CN111664834A - Method/system for estimating elevation position of indoor moving body, storage medium, and apparatus - Google Patents

Abstract

The invention discloses an elevation estimation method/system, storage medium and equipment for an indoor moving body. The collected state data is used to judge and identify the state of the detected moving body, and the elevation position of the detected moving body is estimated according to the judgment and identification result, which is beneficial to Improve the estimation accuracy; and when the measured moving body is identified and judged to be non-planar motion, the various data in the state data are combined with the extended Kalman algorithm to estimate the elevation position of the measured moving body, which can further improve the indoor measured Elevation position estimation accuracy for moving objects.

Description

Elevation position estimation method/system, storage medium and device for indoor moving objects

technical field

The present invention relates to a method for estimating an elevation position, in particular to a method for estimating an elevation position of an indoor mobile unit.

Background technique

In the field of indoor positioning, the height position information in the vertical direction is an important means to achieve three-dimensional spatial positioning. At present, the estimation methods of indoor height mainly include barometric altimetry and inertial integration, both of which are estimated by collecting data from sensors of indoor mobile units (such as personnel). Among them, the barometric altimetry method uses the relationship between the air pressure change and the height to estimate the height displacement of the mobile unit to be measured in motion. Although this method is simple in technology and high in operability, the relationship between air pressure and height is very complicated. In addition, the air pressure will be affected by environmental factors, and the estimated elevation position has a large error; while the inertial integration method calculates the vertical height change by using the component of the acceleration in the vertical direction. Although the estimation accuracy of the method has been improved, the accumulated error is large after long-term use, which will lead to a significant decrease in the estimated elevation position accuracy with the increase of running time.

SUMMARY OF THE INVENTION

In view of the above-mentioned shortcomings of the prior art, the main purpose of the present invention is to provide a method for estimating the elevation position of an indoor person, so as to improve the estimation accuracy of the elevation position of an indoor moving body.

For achieving the above-mentioned purpose and other related purposes, the technical scheme of the present invention is as follows:

A method for calculating the elevation position of an indoor moving body, comprising:

Collect real-time state data of the moving object under test at a specified frequency, the state data at least include air pressure data and acceleration data in the vertical direction;

dividing the acceleration component data in the vertical direction of the acceleration data and the remaining data in the state data according to time series to form a first sub-data set of several specified durations;

Input the first sub-data set of the moving object under test within the current specified time period into a pre-generated motion state recognition classifier, and determine the identification of the moving object under test according to the first sub-data set. Which state of the motion state is the flat motion state and the non-flat motion state?

If the moving object to be tested is in the flat ground motion state, the current output height corresponding to the current specified duration is equal to the previous output height corresponding to the previous specified duration,

If the measured moving body is in the non-flat motion state, the current output height is estimated by fusing all the first sub-data through an extended Kalman filter method.

Optionally, the method for identifying the motion state of the moving body under test by the motion state recognition classifier includes:

extracting categorical features of the first sub-data set,

The motion state of the tested moving body is identified according to the classification feature of each category.

Optionally, the training method of the motion state recognition classifier includes:

The state data of the moving object under test in the flat ground motion state and the non-flat ground motion state are collected at the specified frequency as training data,

Dividing the acceleration data in the training data, the acceleration component data of the vertical component and the rest of the data in the training data, evenly in time series to form a plurality of second sub-data sets of the specified duration;

extracting the classification features of each of the second sub-data sets to form training samples;

The training samples are input into a support vector machine model for training, the parameters of the support vector machine model are optimized, and the parameter-optimized support vector machine model forms the motion state recognition classifier.

Optionally, the classification features include the mean value of acceleration in the vertical direction, the variance of the acceleration in the vertical direction, the difference in air pressure, and the variance in air pressure.

Optionally, the state of the flat ground movement includes stillness, walking and jogging, and the non-flat ground movement state includes upstairs and downstairs.

Optionally, the method for estimating the current output height value by fusing the first sub-data set with the extended Kalman filter method includes:

A system state equation is constructed by using the acceleration component data in the vertical direction, and the current prior height is calculated;

Construct a priori noise covariance matrix calculation formula according to the acceleration component data in the vertical direction and the previous a posteriori noise covariance matrix, and calculate the current priori noise covariance matrix;

Construct a system measurement equation through the air pressure data to calculate the current measurement altitude;

Build a Kalman gain calculation formula according to the prior noise covariance matrix, and calculate the current Kalman gain;

Construct a current output height fusion calculation equation according to the current prior height, the current measured height and the current Kalman gain, and calculate the current output height;

The current a priori noise covariance matrix is obtained by updating the current prior noise covariance matrix according to the current Kalman gain.

Optionally, the system state equation is:

是当前先验高度，

是前一所述指定时长对应的所述前一输出高度，

是相对高度计算函数，所述高度计算函数中的变量sa_i＝[a₁,a₂,…,a_n]^T，a₁,a₂,…,a_n是指当前所述指定时长内所述第一子数据集中按时序排列的所述竖直方向上的加速度数据，表n示当前所述指定时长内采集所述竖直方向上的加速度数据的总次数，a₁表示当前所述指定时长内的第一个所述竖直方向上的加速度数据；a_n表示当前所述指定时长内的最后一个所述竖直方向上的加速度数据；Among them, i refers to the current specified duration, i-1 refers to the previous specified duration,

is the current prior height,

is the previous output height corresponding to the previous specified duration,

is the relative height calculation function, the variables sa _i =[a ₁ ,a ₂ ,...,a _n ] ^T in the height calculation function, a ₁ ,a ₂ ,...,a _n refers to the The acceleration data in the vertical direction arranged in time series in the first sub-data set, n represents the total number of times the acceleration data in the vertical direction is collected within the current specified time period, and a ₁ represents the current specified time period. the first acceleration data in the vertical direction within the duration; _an represents the last acceleration data in the vertical direction within the current specified duration;

The calculation formula of the prior noise covariance matrix is:

是当前所述指定时长对应的当前先验噪声协方差矩阵，

是前一所述指定时长对应的后验噪声协方差矩阵，

是系统状态方程中所述相对高度计算函数

的雅可比矩阵，

是

的转置矩阵，Q是过程噪声协方差；in,

is the current prior noise covariance matrix corresponding to the current specified duration,

is the posterior noise covariance matrix corresponding to the specified duration described in the previous section,

is the relative height calculation function described in the system state equation

The Jacobian matrix of ,

Yes

The transpose matrix of , Q is the process noise covariance;

The system measurement equation is:

表示当前测量高度，P_i表示当前气压值，P₁表示初始气压值，P₀表示标准大气压；in,

represents the current measurement altitude, P _i represents the current air pressure value, P ₁ represents the initial air pressure value, and P ₀ represents the standard atmospheric pressure;

The current Kalman gain calculation formula is:

和

分别为加速度和气压计的噪声方差；where K _i represents the current Kalman gain, R _Q represents the measurement variance,

and

are the noise variances of the accelerometer and barometer, respectively;

The current output height fusion calculation equation is:

The formula for updating the current prior noise covariance matrix to obtain the current posterior noise covariance matrix according to the current Kalman gain is:

An elevation position estimation system for an indoor moving body, comprising:

a state data acquisition module, which is used for real-time acquisition of the state data of the moving object under test, the state data at least including air pressure data and acceleration data in the vertical direction;

a data preprocessing module, configured to equally divide the acceleration component data in the vertical direction of the acceleration data and the remaining data in the state data according to time series to form a first sub-data set of several specified durations;

The motion state recognition classifier is used to read the first sub-data set of the tested moving body within the current specified time period, and determine the identification of the tested moving body according to the first sub-data set. Which state of the motion state is in the flat ground motion state and the non-flat ground motion state;

An output height calculation module, which is used to calculate the current output height corresponding to the current specified duration according to the recognition result of the motion state recognition classifier,

If the measured moving body is in the flat ground motion state, the output height calculation module outputs the previous output height corresponding to the previous specified duration as the current output height;

If the moving object under test is in the non-flat motion state, the output height calculation module estimates the current output height by fusing all the first sub-data through an extended Kalman filter method.

Optionally, the motion state recognition classifier includes:

a feature processing unit, the feature processing unit is used to extract the classification features of the first sub-data set;

A state determination unit, which is used for identifying the motion state of the moving object under test according to the classification feature of each type.

Optionally, the classification features include the mean value of acceleration in the vertical direction, the variance of the acceleration in the vertical direction, the difference in air pressure, and the variance in air pressure.

Optionally, the state of the flat ground movement includes stillness, walking and jogging, and the non-flat ground movement state includes upstairs and downstairs.

Optionally, the output height calculation module includes:

A current prior height calculation unit, configured to calculate the current prior height according to the acceleration component data in the vertical direction;

A current priori noise covariance matrix calculation unit, configured to calculate the current priori noise covariance matrix according to the acceleration component data in the vertical direction and the previous a posteriori noise covariance matrix;

The current measurement altitude calculation unit, which is used to calculate the current measurement altitude according to the air pressure data;

A current Kalman gain calculation unit, configured to calculate the current Kalman gain according to the current noise covariance;

A current output height fusion calculation unit, configured to fuse and calculate the current output height according to the current prior height, the current measured height and the current Kalman gain;

A posteriori noise covariance updating unit, configured to update the current priori noise covariance matrix according to the current Kalman gain to obtain a current posteriori noise covariance matrix.

Optionally, the state data acquisition module includes an accelerometer and a barometer.

A storage medium having a computer program stored thereon, when the program is executed by a processor, implements any one of the above-mentioned methods for estimating the elevation position of an indoor moving body.

A device, comprising a processor and a memory, the memory is used to store a computer program, the processor is used to execute the computer program stored in the memory, so that the device executes any one of the above-mentioned indoor mobile objects. Elevation location estimation method.

The method/system, storage medium and device for estimating the elevation of an indoor moving body of the present invention use the collected state data to judge and identify the state of the moving body under test, and estimate the elevation position of the moving body under test according to the judgment and identification result, which is beneficial to improve Estimation accuracy; and when the measured moving body is identified and judged to be non-planar motion, various data in the state data are combined with the extended Kalman algorithm to estimate the elevation position of the measured moving body, which can further improve the indoor measured movement. Elevation position estimation accuracy of the volume.

Description of drawings

FIG. 1 is a flow chart of the method for estimating the elevation position of an indoor moving body according to the present invention;

Fig. 2 shows the flow chart of estimating the current output height value by the extended Kalman algorithm fusion in the present invention;

Fig. 3 shows the network structure diagram of the elevation position estimation system of the indoor moving body of the present invention;

FIG. 4 is a network structure diagram of the output height calculation module in the present invention.

Detailed ways

In order to make the objectives, technical solutions and advantages of the present invention clearer, the specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention.

In addition, it should be noted that, for the convenience of description, the drawings only show some but not all of the contents related to the present invention. Before discussing the exemplary embodiments in greater detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although a flowchart depicts various operations (or steps) as a sequential process, many of the operations may be performed in parallel, concurrently, or concurrently. Additionally, the order of operations can be rearranged. The process may be terminated when its operation is complete, but may also have additional steps not included in the figures. The processes may correspond to methods, functions, procedures, subroutines, subroutines, and the like.

Referring to Fig. 1, a method for calculating the elevation position of an indoor moving body provided by the present invention includes:

Step 120, collect the state data of the moving object under test in real time at a specified frequency, and the state data at least include air pressure data and acceleration data in the vertical direction;

Step 140, data preprocessing: the acceleration component data of the acceleration data in the vertical direction and the rest of the data in the state data are evenly divided according to time series to form a first sub-data set of several specified durations;

Step 160: Input the first sub-data set of the moving object under test in the current specified time period into a pre-generated motion state recognition classifier, and determine and identify the tested moving body according to the first sub-data set. Which state of the motion state of the moving body is in the flat motion state and the non-flat motion state?

Step 180B, if the moving object to be tested is in the flat ground motion state, the current output height corresponding to the current specified duration is equal to the previous output height corresponding to the previous specified duration,

Step 180A, if the moving object under test is in the non-flat motion state, the current output height is estimated by fusing all the first sub-data through an extended Kalman filter method.

In the actual implementation process, the moving object to be measured may be a human, an animal, or even a robot. In this embodiment, the moving object to be measured is an indoor person. When the moving object to be measured is a human or an animal, the flat ground The motion states include stationary, walking, and jogging, and the non-level motion states include upstairs and downstairs. When performing the above or the following estimation methods, the moving object under test needs to be equipped with a data acquisition module for acquiring the above state data. For example, the data acquisition module may include a barometer for acquiring air pressure and an accelerometer for acquiring vertical acceleration , the acceleration component data in the vertical direction can also be obtained by decomposing the total acceleration. When a sensor is used to collect state data, the sensor can be embedded in a wearable device or a mobile terminal carried by indoor personnel. The position of the sensor can be any part of the moving body to be tested. For example, when the moving body to be tested is an indoor person, a wearable device embedded with the sensor can be worn on the ankle, wrist, waist, thigh, arm or indoor. any other part of the person's body.

In this embodiment, the specified duration can be selected as any length, and the specified frequency can be selected as any frequency, but the shorter the specified duration is, the better the estimation accuracy of the elevation position is improved, but the shorter the specified duration, the greater the amount of data processing The larger the value, the higher the specified frequency, the more conducive to improving the estimation accuracy of the elevation position, but the higher the specified flat rate, which will also lead to an increase in the amount of data processing. For example, the specified duration can be selected as 3 seconds, and the specified frequency can be selected as 100Hz, then in the current 3 seconds, the number of times of collecting the state data is 300 times, that is, there are 300 air pressure data and 300 acceleration data.

In this embodiment, for ease of understanding, the specified duration is defined as t, the number of collection times is defined as n, and i is defined as the current moment. The acceleration data sequence and the air pressure data sequence within the specified time period, the acceleration data sequence in the vertical direction in the first sub-data set can be expressed as (a ₁ , a ₂ ,...,an ), and the air pressure data sequence can _be It is expressed as (p ₁ , p ₂ ,..., _{p n} ), where a ₁ represents the acceleration component data of the first vertical direction collected in the specified time sequence in time series, and an represents the time sequence in the specified time period For the last collected acceleration component data in the vertical direction, p ₁ represents the first air pressure data collected in time sequence within the specified time period, and p _n represents the last air pressure data collected in time sequence within the specified time period.

The method for estimating the elevation of an indoor moving body of the present invention uses the collected state data to judge and identify the state of the moving body under test, and estimates the elevation position of the moving body under test according to the judgment and identification result, which is beneficial to improve the estimation accuracy; When the object is identified and judged to be non-planar motion, various data in the state data are combined with the extended Kalman algorithm to estimate the elevation position of the moving object under test, which can further improve the estimation accuracy of the elevation position of the indoor moving object under test.

In some embodiments, the method for identifying the motion state of the detected moving body by the motion state identification classifier includes:

extracting categorical features of the first sub-data set,

The motion state of the tested moving body is identified according to the classification feature of each category.

In some embodiments, the classification features include vertical acceleration mean, vertical acceleration variance, air pressure difference, and air pressure variance.

The formula for calculating the mean value of the vertical acceleration is:

为被测移动体在指定时间段t内的竖直方向加速度均值；in,

is the mean value of the vertical acceleration of the moving object under test within the specified time period t;

The formula for calculating the vertical acceleration variance is:

为被测移动体的在指定时间段t内的竖直方向加速度方差；in,

is the vertical acceleration variance of the moving object under test within the specified time period t;

The formula for calculating the air pressure difference is:

Δp=p ₁ −p ₂

Among them, Δp represents the air pressure difference of the moving object under test within the specified time period t;

The formula for calculating the air pressure variance is:

表示被测移动体的在指定时间段t内的气压方差，

表示被测移动体的在指定时间段t内的平均气压，

in,

represents the air pressure variance of the moving object under test within the specified time period t,

Represents the average air pressure of the moving object under test within a specified time period t,

In some embodiments, the training method of the motion state recognition classifier includes:

The state data of the moving object under test in the flat ground motion state and the non-flat ground motion state are collected at the specified frequency as training data,

The acceleration data in the training data is evenly divided in time series between the acceleration component data of the vertical component and the rest of the data in the training data to form a number of second sub-data sets of the specified duration;

extracting the classification features of each of the second sub-data sets to form training samples;

The training samples are input into a support vector machine model for training, the parameters of the support vector machine model are optimized, and the parameter-optimized support vector machine model forms the motion state recognition classifier.

In some embodiments, the training sample is a feature matrix X formed by the arrangement of the classification features, and a class label vector Y corresponding to each classification feature is defined, then when optimizing the support vector machine model, the feature matrix X and The class label vector Y is input into the support vector machine model, and the optimal parameters are solved. The solving equations of the optimal parameters are as follows:

Among them, m represents the number of training samples, x _v and x _w represent the v-th sample and the w-th sample, respectively, y _v , y _w represent the labels corresponding to the samples x _v and x _w respectively, β _v and β _w are A vector of parameters to be estimated.

Referring to FIG. 2 , in some embodiments, the method for estimating the current output height value by fusing the first sub-data set with an extended Kalman filter method includes:

Step 181, constructing a system state equation by using the acceleration component data in the vertical direction, and calculating the current prior height;

Step 182: Construct a priori noise covariance matrix calculation formula according to the acceleration component data in the vertical direction and the previous a posteriori noise covariance matrix, and calculate the current priori noise covariance matrix;

Step 183, constructing a system measurement equation based on the air pressure data, and calculating the current measurement altitude;

Step 184, construct a Kalman gain calculation formula according to the prior noise covariance matrix, and calculate the current Kalman gain;

Step 185, construct a current output height fusion calculation equation according to the current prior height, the current measured height and the current Kalman gain, and calculate the current output height;

Step 186: Update the current prior noise covariance matrix according to the current Kalman gain to obtain a current posterior noise covariance matrix.

In some embodiments, the system state equation is:

是当前先验高度，

是前一所述指定时长对应的所述前一输出高度，

是相对高度计算函数，所述高度计算函数中的变量sa_i＝[a₁,a₂,…,a_n]^T，a₁,a₂,…,a_n是指当前所述指定时长内所述第一子数据集中按时序排列的所述竖直方向上的加速度数据，表n示当前所述指定时长内采集所述竖直方向上的加速度数据的总次数，a₁表示当前所述指定时长内的第一个所述竖直方向上的加速度数据；a_n表示当前所述指定时长内的最后一个所述竖直方向上的加速度数据；Among them, i refers to the current specified duration, i-1 refers to the previous specified duration,

is the current prior height,

is the previous output height corresponding to the previous specified duration,

is the relative height calculation function, the variables sa _i =[a ₁ ,a ₂ ,...,a _n ] ^T in the height calculation function, a ₁ ,a ₂ ,...,a _n refers to the The acceleration data in the vertical direction arranged in time series in the first sub-data set, n represents the total number of times the acceleration data in the vertical direction is collected within the current specified time period, and a ₁ represents the current specified time period. the first acceleration data in the vertical direction within the duration; _an represents the last acceleration data in the vertical direction within the current specified duration;

The calculation formula of the prior noise covariance matrix is:

是当前所述指定时长对应的当前先验噪声协方差矩阵，

是前一所述指定时长对应的后验噪声协方差矩阵，

是系统状态方程中所述相对高度计算函数

的雅可比矩阵，

是

的转置矩阵，Q是过程噪声协方差，Q可以通过观测获得；in,

is the current prior noise covariance matrix corresponding to the current specified duration,

is the posterior noise covariance matrix corresponding to the specified duration described in the previous section,

is the relative height calculation function described in the system state equation

The Jacobian matrix of ,

Yes

The transpose matrix of , Q is the process noise covariance, and Q can be obtained by observation;

The system measurement equation is:

表示当前测量高度，P_i表示当前气压值，P_i等于当前第一子数据集中的p_n，P₁表示初始气压值，P₀表示标准大气压；in,

represents the current measurement altitude, _Pi represents the current air pressure value, Pi is equal to _pn in the current first sub-data set, P ₁ represents the initial air pressure value, and P ₀ represents the standard atmospheric pressure _;

The current Kalman gain calculation formula is:

和

分别为加速度和气压计的噪声方差；where K _i represents the current Kalman gain, R _Q represents the measurement variance,

and

are the noise variances of the accelerometer and barometer, respectively;

The current output height fusion calculation equation is:

The formula for updating the current prior noise covariance matrix to obtain the current posterior noise covariance matrix according to the current Kalman gain is:

Using the above method, the air pressure data and acceleration data can be combined with the extended Kalman gain algorithm to estimate the fusion height with high accuracy.

This embodiment also discloses a storage medium on which a computer program corresponding to the above and the following methods is stored, and when the program is executed by a processor, implements any of the above methods for estimating the elevation position of an indoor moving body.

For the storage medium in this embodiment, those of ordinary skill in the art can understand that all or part of the steps of implementing each method embodiment of this specification can be completed by hardware related to a computer program. The aforementioned computer program may be stored in a computer-readable storage medium. When the program is executed, it executes the steps including the method embodiments in this specification; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other media that can store program codes.

Referring to FIG. 3 , corresponding to the above-mentioned algorithm for estimating the elevation position of an indoor moving body, the present embodiment also provides an elevation position estimation system for an indoor moving body, and the system includes:

a state data acquisition module 2, which is used for real-time acquisition of the state data of the moving object under test, the state data at least including air pressure data and acceleration data in the vertical direction;

a data preprocessing module 4, configured to equally divide the acceleration component data in the vertical direction of the acceleration data and the remaining data in the state data according to time series to form a first sub-data set of several specified durations;

A motion state recognition classifier 6, configured to read the first sub-data set of the tested moving body within the current specified time period, and determine and identify the tested moving body according to the first sub-data set Which state of the motion state is in the flat ground motion state and the non-flat motion state;

Output height calculation module 8, which is used to calculate the current output height corresponding to the current specified duration according to the recognition result of the motion state recognition classifier,

If the measured moving body is in the flat ground motion state, the output height calculation module outputs the previous output height corresponding to the previous specified duration as the current output height;

If the moving object under test is in the non-flat motion state, the output height calculation module estimates the current output height by fusing all the first sub-data through an extended Kalman filter method.

In some embodiments, the motion state recognition classifier may include:

a feature processing unit, the feature processing unit is used to extract the classification features of the first sub-data set;

A state determination unit, which is used for identifying the motion state of the moving object under test according to the classification feature of each type.

In other embodiments, the feature processing unit may also be arranged in the data preprocessing module, so that the formed first sub-data set always includes classification features.

In some embodiments, the classification features include vertical acceleration mean, vertical acceleration variance, air pressure difference, and air pressure variance.

In some embodiments, the flat motion states include stationary, walking, and jogging, and the non-flat motion states include upstairs and downstairs.

Referring to FIG. 4, in some embodiments, the output height calculation module includes:

The current prior height calculation unit 81 is configured to calculate the current prior height according to the acceleration component data in the vertical direction;

The current priori noise covariance matrix calculation unit 82 is configured to calculate the current priori noise covariance matrix according to the acceleration component data in the vertical direction and the previous posterior noise covariance matrix;

The current measurement altitude calculation unit 83, which is used to calculate the current measurement altitude according to the air pressure data;

The current Kalman gain calculation unit 84 is configured to calculate the current Kalman gain according to the current noise covariance;

The current output height fusion calculation unit 85 is configured to fuse and calculate the current output height according to the current prior height, the current measured height and the current Kalman gain;

A posteriori noise covariance updating unit 86, configured to update the current priori noise covariance matrix according to the current Kalman gain to obtain a current posteriori noise covariance matrix.

In some embodiments, the state data acquisition module includes an accelerometer and a barometer.

The present invention also provides a device including a processor and a memory, the memory is used for storing a computer program, and the processor is used for executing the computer program stored in the memory, so that the device can execute any one of the above Elevation position estimation method for indoor moving objects.

The device provided in this embodiment includes a processor, a memory, a transceiver, and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and communicate with each other. The memory is used for storing computer programs, and the communication interface is used for communication. , the processor and the transceiver are used for running a computer program, so that the device executes each step in the above method for estimating the elevation position of an indoor moving body.

In this embodiment, the memory may include random access memory (Random Access Memory, RAM for short), and may also include non-volatile memory (non-volatile memory), such as at least one disk memory.

The above-mentioned processor may be a general-purpose processor, including a central processing unit (Central Processing Unit, referred to as CPU), a network processor (Network Processor, referred to as NP), etc.; may also be a digital signal processor (Digital Signal Processing, referred to as DSP) , Application Specific Integrated Circuit (ASIC for short), Field-Programmable Gate Array (FPGA for short) or other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components.

The above-mentioned embodiments merely illustrate the principles and effects of the present invention, but are not intended to limit the present invention. Anyone skilled in the art can make modifications or changes to the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or changes made by those with ordinary knowledge in the technical field without departing from the spirit and technical idea disclosed in the present invention should still be covered by the claims of the present invention.

Claims (15)

1. a method for estimating the elevation position of an indoor moving body, is characterized in that, comprising: Collect real-time state data of the moving object under test at a specified frequency, and the state data at least include air pressure data and acceleration data; The acceleration component data in the vertical direction of the acceleration data and the remaining data in the state data are equally divided according to time series to form a first sub-data set of several specified durations; Input the first sub-data set of the moving object under test within the current specified time period into a pre-generated motion state recognition classifier, and determine the identification of the moving object under test according to the first sub-data set. Which state of the motion state is the flat motion state and the non-flat motion state? If the moving object to be tested is in the flat ground motion state, the current output height corresponding to the current specified duration is equal to the previous output height corresponding to the previous specified duration, If the detected moving body is in the non-flat motion state, the current output height is estimated by fusing all the first sub-data sets through an extended Kalman filter method. 2. The method for estimating the elevation position of an indoor moving body according to claim 1, wherein the method for identifying the motion state of the measured moving body by the motion state identification classifier comprises: extracting categorical features of the first sub-data set, Identify the motion state of the moving object under test according to the classification feature. 3. The method for estimating the elevation position of an indoor moving object according to claim 2, wherein the training method of the motion state recognition classifier comprises: The state data of the moving object under test in the flat ground motion state and the non-flat ground motion state are collected at the specified frequency as training data, Dividing the acceleration data in the training data, the acceleration component data of the vertical component and the rest of the data in the training data, evenly in time series to form a plurality of second sub-data sets of the specified duration; extracting the classification features of each of the second sub-data sets to form training samples; The training samples are input into a support vector machine model for training, the parameters of the support vector machine model are optimized, and the parameter-optimized support vector machine model forms the motion state recognition classifier. 4 . The method for estimating the elevation position of an indoor moving object according to claim 3 , wherein the classification features include vertical acceleration mean, vertical acceleration variance, air pressure difference, and air pressure variance. 5 . 5 . The method for estimating the elevation position of an indoor moving object according to claim 1 , wherein the state of the flat ground motion includes stillness, walking and jogging, and the non-flat ground motion state includes upstairs and downstairs. 6 . 6. The method for estimating the elevation position of an indoor moving object according to claim 1, wherein the method for estimating the current output height value by fusing the first sub-data set with an extended Kalman filter method comprises: A system state equation is constructed by using the acceleration component data in the vertical direction, and the current prior height is calculated; Construct a priori noise covariance matrix calculation formula according to the acceleration component data in the vertical direction and the previous a posteriori noise covariance matrix, and calculate the current priori noise covariance matrix; Construct a system measurement equation through the air pressure data to calculate the current measurement altitude; Build a Kalman gain calculation formula according to the prior noise covariance matrix, and calculate the current Kalman gain; Construct a current output height fusion calculation equation according to the current prior height, the current measured height and the current Kalman gain, and calculate the current output height; The current a priori noise covariance matrix is obtained by updating the current prior noise covariance matrix according to the current Kalman gain. 7. The elevation position estimation method of indoor moving body according to claim 6, is characterized in that: The state equation of the system is:

Among them, i refers to the current specified duration, i-1 refers to the previous specified duration,

is the current prior height,

is the previous output height corresponding to the previous specified duration,

is the relative height calculation function, the variables sa _i =[a ₁ ,a ₂ ,...,a _n ] ^T in the height calculation function, a ₁ ,a ₂ ,...,a _n refers to the The acceleration data in the vertical direction arranged in time series in the first sub-data set, n represents the total number of times the acceleration data in the vertical direction is collected within the current specified time period, and a ₁ represents the current specified time period. the first acceleration data in the vertical direction within the duration; _an represents the last acceleration data in the vertical direction within the current specified duration; The calculation formula of the prior noise covariance matrix is:

in,

is the current prior noise covariance matrix corresponding to the current specified duration,

is the posterior noise covariance matrix corresponding to the specified duration described in the previous section,

is the relative height calculation function described in the system state equation

The Jacobian matrix of ,

Yes

The transpose matrix of , Q is the process noise covariance; The system measurement equation is:

in,

represents the current measurement altitude, P _i represents the current air pressure value, P ₁ represents the initial air pressure value, and P ₀ represents the standard atmospheric pressure; the current Kalman gain calculation formula is:

where K _i represents the current Kalman gain, R _Q represents the measurement variance,

and

are the noise variances of the accelerometer and barometer, respectively; The current output height fusion calculation equation is:

The formula for updating the current prior noise covariance matrix to obtain the current posterior noise covariance matrix according to the current Kalman gain is:

8. A system for estimating an elevation position of an indoor moving body, comprising: a state data acquisition module, which is used for real-time acquisition of the state data of the moving object under test, the state data at least including air pressure data and acceleration data; a data preprocessing module, configured to equally divide the acceleration component data in the vertical direction of the acceleration data and the remaining data in the state data according to time series to form a first sub-data set of several specified durations; The motion state recognition classifier is used to read the first sub-data set of the tested moving body within the current specified time period, and determine the identification of the tested moving body according to the first sub-data set. Which state of the motion state is in the flat ground motion state and the non-flat ground motion state; An output height calculation module, which is used to calculate the current output height corresponding to the current specified duration according to the recognition result of the motion state recognition classifier, If the measured moving body is in the flat ground motion state, the output height calculation module outputs the previous output height corresponding to the previous specified duration as the current output height; If the moving object under test is in the non-flat motion state, the output height calculation module estimates the current output height by fusing all the first sub-data through an extended Kalman filter method. 9. The system for estimating the elevation position of an indoor moving object according to claim 8, wherein the motion state recognition classifier comprises: a feature processing unit, the feature processing unit is used to extract the classification features of the first sub-data set; A state determination unit, which is used for identifying the motion state of the moving object under test according to the classification feature of each type. 10 . The system for estimating the elevation position of an indoor moving object according to claim 9 , wherein the classification features include vertical acceleration mean, vertical acceleration variance, air pressure difference, and air pressure variance. 11 . 11 . The system for estimating the elevation position of an indoor moving object according to claim 8 , wherein the state of the flat ground motion includes stillness, walking and jogging, and the non-flat ground motion state includes upstairs and downstairs. 12 . 12. The system for estimating the elevation position of an indoor moving body according to claim 8, wherein the output height calculation module comprises: A current prior height calculation unit, configured to calculate the current prior height according to the acceleration component data in the vertical direction; A current priori noise covariance matrix calculation unit, configured to calculate the current priori noise covariance matrix according to the acceleration component data in the vertical direction and the previous a posteriori noise covariance matrix; The current measurement altitude calculation unit, which is used to calculate the current measurement altitude according to the air pressure data; A current Kalman gain calculation unit, configured to calculate the current Kalman gain according to the current noise covariance; A current output height fusion calculation unit, configured to fuse and calculate the current output height according to the current prior height, the current measured height and the current Kalman gain; A posteriori noise covariance updating unit, configured to update the current priori noise covariance matrix according to the current Kalman gain to obtain a current posteriori noise covariance matrix. 13. The system for estimating the elevation position of an indoor moving object according to claim 8, wherein the state data acquisition module comprises an accelerometer and a barometer. 14. A storage medium on which a computer program is stored, characterized in that: when the program is executed by a processor, the method for estimating the elevation position of an indoor moving body according to any one of claims 1 to 7 is implemented. 15. A device, characterized by comprising a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the device executes the method according to claim 1 to The method for estimating the elevation position of an indoor moving body according to any one of 7.

Images