A, B, and C are the three components of the normal vector perpendicular to the plane of the backrest, which defines the plane equation of the backrest plane. At the same time, only the position where the sphere intersects with the front of the backrest (that is, the side facing the user) is used as the response sound zone. Because the area on the back of the backrest is not within the sofa, it is the area that needs to be excluded. It can be understood that without adjusting the backrest tilt angle, the backrest plane is the yoz plane and the normal vector is (1,0,0), that is, A=1, B=0, C=0. And by bringing a coordinate on the backrest plane into the equation of the plane, D can be obtained.
本实施例中,通过公式二设置的响应音区范围相较于初始范围缩小了范围,提升了语音信号质量,同时还可以在不改变用户坐姿情况下,留有给头部活动的余量,无需频繁修改响应音区范围,降低对计算机资源的开销。In this embodiment, the range of the response sound zone set by Formula 2 is narrowed compared to the initial range, which improves the quality of the voice signal. At the same time, it can also leave room for head movement without changing the user's sitting posture. There is no need to frequently modify the response range, reducing the cost of computer resources.
进一步的,所述方法还包括:Further, the method also includes:
检测到用户坐姿发生第二类改变时,基于第四音区修正算法计算最新用户坐姿对应的响应音区;其中,用户坐姿发生第二类改变是指,用户坐姿改变前后,靠背上的第二压力传感器阵列均被用户触发并且坐垫平面上的第一压力传感器阵列均被用户触发。When a second type of change in the user's sitting posture is detected, the response sound zone corresponding to the latest user's sitting posture is calculated based on the fourth sound zone correction algorithm; wherein, the second type of change in the user's sitting posture means that before and after the user's sitting posture changes, the second pressure sensor array on the backrest is triggered by the user and the first pressure sensor array on the seat cushion plane is triggered by the user.
所述基于第四音区修正算法计算最新用户坐姿对应的响应音区,具体包括:The calculating of the response sound zone corresponding to the latest user sitting posture based on the fourth sound zone correction algorithm specifically includes:
步骤S301,获取在用户坐姿改变前,靠背平面上的第二压力传感器阵列中被用户触发的压力传感器组成的几何图形的重心点P2的坐标。Step S301: Obtain the coordinates of the center point P2 of the geometric figure composed of the pressure sensors triggered by the user in the second pressure sensor array on the backrest plane before the user changes his sitting posture.
参照图7,在用户坐姿发生第二类改变之前,此时用户坐姿数据为M1=[000000011110],M2=[0000000001100110]。Referring to Figure 7, before the second type of change in the user's sitting posture occurs, the user's sitting posture data at this time is M1=[000000011110], M2=[0000000001100110].
步骤S302,获取在用户坐姿改变后,靠背平面上的第二压力传感器阵列中被用户触发的压力传感器组成的几何图形的重心点P3的坐标。Step S302, obtaining the coordinates of the center of gravity point P3 of a geometric figure composed of pressure sensors triggered by the user in the second pressure sensor array on the backrest plane after the user's sitting posture changes.
参照图8在用户坐在发生第二改变之后,此时用户坐姿数据为M1=[000110011000],M2=[0000011001100000]。Referring to Figure 8, after the second change in user sitting occurs, the user sitting posture data at this time is M1=[000110011000], M2=[0000011001100000].
步骤S303,用P3减去P2,得到平移偏移量。Step S303: Subtract P2 from P3 to obtain the translation offset.
步骤S304,将当前响应音区的所有坐标点加上平移偏移量,得到最新坐姿对应的响应音区的所有坐标点。Step S304: Add the translation offset to all the coordinate points of the current response sound area to obtain all the coordinate points of the response sound area corresponding to the latest sitting posture.
图7为用户坐姿发生第二类改变前的示意图,图8为用户坐姿发生第二类改变后的示意图,可知,用户坐姿发生第二类改变前后都维持第二姿态,但是用户的背部(背包会带动头部移动)在靠背平面上挪动了位置,通过用户坐姿数据可以判断出用户坐姿发生了改变。已知各传感器的坐标,根据求几何图形重心的方法可以容易求出P2和p3,此处不再赘述。Figure 7 is a schematic diagram before the second type of change in the user's sitting posture. Figure 8 is a schematic diagram after the second type of change in the user's sitting posture. It can be seen that the user maintains the second posture before and after the second type of change in the user's sitting posture, but the user's back (backpack) (will drive the head to move) has moved its position on the backrest plane. From the user's sitting posture data, it can be judged that the user's sitting posture has changed. Knowing the coordinates of each sensor, P2 and p3 can be easily obtained based on the method of finding the center of gravity of geometric figures, which will not be described again here.
可以理解的是,在步骤S301至步骤S304中,无需在用户坐姿发生第二类改变后重新使用公式二求出最新坐姿对应的响应音区的所有坐标点,而是直接利用平移偏移量与发生第二类改变前的响应音区(即步骤S304中的当前响应音区)进行简单加和运算即可,降低了更新响应音区的运算量,节省了计算机资源。It can be understood that in steps S301 to S304, there is no need to reuse Formula 2 after the second type of change in the user's sitting posture to find all the coordinate points of the response sound zone corresponding to the latest sitting posture, but directly use the translation offset and The response sound area before the second type of change occurs (that is, the current response sound area in step S304) can be simply added, which reduces the calculation amount of updating the response sound area and saves computer resources.
在一个实施例中,所述方法还包括:In one embodiment, the method further includes:
检测到用户坐姿发生第三类改变时,确定改变后最新的用户坐姿;第三类改变是指,用户坐姿改变前后,靠背上的第二压力传感器阵列分别处于被用户触发和不被用户触发的状态。即用户的坐姿在第一姿态和第二姿态直接切换的变化。When a third type of change in the user's sitting posture is detected, the latest user's sitting posture after the change is determined; the third type of change means that before and after the user's sitting posture changes, the second pressure sensor array on the backrest is in a state that is triggered by the user and is not triggered by the user respectively. state. That is, the user's sitting posture directly switches between the first posture and the second posture.
如果最新的用户坐姿为仅有坐垫平面上的第一压力传感器阵列被用户触发时的用户坐姿,则基于第二音区修正算法计算最新用户坐姿对应的响应音区;If the latest user sitting posture is the user's sitting posture when only the first pressure sensor array on the seat cushion plane is triggered by the user, then the response sound zone corresponding to the latest user sitting posture is calculated based on the second sound zone correction algorithm;
如果最新的用户坐姿为坐垫平面上的第一压力传感器阵列和靠背上的第二压力传感器阵列二者同时被用户触发时的用户坐姿,则基于第四音区修正算法计算最新用户坐姿对应的响应音区。If the latest user's sitting posture is the user's sitting posture when both the first pressure sensor array on the seat cushion plane and the second pressure sensor array on the backrest are triggered by the user at the same time, then the response corresponding to the latest user's sitting posture is calculated based on the fourth sound zone correction algorithm Sound area.
具体的,实施例3中,如图9所示,用户从第一姿态切换为第二姿态,即发生了第三类改变。具体的,可以根据用户坐姿数据来判断发生了第三类改变。参照图10,在用户坐姿发生第三类改变之前,此时用户坐姿数据为M1=[000000011110],M2=[0000000000000000]。参照图11在用户坐在发生第三改变之后,此时用户坐姿数据为M1=[000000011110],M2=[0000000001100110]。Specifically, in Embodiment 3, as shown in Figure 9, the user switches from the first posture to the second posture, that is, the third type of change occurs. Specifically, it can be judged that the third type of change has occurred based on the user's sitting posture data. Referring to Figure 10, before the third type of change in the user's sitting posture occurs, the user's sitting posture data at this time is M1=[000000011110], M2=[0000000000000000]. Referring to Figure 11, after the third change in user sitting occurs, the user sitting posture data at this time is M1=[000000011110], M2=[0000000001100110].
根据实施例1和2可知,已经分别得到该用户在第一姿态或者第二姿态下的响应音区范围了,所以可以直接使用步骤S201至S203(即第二音区修正算法)或步骤S301至S304来计算最新用户坐姿对应的响应音区,以节省计算机开销。According to Embodiments 1 and 2, the user's response range range in the first posture or the second posture has been obtained respectively, so steps S201 to S203 (i.e., the second range correction algorithm) or steps S301 to S301 can be directly used. S304 to calculate the response sound area corresponding to the latest user's sitting posture to save computer overhead.
具体的,实施例4中,如图12所示,用户从第二姿态切换为第一姿态,即发生了第三类改变。具体的,可以根据用户坐姿数据来判断发生了第三类改变。参照图13,在用户坐姿发生第三类改变之前,此时用户坐姿数据为M1=[000110011000],M2=[0000011001100000]。参照图14在用户坐在发生第三改变之后,此时用户坐姿数据为M1=[000110011000],M2=[0000000000000000]。Specifically, in Embodiment 4, as shown in Figure 12, the user switches from the second posture to the first posture, that is, the third type of change occurs. Specifically, it can be judged that the third type of change has occurred based on the user's sitting posture data. Referring to Figure 13, before the third type of change in the user's sitting posture occurs, the user's sitting posture data at this time is M1=[000110011000], M2=[0000011001100000]. Referring to Figure 14, after the third change in user sitting occurs, the user sitting posture data at this time is M1=[000110011000], M2=[0000000000000000].
根据实施例1和2可知,已经分别得到该用户在第一姿态或者第二姿态下的响应音区范围了,所以可以直接使用步骤S201至S203(即第二音区修正算法)或步骤S301至S304来计算最新用户坐姿对应的响应音区,以节省计算机开销。According to Examples 1 and 2, the response sound zone range of the user in the first posture or the second posture has been obtained respectively, so steps S201 to S203 (i.e., the second sound zone correction algorithm) or steps S301 to S304 can be directly used to calculate the response sound zone corresponding to the latest user sitting posture to save computer overhead.
实施例5Example 5
如图15所示,在本实施例中,所述方法还包括:As shown in Figure 15, in this embodiment, the method further includes:
检测到用户坐姿发生第四类改变时,基于第五音区修正算法计算最新用户坐姿对应的响应音区;其中,用户坐姿发生第四类改变是指,在靠背上的第二压力传感器阵列被用户触发的情况下,用户坐姿改变前后,靠背的倾斜角度发生变动;When it is detected that the fourth type of change in the user's sitting posture occurs, the response sound zone corresponding to the latest user's sitting posture is calculated based on the fifth sound zone correction algorithm; wherein the fourth type of change in the user's sitting posture means that the second pressure sensor array on the backrest is When triggered by the user, the tilt angle of the backrest changes before and after the user changes his sitting posture;
所述基于第五音区修正算法计算最新用户坐姿对应的响应音区,具体包括:The step of calculating the response sound zone corresponding to the latest user sitting posture based on the fifth sound zone correction algorithm specifically includes:
基于靠背倾斜角度的变动,确定变动前后靠背所在平面的旋转矩阵;Based on the change in the inclination angle of the backrest, determine the rotation matrix of the plane where the backrest is located before and after the change;
基于所述旋转矩阵将当前响应音区的所有坐标点转换得到最新坐姿对应的响应音区的所有坐标点;Based on the rotation matrix, all coordinate points of the current response sound zone are converted to obtain all coordinate points of the response sound zone corresponding to the latest sitting posture;
其中,所述旋转矩阵为: Wherein, the rotation matrix is:
α为靠背倾斜角度的变动值,0°≤α≤60°。α is the change value of the backrest tilt angle, 0°≤α≤60°.
可以理解的是,用户躺在沙发靠背之后,不再挪动背包或者臀部的情况下,随着靠背倾斜角度的变动,用户的嘴巴位置也会发生改变,因此在靠背倾斜角度发生变动时,也需要修正响应音区的位置。It is understandable that after the user lies on the back of the sofa and no longer moves the backpack or buttocks, the position of the user's mouth will also change as the inclination angle of the backrest changes. Therefore, when the inclination angle of the backrest changes, it is also necessary to Correct the position of the response zone.
具体的,本实施例中,靠背绕在y轴转动了α角,此时可以采用平面绕y轴旋转的旋转矩阵来计算倾斜角度变动前后的映射关系,加快对于第四类变动的音区修正过程。Specifically, in this embodiment, the backrest rotates around the y-axis by an angle α. At this time, the rotation matrix of the plane rotating around the y-axis can be used to calculate the mapping relationship before and after the tilt angle change, speeding up the sound zone correction for the fourth type of change. process.
举例而言,用户背靠在靠背上属于第二姿态,已经在实施例2中得到了对应的响应音区(即靠背平面为调教倾角时的响应音区),如果靠背转动了α角,则可以使用旋转矩阵直接对实施例2中得到的响应音区的所有坐标进行变换,就可以转换出调节倾角后的响应音区。例如,α=30度,点P(1, 2, 3)绕Y轴旋转30度后的新坐标大约是(2.37, 2, 2.10)。For example, the user leaning on the backrest belongs to the second posture, and the corresponding response sound zone (i.e., the response sound zone when the backrest plane is the adjustment angle) has been obtained in Example 2. If the backrest is rotated by an angle α, the rotation matrix can be used to directly transform all the coordinates of the response sound zone obtained in Example 2, and the response sound zone after the adjustment angle can be converted. For example, α=30 degrees, the new coordinates of point P(1, 2, 3) after rotating 30 degrees around the Y axis are approximately (2.37, 2, 2.10).
综上,如图17所示,用户坐姿变化分为四种类型,每种类型的变换会采用相应的音区修正算法来动态修正响应音区的范围,即上述的5中实施例,以获得高质量的语音信号,从而提高语音识别准确性。In summary, as shown in FIG17 , user sitting posture changes are divided into four types. Each type of change will adopt a corresponding sound zone correction algorithm to dynamically correct the range of the response sound zone, that is, the above-mentioned 5 embodiments, to obtain a high-quality voice signal, thereby improving the accuracy of voice recognition.
另一方面,本发明提供了一种沙发椅控制器,包括存储器、处理器及存储在存储器上的计算机程序,所述处理器执行所述计算机程序以实现上述方法的步骤。On the other hand, the present invention provides a sofa chair controller, which includes a memory, a processor and a computer program stored on the memory. The processor executes the computer program to implement the steps of the above method.
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程,是可以通过计算机程序来指令相关的硬件来完成,所述的程序可存储于一非易失性计算机可读取存储介质中,该程序在执行时,可包括如上述各方法的实施例的流程。其中,本申请所提供的各实施例中所使用的对存储器、存储、数据库或其它介质的任何引用,均可包括非易失性和/或易失性存储器。非易失性存储器可包括只读存储器(ROM)、可编程ROM(PROM)、电可编程ROM(EPROM)、电可擦除可编程ROM(EEPROM)或闪存。易失性存储器可包括随机存取存储器(RAM)或者外部高速缓冲存储器。作为说明而非局限,RAM以多种形式可得,诸如静态RAM(SRAM)、动态RAM(DRAM)、同步DRAM(SDRAM)、双数据率SDRAM(DDRSDRAM)、增强型SDRAM(ESDRAM)、同步链路(Synchlink)DRAM(SLDRA)、存储器总线(Rambus)直接RAM(RDRA)、直接存储器总线动态RAM(DRDRAM)、以及存储器总线动态RAM(RDRAM)等。Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing relevant hardware through computer programs. The programs can be stored in a non-volatile computer-readable storage medium. , when the program is executed, it may include the processes of the above-mentioned method embodiments. Any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and/or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Synchlink DRAM (SLDRA), memory bus (Rambus) direct RAM (RDRA), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
以上实施例的各技术特征可以进行任意的组合,为使描述简洁,未对上述实施例中的各个技术特征所有可能的组合都进行描述,然而,只要这些技术特征的组合不存在矛盾,都应当认为是本说明书记载的范围。The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
Claims (10)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410121082.9A CN117809637A (en) | 2024-01-29 | 2024-01-29 | Control device and control method for intelligent desktop |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410121082.9A CN117809637A (en) | 2024-01-29 | 2024-01-29 | Control device and control method for intelligent desktop |
Publications (1)
Publication Number | Publication Date |
---|---|
CN117809637A true CN117809637A (en) | 2024-04-02 |
Family
ID=90428358
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202410121082.9A Pending CN117809637A (en) | 2024-01-29 | 2024-01-29 | Control device and control method for intelligent desktop |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117809637A (en) |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN201965906U (en) * | 2010-04-01 | 2011-09-07 | 北京邮电大学 | Speech signal processing device for digital watermark technology |
CN102800312A (en) * | 2011-05-24 | 2012-11-28 | 鸿富锦精密工业(深圳)有限公司 | Voice control system and method |
CN102881287A (en) * | 2012-09-20 | 2013-01-16 | 熊猫电子集团有限公司 | Voice triggering method and circuit |
US20140257813A1 (en) * | 2013-03-08 | 2014-09-11 | Analog Devices A/S | Microphone circuit assembly and system with speech recognition |
CN104301650A (en) * | 2013-07-16 | 2015-01-21 | 鸿富锦精密工业(深圳)有限公司 | signal transmission system |
CN106782519A (en) * | 2016-12-23 | 2017-05-31 | 深圳先进技术研究院 | A kind of robot |
CN108491285A (en) * | 2018-02-23 | 2018-09-04 | 广东欧珀移动通信有限公司 | Data guard method and Related product applied to testing for electrical equipment |
CN212064348U (en) * | 2020-06-04 | 2020-12-01 | 河南莱拓智能科技有限公司 | Desk type microphone with DSP noise reduction function |
-
2024
- 2024-01-29 CN CN202410121082.9A patent/CN117809637A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN201965906U (en) * | 2010-04-01 | 2011-09-07 | 北京邮电大学 | Speech signal processing device for digital watermark technology |
CN102800312A (en) * | 2011-05-24 | 2012-11-28 | 鸿富锦精密工业(深圳)有限公司 | Voice control system and method |
CN102881287A (en) * | 2012-09-20 | 2013-01-16 | 熊猫电子集团有限公司 | Voice triggering method and circuit |
US20140257813A1 (en) * | 2013-03-08 | 2014-09-11 | Analog Devices A/S | Microphone circuit assembly and system with speech recognition |
CN104301650A (en) * | 2013-07-16 | 2015-01-21 | 鸿富锦精密工业(深圳)有限公司 | signal transmission system |
CN106782519A (en) * | 2016-12-23 | 2017-05-31 | 深圳先进技术研究院 | A kind of robot |
CN108491285A (en) * | 2018-02-23 | 2018-09-04 | 广东欧珀移动通信有限公司 | Data guard method and Related product applied to testing for electrical equipment |
CN212064348U (en) * | 2020-06-04 | 2020-12-01 | 河南莱拓智能科技有限公司 | Desk type microphone with DSP noise reduction function |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111868824B (en) | Device and method for context-aware control | |
US10638842B2 (en) | Smart seat and control method for the same | |
CN110678921B (en) | Dynamic text-to-speech responses from smart speakers | |
JP7578219B2 (en) | Managing the playback of multiple audio streams through multiple speakers | |
US12069436B2 (en) | Ear-worn electronic device employing acoustic environment adaptation for muffled speech | |
Wolfe et al. | Benefit of a commercially available cochlear implant processor with dual-microphone beamforming: a multi-center study | |
US11381417B2 (en) | User terminal and sleep management method | |
US11546688B2 (en) | Loudspeaker device, method, apparatus and device for adjusting sound effect thereof, and medium | |
CN106688247A (en) | Determination of room reverberation for signal enhancement | |
CN106257936B (en) | In-situ fitting systems for hearing aids and hearing aid systems | |
WO2022127477A1 (en) | Method and apparatus for controlling smart playing terminal, and storage medium and smart wearable device | |
CN108803896A (en) | Control method, apparatus, terminal and the storage medium of screen | |
CN113012574B (en) | Screen curvature adjustment method, device, curved display and storage medium | |
CN114442503B (en) | Smart home linkage management method and system | |
CN114625019A (en) | A method and device for controlling smart home equipment based on the Internet of Things | |
CN117809637A (en) | Control device and control method for intelligent desktop | |
US20220279290A1 (en) | Ear-worn electronic device employing user-initiated acoustic environment adaptation | |
US20210280188A1 (en) | Information processing method, information processing system, and non-transitory computer-readable recording medium recording information processing program | |
CN117636858B (en) | Intelligent furniture controller and control method | |
JP6767017B1 (en) | Portable intelligent noise suppression and audio signal broadcasting device | |
CN109819112A (en) | A kind of control method of alarm clock, device, storage medium and electronic equipment | |
CN115076877A (en) | Method and device for controlling air conditioner, air conditioner and storage medium | |
JP2023153610A (en) | air conditioning system | |
CN108834231A (en) | Ground pad control method and smart machine | |
US20240335163A1 (en) | Electronic apparatus and method for controlling thereof |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |