JP7032721B2 - Channel state prediction device, channel state prediction method, wireless communication device and wireless communication method - Google Patents

Description

Application of Article 30, Paragraph 2 of the Patent Act (1) Published in the Proceedings of the 2017 IEICE General Conference on March 7, 2017 (2) Electronic Information and Communication on March 25, 2017 Presented at the 2017 General Conference Lecture

The present invention relates to a channel state prediction device, a channel state prediction method, a wireless communication device using the channel state prediction device, and a wireless communication method.

LTE (Long Term Evolution) Release 8 (Rel-8), which is a conventional wireless communication system, for example, a wireless communication system standardized by 3GPP (3rd Generation Partnership Project), communicates using a band of up to 20 MHz. It is possible to do.

Furthermore, in LTE-A (Long Term Evolution-Advanced), which is an advanced version of LTE, the bandwidth supported by LTE is the basic unit in order to realize higher speed transmission while ensuring backward compatibility with LTE. Carrier Aggregation (CA: Carrier Aggregation) technology is adopted in which a plurality of component carriers (CC: Component Carriers) are bundled and used at the same time, and a wide band transmission of 100 MHz width can be realized by using a maximum of 5 CC (100 MHz width). However, such carrier aggregation is transmission using different channels in adjacent frequency bands.

Despite the above-mentioned speedups, the demand for mobile communication traffic has increased sharply in recent years with the spread of high-performance mobile terminals such as smartphones.

As a result, in addition to the conventional expansion of the use of wireless LAN (Local Area Network), offloading to wireless LAN has progressed due to the increase in mobile data traffic due to the spread of smartphones, and unlicensed bands (2.4 GHz band, 5 GHz band). ) Is experiencing a surge in traffic.

In addition, due to the development of the IoT (Internet Of Things) / M2M (Machine to Machine) society, there is concern that the above frequency bands and the 920 MHz band will become even tighter, and improving the frequency utilization efficiency of these frequency bands has become an urgent issue. There is.

Here, since the usage status of the radio resource fluctuates depending on the time, place, frequency band, radio channel, etc., a situation may occur in which only a part of the frequency band (or the radio channel) is congested.

However, the existing self-employed wireless system (for example, IEEE802.11 wireless LAN) uses a single frequency band or determines one band to be used in advance before communication. For example, IEEE802.11n is used after setting whether to use the 2.4 GHz band or the 5 GHz band. Therefore, even if the existing self-employed wireless system as a whole has free wireless resources, congestion may occur.

Here, cognitive radio technology is attracting attention in order to make effective use of wireless communication resources. Cognitive radio technology means that a wireless terminal recognizes the usage status of surrounding radio waves and changes the wireless communication resource to be used according to the usage status. Cognitive radio technology includes a heterogeneous type in which different wireless communication standards are selected and used according to the situation, and a frequency sharing type in which a wireless terminal searches for a free frequency and secures a necessary communication band.

In the heterogeneous type, the cognitive radio recognizes multiple radio systems operating in the vicinity, obtains information on the utilization of each system and feasible transmission quality, and connects to the appropriate radio system. That is, the heterogeneous type cognitive radio indirectly enhances the utilization efficiency of frequency resources by increasing the utilization efficiency of wireless systems existing in the vicinity.

On the other hand, in the frequency sharing type, the cognitive radio has a frequency resource (this is called a white space) that is temporarily or locally not used in the frequency band in which another radio system is operated. Is detected and signal transmission is performed using this. That is, the frequency sharing type cognitive radio directly enhances the utilization efficiency of frequency resources in a certain frequency band.

Then, as a method for solving the problem of increased traffic in the unlicensed band as described above, a plurality of wireless LAN standards having different frequency bands (for example, 2.4 GHz band wireless LAN standard and 5 GHz band wireless LAN standard) are used. A heterogeneous cognitive radio approach that is selected or used in parallel can be considered (for example, Patent Document 1 and Patent Document 2).

However, in this heterogeneous cognitive radio approach, it is necessary to divide the transmission data as appropriate and to divide in advance which frequency band each is to be transmitted. As a result, there are new problems such as the transmission delay greatly differing depending on the frequency band used depending on the degree of congestion in each frequency band, and the order in which data arrives at the destination is changed.

Therefore, it is desirable to realize cognitive wireless communication by sharing a frequency with an existing system in a plurality of frequency bands that are largely separated from each other, for example, a 2.4 GHz band wireless LAN and a 5 GHz band wireless LAN.

However, it is not always clear what kind of data should be distributed and how to determine the transmission timing when communicating in parallel in a plurality of frequency bands separated from each other.

Therefore, we observe the usage status (availability of each radio channel, etc.) of multiple frequency bands (one or more radio channels in each frequency band), and based on such observation results, we use the transmission timing and use. It is conceivable to control the radio channel. Specifically, the timing of transmission and the wireless channel to be used are determined according to how long each wireless channel is available or when it is available.

However, for such transmission control, it is necessary to predict the future usage status from the past observation results.

At this time, considering that the availability and busy probability are frequently predicted each time a frame transmission is attempted, the prediction algorithm needs to maintain a low calculation amount and good accuracy.

In a situation where many wireless devices perform wireless transmission and various communication traffics are mixed, it is considered that the wireless resource usage pattern is complicated. It is very difficult to model with a simple expression for such a situation. Therefore, designing prediction algorithms and reducing complexity are important issues.

Here, as the conventional technology for predicting the usage status of wireless communication, there are the following.

Non-Patent Document 1 proposes a prediction algorithm for radio resource utilization (COR: channel occupation ratio) based on the AR (autoregressive) method. However, it does not mention the prediction of the probability of occurrence for the duration of the busy / idle state on the real-time axis.

In Non-Patent Document 2, traffic is modeled using a hidden Markov model to predict COR. Further, in Non-Patent Document 3, COR is analyzed and predicted by using an autoregressive integrated moving average (ARIMA) model. However, the probability of occurrence for the start / end timing and duration of busy / idle is not predicted.

Non-Patent Document 4 examines a wireless channel utilization model based on actual measurement results. It also outlines the channel situation prediction method. However, no effective prediction method for the start / end timing of busy / idle has been shown.

JP-A-2011-21433 Japanese Patent Application Laid-Open No. 2013-187561

S. Kaneko, S. Nomoto, T. Ueda, S. Nomura, and K. Takeuchi, ”Predicting radio resource availability in cognitive radio-an experimental examination,” in Proc. CrownCom, Singapore, pp. 1-8, May 2008 .. E. Chatziantoniou, B. Allen and V. Velisavljevie, "An HMM-based spectrum occupancy predictor for energy efficient cognitive radio," IEEE PIMRC, London, pp. 601-605, Sept. 2013. Z. Wang and S. Salous, ”Spectrum occupancy statistics and time series models for cognitive radio,” J. Signal Process. Syst., Vol. 62, no. 2, pp. 145-155, Feb. 2011. Y. Chen and H.S. Oh, "A Survey of Measurement-Based Spectrum Occupancy Modeling for Cognitive Radios," IEEE Communications Surveys & Tutorials, vol. 18, no. 1, pp. 848-859, First quarter 2016.

However, as described above, in order to communicate in parallel in a plurality of frequency bands separated from each other, the usage status of the multiple frequency bands is observed, and based on the observation result, one or more momentarily unused. It is necessary to transmit radio frames at the same time in various frequency bands and radio channels. In this case, the data is mapped to a plurality of bands and transmitted, and the receiving side collectively receives the plurality of bands and integrates the data.

By doing so, even if the congestion situation is biased among the bands, the transmission opportunity can be secured, so that the improvement of the frequency utilization efficiency and the reduction of the transmission delay can be expected, and the problem that the arrival order of the data is changed does not occur.

However, in order to perform efficient frame transmission using multiple frequency bands, it is important to appropriately determine the transmission timing and the radio channel to be used, and for this, which frequency band / radio channel is used at which timing. It is necessary to predict with high accuracy whether or not will be available. Therefore, from the observation results of the usage status of a plurality of frequency bands, an effective prediction method regarding the start / end timing of the channel empty / busy state (idle / busy) is required for each frequency band.

The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to determine the usage status of a plurality of frequency bands when communicating in parallel in a plurality of frequency bands separated from each other. From the observation results, a channel state prediction device, a channel state prediction method, and a wireless communication device and wireless communication using the channel state prediction device, which can effectively predict the start / end timing of the channel empty / blocked state for each frequency band. Is to provide a method.

According to one aspect of the present invention, it is a channel state prediction device for predicting the duration of a busy state and an idle state of a radio channel that performs random access control in a target frequency band, and is a radio in the frequency band. The channel usage status observation unit that observes the channel usage status at multiple measurement points, and the self-return model for multiple measurement points according to the observed usage points at multiple measurement points, use the channel at the next point in time. It is equipped with a channel usage prediction unit that predicts the situation and generates forecast information, and the channel usage prediction unit determines the busy / idle state of the radio channel and measures the duration of the busy / idle state. Duration prediction error probability acquisition that predicts the next busy / idle occurrence time from the idle duration acquisition unit and the acquired busy / idle duration, and calculates the probability of occurrence of a prediction error with respect to the value measured by the channel usage observation unit. The duration prediction error probability acquisition unit learns the parameters of the prediction formula of the self-return model from the acquired busy / idle duration, and predicts the next busy / idle occurrence time using the learned parameters. Then, the duration occurrence probability prediction unit that calculates the prediction error occurrence probability for the measured value by the channel usage status observation unit and predicts the occurrence probability of the next busy / idle duration based on the prediction error occurrence probability is further added. Including .

According to another aspect of the present invention, it is a channel state prediction method for predicting the duration of the busy state and the idle state of a radio channel that performs random access control in the target frequency band, and is a radio in the frequency band. Channel usage at multiple measurement points in a channel usage observation step, and a self-return model for multiple measurement points according to the observed usage at multiple measurement points, to use the channel at the next point in time. A step of predicting a situation and generating predictive information is provided, and the step of generating predictive information determines the busy / idle state of the radio channel and measures the duration of the busy / idle state. The acquisition step and the duration prediction error probability acquisition step that predicts the next busy / idle occurrence time from the acquired busy / idle duration and calculates the probability of occurrence of a prediction error with respect to the measured value in the channel usage observation step. Including, in the duration prediction error probability acquisition step, the parameters of the prediction formula of the self-return model are learned from the acquired busy / idle duration, and the next busy / idle occurrence time is predicted using the learned parameters. It further includes a duration occurrence probability prediction step that calculates the occurrence probability of a prediction error with respect to the measured value in the channel usage observation step and predicts the occurrence probability of the next busy / idle duration based on the occurrence probability of the prediction error.

According to still another aspect of the present invention, it is a wireless communication device for transmitting a signal by using a plurality of radio channels that perform random access control in each of a plurality of frequency bands separated from each other. A digital signal processing unit for dividing data into a plurality of partial data corresponding to each of a plurality of frequency bands and generating a digital signal of a transmission packet for each frequency band, and a digital signal processing unit provided for each frequency band are provided. Observed by a plurality of high-frequency signal processing units for converting signals into high-frequency signals for each corresponding frequency band, and a channel usage status observation unit that observes the usage status of radio channels in multiple frequency bands at multiple measurement points. It is equipped with a channel usage status prediction unit that predicts the channel usage status at the next time point and generates prediction information by a self-return model for multiple measurement time points according to the usage status at multiple measurement time points. The situation prediction unit determines the busy / idle state of the radio channel, measures the busy / idle duration, and the busy / idle duration acquisition unit, and the next busy / idle generation from the acquired busy / idle duration. The duration prediction error probability acquisition unit includes a duration prediction error probability acquisition unit that predicts the time and calculates the probability of occurrence of a prediction error for the measured value by the channel usage observation unit, and the duration prediction error probability acquisition unit is the acquired busy / idle duration. Learn the parameters of the prediction formula of the self-return model, predict the next busy / idle occurrence time using the learned parameters, calculate the probability of occurrence of prediction error with respect to the measured value by the channel usage observation unit, and predict. It further includes a duration occurrence probability predictor that predicts the probability of the next busy / idle duration based on the error occurrence probability, and controls the digital signal processing unit and the high frequency signal processing unit based on the prediction information. It is provided with an access control unit that synchronously transmits each partial data as a packet for each of a plurality of frequency bands by a plurality of radio channels at the same timing.

According to still another aspect of the present invention, it is a wireless communication method for transmitting a signal by using a plurality of radio channels that perform random access control in each of a plurality of frequency bands separated from each other. The step of dividing the data into multiple partial data corresponding to each of the multiple frequency bands and generating the digital signal of the transmission packet for each frequency band, and for each frequency band, for each frequency band corresponding to the digital signal. A step of converting to a high-frequency signal of The self-return model for the measurement time point includes a step of predicting the channel usage status at the next time point and generating prediction information, and the step of generating the prediction information determines the busy / idle state of the radio channel. Prediction error for the measured value in the channel usage observation step by predicting the next busy / idle occurrence time from the busy / idle duration acquisition step that measures the duration of the busy / idle state and the acquired busy / idle duration. In the duration prediction error probability acquisition step, which includes the duration prediction error probability acquisition step for calculating the occurrence probability of, the parameters of the prediction formula of the self-return model are learned from the acquired busy / idle duration, and the learned parameters. Predicts the next busy / idle occurrence time using, calculates the probability of occurrence of a prediction error with respect to the measured value in the channel usage observation step , and based on the probability of occurrence of the prediction error, of the next busy / idle duration Duration for predicting the occurrence probability A step of synchronously transmitting each partial data as a packet for each of a plurality of frequency bands at the same timing by a plurality of radio channels based on the prediction information, including a duration occurrence probability prediction step. And.

According to the present invention, from the observation result of the usage status of a plurality of frequency bands, the start / end timing of the empty / blocked state of the channel is effectively predicted for each frequency band, and the transmission data is mapped to the plurality of frequency bands. However, it is possible to perform data transmission by adjusting the transmission timing.

It is a conceptual diagram for demonstrating the structure of the wireless communication system of this embodiment. It is a figure for demonstrating a specific example for mapping transmission data into a plurality of bands and transmitting, and collectively receiving and integrating on a receiving side. It is a functional block diagram for demonstrating the configuration of the transmission apparatus 1000 of this embodiment. It is a timing chart for demonstrating the operation of the channel usage condition observation unit 1060, the channel usage condition statistics prediction unit 1070, and the access control unit 1080. It is a flowchart for demonstrating the operation of the channel use condition observation unit 1060, the channel use condition statistics prediction unit 1070, and the access control unit 1080 of the transmission apparatus 1000. It is a functional block diagram explaining the structure for predicting the occurrence probability of the busy / idle duration in the channel usage statistics prediction unit 1070. It is a conceptual diagram for explaining an autoregressive model method. It is a conceptual diagram for demonstrating the procedure of getting a busy / idle duration. It is a conceptual diagram which shows the procedure for learning the parameter of a prediction formula. It is a flowchart which shows the procedure of the prediction of the busy / idle continuation probability. It is a figure explaining the structure of the experiment performed for the prediction process. It is a figure which shows the calculation result of the complementary cumulative distribution function of the prediction error of a busy state and an idle state. It is a figure which shows the occurrence probability RB ( ^{X B} _new ) and ^RI ( ^XI _new ) of a busy / idle duration. It is a functional block diagram for demonstrating the example of the more detailed structure of the transmission apparatus 1000. It is a functional block diagram for demonstrating the configuration of the receiving apparatus 2000. It is a functional block diagram for demonstrating the example of the more detailed structure of the receiving apparatus 2000.

Hereinafter, the configurations of the wireless communication system and the wireless communication device according to the embodiment of the present invention will be described. In the following embodiments, the components and processing steps having the same reference numerals are the same or equivalent, and the description thereof will not be repeated if they are not necessary.

In the following, as an example of explaining the receiving device of the present invention, a plurality of existing unlicensed bands (for example, 920 MHz band used for IoT and the like, used for wireless LAN), which are largely separated from each other as described above, 2 1. An embodiment of a transmission device in a wireless communication system capable of cognitive wireless communication by sharing a frequency with an existing system in the 4 GHz band and the 5 GHz band) will be described.

However, the wireless communication device of the present invention is not necessarily limited to such a case, and more generally, using a plurality of frequency bands separated from each other, the simultaneous and parallel timings are synchronized by the same wireless method. It can be applied to a receiving device that communicates with. Further, in the wireless communication device of the present invention, as will be described later, it may be applied to a receiving device that uses a plurality of frequency bands separated from each other and simultaneously communicates at a timing synchronized by different wireless methods. It is possible.

Then, in the present embodiment, as described below, modeling and prediction methods based on machine learning are used in combination to obtain a channel empty / busy state (idle / busy) in each frequency band / channel in the actual environment. By predicting the time to, we provide a probability distribution for the duration of the empty / closed state. Further, the method of calculating the probability distribution with respect to the duration of the empty / closed state can be applied to a configuration in which communication is performed using at least one frequency band among a plurality of frequency bands.

Further, in the following, "carrier sense" means sensing for detecting the presence or absence of a signal of a target wireless channel and determining the transmission timing by virtual carrier sense accompanied by power detection or decoding of a received signal. The term "channel sensing" means, in addition to sensing as a carrier sense, sensing that monitors communication or the like in order to grasp the usage status of the target channel.

FIG. 1 is a conceptual diagram for explaining the configuration of the wireless communication system of the present embodiment.

With reference to FIG. 1, on the transmission side, assuming that three frequency bands of 920 MHz band, 2.4 GHz band, and 5 GHz band are used, it is assumed that one wireless channel is used in each band, and the transmission frame is used. To configure.

Although a plurality of channels may be used in each frequency band, it will be described below assuming that one channel is used for each frequency band.

In this embodiment, wireless access control having the following features is performed.

That is, first, the transmitting side observes the usage status of a plurality of frequency bands (vacancy status of each radio channel, etc.) by a method as described later.

Subsequently, the transmitting side simultaneously transmits a radio packet (frame) in one or more unused frequency bands / radio channels at a certain timing. At this time, the transmission data is mapped to a plurality of bands and transmitted.

On the other hand, on the receiving side, multiple bands are collectively received and the data is integrated.

With such a configuration for transmission and reception, transmission opportunities can be secured even if the congestion situation is biased between bands, so improvement in frequency utilization efficiency and reduction in transmission delay can be expected, and the order of data arrival is changed. No problem occurs.

FIG. 2 is a diagram for explaining a specific example for mapping transmission data to a plurality of bands and transmitting the data, and collectively receiving and integrating the transmission data on the receiving side.

As shown in FIG. 2, the transmission data is divided into each band by the number of symbols proportional to the transmission rate Ri of each band using the transmission series, and is assigned to each band by serial / parallel conversion.

For example, if (5 GHz band transmission rate: 2.4 GHz band transmission rate: 920 MHz band transmission rate) = (R1: R2: R3) = (3: 2: 1), the transmission data series is divided into 6 symbols. The 5 GHz band (ch1), 2.4 GHz band (ch2), and 920 MHz band (ch3) are assigned 3 symbols, 2 symbols, and 1 symbol. It should be noted that when the transmission series is divided and assigned, it is not limited to such a case, and more generally, when m frequency bands are used, the ratio of the transmission rates of the frequency bands is set to (R1: R2). : ...: Rm) (Ratio is expressed as a contract), the transmission series is divided into (R1 + R2 + ... + Rm) x n (m, n: natural number) symbols, and (m, n: natural number) is assigned to each channel. R1 × n) symbols, (R2 × n) symbols, ..., (Rm × n) symbols may be assigned.

After such allocation, a physical header is attached to the transmission symbol for each band to form a packet, and these packets are transmitted simultaneously and in parallel at the same timing.

The number of symbols allocated to each band on the transmitting side is stored as information in this physical header.

On the receiving side, synchronization and demodulation processing are performed using the physical headers on each band. Each demodulated series is combined by parallel / serial conversion in the reverse process of the transmission side, and the frame is decoded.
[Distractor configuration]
FIG. 3 is a functional block diagram for explaining the configuration of the transmission device 1000 of the present embodiment.

With reference to FIG. 3, the transmission device 1000 performs an error correction coding unit 1110 for performing error correction coding processing on a series of transmission data, and an interleaving process on the data after error correction coding. For the interleaving unit 1112 to be performed, the serial / parallel conversion (hereinafter referred to as S / P conversion) unit 1010 for performing the processing assigned to each frequency band as described with reference to FIG. 2, and the data after the S / P conversion. Wireless frame generator 1020.1 to 1020.3 for executing digital processing that forms a wireless frame (packet) for communication by a predetermined wireless communication method, such as mapping processing and addition of a physical header for each frequency band. And, for the digital signals from the wireless frame generators 1020.1 to 1020.3, digital-to-analog conversion processing and modulation processing to a predetermined modulation method (for example, orthogonal modulation for a predetermined multi-value modulation method, respectively). High-frequency processing unit (RF unit) 1040.1 to 1040.3 that executes processing), up-conversion processing, power amplification processing, etc., and antenna 1050 for transmitting high-frequency signals of RF unit 1040.1 to 1040.3, respectively. .1 to 1050.3 are included. The operation of the RF units 1040.1 to 1040.3 is controlled based on the clock from the local oscillator 1030 provided in common with them.

Further, the transmitter 1000 has a channel usage status observation unit 1060 for observing the usage status (vacancy status of each radio channel, etc.) of each frequency band (one or more radio channels in each frequency band), and a channel usage status. Based on the observation of the observation unit 1060, the channel usage status statistic prediction unit 1070 that predicts the channel usage status at a predetermined timing, the processing timing of the radio frame generation units 1020.1 to 1020.3, and the transmission in the RF unit. It includes an access control unit 1080 that controls the timing so as to simultaneously transmit radio packets in an unused frequency band / radio channel for a predetermined period at the same controlled transmission timing.

Here, the channel usage status observation unit 1060 is configured to execute the carrier sense and channel sensing described above.

Here, the access control unit 1080 selects and uses a channel found to be usable according to the result of carrier sensing of the target band as a candidate at the time of transmission, and is unused at the same controlled transmission timing. Wireless packets will be transmitted simultaneously in the frequency band and wireless channel.

A detailed operation example of the channel usage statistics prediction unit 1070 will be described later.

As described with reference to FIG. 2, the transmission device 1000 having such a configuration maps the data to a plurality of bands and transmits the data, and the receiving side collectively receives the plurality of bands and integrates the data.

FIG. 4 is a timing chart for explaining the operations of the channel usage status observation unit 1060, the channel usage status statistic prediction unit 1070, and the access control unit 1080.

With reference to FIG. 4, the channel usage status observation unit 1060 observes the usage status of each frequency band (for example, the availability status and busy probability of each radio channel), and the channel usage status statistic prediction unit 1070 observes each frequency. The latest usage status of the band is predicted, and from the result, the access control unit 1080 determines transmission parameters such as transmission timing, frequency band used, and wireless channel so that good communication can be performed.

That is, as will be described in detail later, when communication is performed using, for example, three frequency bands, the channel usage statistics prediction unit 1070 uses the current time as a reference, for example, two bands at time t2. It is predicted that it can be transmitted by using it, and if it is time t3, it is predicted that 3 bands can be used. The access control unit 1080 determines the transmission start timing and the frequency band used based on the prediction result of the usage status in order to perform efficient transmission.

For example, in the conventional random access control in a wireless LAN or the like, transmission is performed immediately when a transmission opportunity is obtained by CSMA / CA and random backoff.

On the other hand, the access control unit 1080 of the present embodiment waits for transmission until a plurality of frequency bands / wireless channels can be used simultaneously even if a transmission opportunity is obtained in some wireless channels, if necessary. , Is performed.
(Configuration and operation of channel usage statistics prediction unit 1070)
Hereinafter, the configuration and operation of the channel usage statistics prediction unit 1070 will be described.

The channel usage status statistic prediction unit 1070 receives information to be observed / measured as the usage status of the radio channel as an input, calculates the usage status statistic of each radio channel calculated from the observation / measurement result, and starts transmission. Calculate the usage information required for the decision.

The following are examples of items to be observed / measured as the usage status of radio channels and usage status statistics of each radio channel calculated from the results of the above observation / measurement.

i) State of each radio channel (busy or idle: physical carrier sense result)
ii) busy duration of each radio channel ii-1) Frame length described in the physical header of the frame being received ii-2) Value of NAV described in the MAC header of the frame being received (virtual carrier sense) result)
NAV is a Network Allocation Vector (transmission prohibition period).

The following are examples of usage statistics for each radio channel calculated from the results of the above observations and measurements.

i) Probability of being busy (temporal utilization rate)
ii) Probability distribution of duration of busy state and idle state (the derivation method will be described later).
iii) Probability distribution of occurrence probability of idle / busy state duration with respect to immediately preceding busy / idle state duration (for example, probability density function (PDF) and cumulative probability (CDF))
iv) Occurrence pattern of busy state and idle state (period and duty ratio: when background traffic is periodic)
For the sake of explanation of terms, a general method for avoiding a collision of transmissions from each terminal in a wireless LAN will be briefly described.

In a wireless LAN channel, a method called "CSMA (Carrier Sense Multiple Access)" is adopted because packets collide with each other and efficient communication cannot be established unless they wait for transmission.

In the case of wireless, it is not possible to tell whether a collision will occur just by monitoring the strength of the radio waves. This is because radio waves are greatly attenuated depending on the distance, so if the person causing the collision is far away, the radio waves may not be detected.

Therefore, be sure to set a "waiting time (DIFS: Distributed access Inter Frame Space)" before transmission, and confirm that there is no other transmission signal before transmission. Such a method is called "CA (Collision Avoidance)".

Then, after transmission, it always waits for "ACK (ACKnowledgement, arrival confirmation response)", and if ACK does not return, it is determined that a collision or the like has occurred and re-transmission is performed.

In addition to this, as an access control mechanism peculiar to wireless LAN, for example, there is "RTS / CTS (Request to Send / Clear to Send)" devised as a countermeasure against hidden terminals. Here, the hidden terminal is a terminal that is out of the radio range from oneself but is within the radio range of the communication partner. Its existence cannot be known directly, but it causes interference.

Assuming that the reach of the radio wave is Lm, consider a situation in which the communication partner B (access point) of the wireless terminal A is Lm ahead and another wireless terminal C is further ahead of that Lm.

At this time, since the radio wave of the terminal C does not reach the terminal A, the terminal C does not know the existence of the terminal C even if the terminal A checks whether another terminal is transmitting a signal (even if the carrier senses). It becomes a hidden terminal of terminal A. If no measures are taken, even if the terminal C is transmitting to the access point B, the terminal A may also transmit the data to the access point B. This causes a collision at the access point B and is a factor that lowers the throughput.

RTS / CTS is a mechanism in which all wireless devices issue an "RTS (transmission request)" packet before transmission, and if the receiving side can also receive, respond with "CTS (receivable)". In the above example, the terminal C first transmits the RTS to the access point B. However, this RTS does not reach the terminal A.

The access point B notifies the terminal C that it can be received by transmitting the CTS. Since this CTS also reaches the terminal A, the terminal A senses that communication is being performed and postpones the transmission. In the RTS / CTS packet, the scheduled occupancy period of the channel is written, and communication is suspended during that period. This period is called "NAV (Network Allocation Vector, transmission prohibition period)".

FIG. 5 is a flowchart for explaining the operations of the channel usage status observation unit 1060, the channel usage status statistic prediction unit 1070, and the access control unit 1080 of the transmission device 1000 described with reference to FIG.

With reference to FIG. 5, first, the channel usage status observation unit 1060 performs carrier sense in a plurality of bands, calculates usage status information, and updates the usage status information stored in a storage device (not shown) ( S100).

That is, the channel usage status observation unit 1060 determines the busy / idle state of each radio channel to be used in each of the plurality of frequency bands, and measures the duration thereof.

The access control unit 1080 determines whether there is data to be transmitted (S102), and if there is no data to be transmitted yet (N in S102), the process returns to S100.

On the other hand, when there is data to be transmitted (Y in S102), the access control unit 1080 first determines whether or not the wireless channel for which the transmission opportunity has been obtained satisfies the "immediate transmission condition" as described below. ..

That is, although the access control unit 1080 normally determines whether the transmission timing has arrived based on the prediction result of the channel usage statistics prediction unit 1070, it is actually busy / idle. ) Since there is a risk that the transmission opportunity will not be obtained as expected due to an error in the prediction of the occurrence of the state, for example, if it is determined that the combination of the following conditions is satisfied for the wireless channel for which the transmission opportunity can be secured, the wireless channel Is used to control the transmission to start immediately (S104). That is, in this case, if the access control unit 1080 determines that the predetermined communication quality can be achieved by performing transmission on the wireless channel that has a transmission opportunity, the access control unit 1080 immediately does not wait for the transmission timing based on the prediction result. Controls the wireless transmission of.

a1) Total transmission rate is above a predetermined value a2) Transmission delay of transmission data is below a predetermined value when transmission is started immediately a3) When transmission is started immediately, throughput is increased by a predetermined amount or more a4) Transmission opportunity was secured. When transmission is performed on a wireless channel, the distribution and / or average of the usage rate among the wireless channels becomes smaller. A5) When transmission is performed on a wireless channel for which a transmission opportunity has been secured, the energy consumption required for transmission is less than a predetermined amount a6) Transmission opportunity is obtained on a predetermined radio channel a7) When loss of transmission opportunity is not allowed A1) to any one of the above conditions a1) to a7) are satisfied, or conditions a1) to a7 ) Satisfyes the predetermined combination (combination of two or more conditions), the access control unit 1080 immediately starts transmission using the radio channel for which the transmission opportunity is secured. That is, the access control unit 1080 determines the transmission parameter and maps the transmission data (S108), controls the S / P conversion unit 1010 and the wireless frame generation units 1020.1 to 1020.3, and selects the frequency. The frame is transmitted on the band and the radio channel (S110), and the process is returned to step S100.

Here, the "transmission parameters" include "band used and radio channel used", "transmission rate used in each radio channel", and "in the case of each radio channel (Orthogonal Frequency Division Multiplexing)". Transmission power of each subcarrier is also possible).

If a predetermined condition is satisfied, it may be possible to transmit using only a part of the radio channels instead of the radio channels of all available frequency bands.

Further, for the determination of the transmission rate and the transmission power, existing methods as described in the following documents can be used.

Literature: Tomoaki Yoshinori, Seiichi Sampei, Norihiko Morinaga, "OFDM Adaptive Modulation Method Using Multi-Level Transmission Power Control for High-Speed Data Transmission," IEICE Journal (B), J84-B, 7, pp. 1141-1150, July 2001 In addition, the propagation path information necessary for determining the transmission rate and transmission power can be obtained by, for example, the following method.

i) Use the propagation path estimation result performed when receiving the frame received in the reverse communication.

ii) Use the propagation path feedback method specified by the IEEE 802.11 wireless LAN.

Subsequently, when the access control unit 1080 does not satisfy the "immediate transmission condition" (N in S104), whether or not the transmission timing has arrived based on the prediction result of the channel usage status statistic prediction unit 1070 as described above. (S106)
For the determination of the transmission start timing, the prediction information based on the usage status information is used as follows.

b1) Prediction of the time required for a busy radio channel to become idle (Prediction of "how long should I wait?")
For this, by using the following information, it is possible to predict "how long should I wait?".

b1-1) The length of the frame or NAV that is the cause of busyness (if already known)
b1-2) Whether or not a busy state is generated for each radio channel after an arbitrary time (in the case of periodic background traffic, from the busy (busy) state and idle (idle) state cycle and duty. Predictable)
b1-3) Idle (idle) occurrence probability for future waiting time based on the past busy (busy) duration (can be calculated from the busy (busy) state and idle (idle) state CDF)
b2) Prediction of the time required for an idle wireless channel to become busy (prediction of "how long can you wait?")
b2-1) Presence or absence of busy state of each radio channel after an arbitrary time (If it is periodic background traffic, it can be predicted from the busy (busy) state and idle (idle) state cycles and duty)
b2-2) Busy occurrence probability for future waiting time based on the idle (idle) duration so far (can be calculated from the CDF in the busy state and the idle state)
The access control unit 1080 "predicts the time required for the wireless channel in the busy state to become idle" as described above, and "the idle wireless channel is busy". By combining with "estimation of the time required until becomes", the transmission timing "the expected value of the transmission speed is maximized" is calculated.

That is, by combining the above two predictions, the access control unit 1080 assumes that the transmission rate of each radio channel has a predetermined value, and is in an idle state at each time timing in a predetermined time range from the present time. It is possible to predict the maximum value of the amount of data that can be transmitted by the wireless channel. When such a predicted amount of data that can be transmitted (transmission speed) exceeds a value corresponding to a predetermined transmission speed, the data is transmitted at the transmission timing.

That is, if the access control unit 1080 determines that such a transmission timing has arrived (Y in S106), the access control unit 1080 determines the transmission parameter and maps the transmission data (S108), and performs S / P conversion. The unit 1010 and the radio frame generation units 1020.1 to 1020.3 are controlled to transmit a frame in the selected frequency band and radio channel (S110), and the process is returned to step S100.

If the access control unit 1080 determines in step S106 that the predicted amount of data that can be transmitted (transmission speed) does not exceed the value corresponding to the predetermined transmission speed, the access control unit 1080 waits for transmission to exceed the current value. Assuming that transmission opportunities may be obtained with the number of radio channels, transmission is waited for (N in step S106), and the process returns to step S100. This is because it is considered that improvement of frequency utilization efficiency and reduction of transmission delay can be achieved by performing such a standby operation.

As a criterion for the access control unit 1080 to determine whether or not the transmission timing has arrived, the following may be adopted.

c1) Minimize the time required to complete transmission of transmission data (because if transmission ends early, many frames can be transmitted and less resources are wasted).
c2) Minimize the amount of free resources generated until the transmission is completed c3) Maximize the amount of data that can be transmitted within a specified time c4) Minimize the distribution and average of the usage rate of each wireless channel (effective use of wireless resources) However, in order to eliminate extremely crowded channels.)
c5) Minimize the required transmission energy under the condition that transmission of transmission data is completed within a certain time c6) Transmission outage (transmission failure / loss of transmission opportunity) probability is less than a predetermined value c7) Use of a specific radio channel by itself The rate is less than or equal to the specified value (in order to keep the time utilization rate of the frequency band within the limit in the frequency band where the transmission time is limited, such as the 920MHz band).
(Prediction of probability of occurrence of busy / idle duration)
The procedure for predicting the probability of occurrence of the busy / idle duration by the channel usage statistics prediction unit 1070 will be described below.

FIG. 6 is a functional block diagram illustrating a configuration for predicting the occurrence probability of the busy / idle duration in the channel usage status statistic prediction unit 1070.

The channel usage statistics prediction unit 1070 determines the busy / idle state of the channel based on the information from the channel usage observation unit 1060, and acquires the busy / idle duration to measure the duration of the busy / idle state. Part 1072 and the acquired busy / idle duration are used to learn the parameters of the prediction formula described later, and the learned parameters are used to predict the next busy / idle occurrence time, and the probability of occurrence of a prediction error with respect to the measured value P ( It includes a duration prediction error probability acquisition unit 1074 that evaluates Err), and a duration occurrence probability prediction unit 1076 that predicts the occurrence probability R (X + Err) of the next busy / idle duration.

Hereinafter, the operation of each part of the channel usage statistics prediction unit 1070 will be described.

First, as a premise, the channel usage statistics prediction unit 1070 predicts the busy / idle duration based on the autoregressive model (AR) method.

Here, in the AR method, the next busy / idle duration is predicted from the latest p busy / idle durations Xi, ..., Xi-p by the following prediction formula.

Here, if the immediately preceding channel state is busy (Xi is the idle duration), the idle duration is predicted as Xi + 1, and the immediately preceding channel state is idle (Xi is the busy duration). In the case of), the busy duration shall be predicted as Xi + 1.

Further, the coefficients a ₀ , ... Ap of the above prediction formula ( _p -th order) are learned by using the training section.

The procedure for predicting the probability of occurrence for busy / idle duration is summarized below.

1) Obtain the busy / idle duration by sensing.

2) Learn the parameters of the prediction formula from the acquired busy / idle duration.

3) Predict the next busy / idle occurrence time X using the learned parameters.

4) Evaluate the probability P (Err) of the prediction error with respect to the measured value.

5) Predict the probability R (X + Err) of the next busy / idle duration by the following equation.

FIG. 7 is a conceptual diagram for explaining the autoregressive model method.

In the autoregressive model for time series prediction, the parameters (AR coefficient (coefficients a ₁ , ..., a _p ) and the variance of the prediction error) are obtained from the sample time series (self-covariance function estimated from it). Ask from. Here, the Levinson-Durbin algorithm, which is a high-speed algorithm for solving the Yule-Walker equation that satisfies the autocovariance function, is generally used.

Such a procedure is a well-known procedure, for example, as disclosed in Non-Patent Document 1 described above.

As for the method of predicting the busy / idle duration, a prediction based on machine learning other than the AR method may be used. For example, Markov Chain based HMM (Markov Chain based HMM) prediction method, autoregressive moving average prediction method (ARMA prediction method, ARMA: Autoregressive moving average). Here, the ARMA prediction method is disclosed in Non-Patent Document 4.

Hereinafter, the procedure of the autoregressive model method will be described in more detail.

(1) The busy / idle duration acquisition unit 1072 acquires the busy / idle duration.

FIG. 8 is a conceptual diagram for explaining a procedure for acquiring a busy / idle duration.

The channel usage status observation unit 1060 performs carrier sense in each band / channel, and the busy / idle duration acquisition unit 1072 determines whether the channel is busy / idle. Further, the busy / idle duration acquisition unit 1072 measures the busy / idle duration based on the busy / idle determination result. The measurement may be continuous time, or may be a discrete value with a fixed width as a unit for the purpose of information compression.

For example, FIG. 8 shows an example in which WLAN is assumed and processing is performed using discrete values in units of 9 μs, which is the contention slot length. In this case, the busy state and the duration of the idle can be expressed by the number of points in this unit time. In the example of FIG. 8, it corresponds to 1 point = 9 μs. In the following, time will be expressed using points.

(2) Learn the parameters a ₁ , ..., a _p of the prediction formula from the acquired busy / idle duration series.

FIG. 9 is a conceptual diagram showing a procedure for learning the parameters of the prediction formula.

Subsequently, the duration prediction error probability acquisition unit ₁₀₇₄ learns the parameters a ₁ , ..., Ap from the duration measurement results of N busy / idle respectively. Therefore, no special data needs to be sent for learning.

As described above, the duration prediction error probability acquisition unit 1074 calculates the parameters a ₁ , ..., a _p by solving the normal equation generated from the least squares equation by using the Levinson-Durbin induction method.

(3) The next busy / idle occurrence time X is predicted using the learned parameters.

Using the coefficients a ₁ , ..., a _p obtained from the learning results, the next busy / idle generation time X (= X _{i + 1} ) is predicted by the above-mentioned prediction formula of the AR method.

(4) The probability of occurrence of prediction error P (Err) with respect to the measured value, which is a complementary cumulative distribution function (CCDF) of the probability of occurrence of prediction error Err, is evaluated.

Here, the following relationship holds.

(5) The probability R (Xnew) of the next busy / idle duration Xnew is predicted from the prediction result X of the busy / idle duration and the prediction error occurrence probability P (Err).

The duration occurrence probability prediction unit 1076 predicts the occurrence probabilities RB ( ^{X B} _new ) and ^RI ( ^XI _new ) of the next busy / idle duration by the following procedure.

The access control unit 1080 "whether to transmit now" based on the probability of occurrence of the next busy / idle duration RB ( ^{X B} _new ) and ^RI ( ^XI _new ) calculated as described above. ? ”Judgment is carried out.

FIG. 10 is a flowchart showing a procedure for predicting a busy / idle continuation probability.

First, the channel usage status observation unit 1060 measures the busy / idle duration by the carrier sense, and the busy / idle duration acquisition unit 1072 acquires the busy / idle duration (S200).

Subsequently, if the prediction formula parameters are being learned / updated (Yes in S202), the duration prediction error probability acquisition unit 1074 will use the parameters a ₁ , ..., a _p from the duration measurement results of N busy / idle respectively. Is learned, the prediction formula is updated (S204), and the process returns to step S200.

On the other hand, if the prediction formula parameter is not being learned / updated (No in S202), the duration prediction error probability acquisition unit 1074 determines whether the prediction formula parameter needs to be updated (S210).

Here, the case where the prediction expression parameter needs to be updated may be, for example, as follows.

i) When the number of active terminals (the number of terminals sending and receiving data) changes.

ii) When new traffic is generated or transmission of some traffic is completed.

iii) When the MCS used is changed. (Because the distribution of busy duration changes.)
iv) When the variance of the prediction error of the busy / idle duration becomes larger than the predetermined amount (or ratio) compared to the time when the learning is completed.

Here, MCS (Modulation and Coding Scheme) represents a modulation method / channel coding rate. In order to change the transmission rate according to the state of the communication path, a set of different modulation methods / channel coding rates is generally prepared as a table.

Returning to FIG. 10 again, when the predictive parameter needs to be updated (Yes in S210), the process returns to step S200.

On the other hand, when it is not necessary to update the prediction expression parameter (No in S210), the duration prediction error probability acquisition unit 1074 predicts the busy / idle time (S212).

Here, the busy / idle duration predictions are alternated. For example, if the previous state is busy, the idle duration is predicted.

Further, the duration prediction error probability acquisition unit 1074 uses the predicted value of the busy / idle duration calculated in step S212 and the busy / idle duration measured by the carrier sense in step S200 to predict the predicted value. The error (prediction error) between and the actual value is calculated (S214).

Further, the duration prediction error probability acquisition unit 1074 updates the complementary cumulative distribution function (CCDF) of the busy / idle duration prediction error based on the calculated prediction error (S216). Here, the CCDF of the prediction error is separately stored in a storage unit (not shown) for busy and idle.

The duration occurrence probability prediction unit 1076 calculates the occurrence probabilities RB ( ^{X B} _new ) and ^RI ( ^XI _new ) of the next busy / idle duration based on the updated complementary cumulative distribution function (S218). ), The probability of occurrence for the busy / idle duration is output to the access control unit 1080 (S220).
(Results of measurement and prediction experiments)
FIG. 11 is a diagram illustrating the configuration of an experiment performed for the prediction process as described above.

1KP video signal or 320k audio data is uploaded from the terminal STA1 to the access point AP using Ethernet (registered trademark), and from the access point AP using the indoor 5GHz band WLAN, the terminal. It shall be downloaded to STA2. At this time, another sensing device detects a busy / idle state, and based on this detection result, a complementary cumulative distribution function (CCDF) of the prediction error of the duration is calculated for the busy state and the idle state, and the busy and idle states are calculated. Predict the probability of idle duration.

Here, it is assumed that the sensing is performed every 9 μs, and the training data for calculating the parameters a ₁ , ..., _Ap is 200 samples (N = 100).

Further, as the order of the prediction formula, the second order is adopted.

FIG. 12 is a diagram showing the calculation result of the complementary cumulative distribution function of the prediction error of the busy state and the idle state.

In particular, in the idle state, the prediction error is within 20 points and is suppressed to 50% or less.

FIG. 13 is a diagram showing the probability of occurrence of busy / idle duration RB ( ^{X B} _new ) and ^RI ( ^XI _new ).

It can be seen that a near-accurate prediction is achieved when the busy period is 1400 points, 2000 points, or when the idle period is 6300 points or 5600 points.
[Detailed configuration of wireless communication device]
FIG. 14 is a functional block diagram for explaining an example of a more detailed configuration of the transmission device 1000.

The functional block diagram shown in FIG. 14 shows, as an example, a configuration of a transmission device according to a wireless communication method similar to the wireless communication standard 802.11a.

That is, although the wireless communication standard 802.11a is a wireless LAN communication method in the 5 GHz band, in FIG. 14, the wireless communication method has the same configuration except that the frequency band is different even in the 2.4 GHz and 920 MHz bands. The receiver according to the above shall be used.

Therefore, in each frequency band, the configuration of the preamble portion of the packet is common to a plurality of frequency bands.

However, it is not always necessary that the wireless communication methods of each frequency band have the same configuration, and even if the wireless communication method (signal format, symbol length, subcarrier interval, etc.) is different for each frequency band. good. In this case, at least a single transmission series may be divided into each band and transmitted at the same time, and the configuration of the RF portion may be basically the same except that the frequency bands are different, and the configuration of the preamble portion of the packet is sufficient. (Preamble length, etc.) may be different for each of the plurality of frequency bands.

FIG. 14 schematically shows a configuration related to transmission in the 5 GHz band. Since the same wireless communication method as the wireless communication standard 802.11a is assumed, the signal to be transmitted shall be OFDM (Orthogonal Frequency Division Multiplexing) modulation.

With reference to FIG. 14, the radio frame generation unit 1020.3 receives the transmission data distributed from the S / P conversion unit 1010, and executes the mapping unit 1122 for executing the mapping process and the inverse Fourier transform process. It includes an IFFT unit 1130 for the purpose, a GI addition unit 1140 for adding a guard interval portion, and a digital-to-analog converter (DAC) 1150 for converting a digital signal into an analog signal of an I component and a Q component.

The high frequency processing unit 1040.3 includes an orthogonal modulator 1210 for modulating the signal from the DAC 1150 into a predetermined multi-valued modulation signal, an upconverter 1220 that upconverts the output of the orthogonal modulator 1210, and an output of the upconverter 1220. Includes a power amplifier 1230 for power amplifying and transmitting from the antenna 1050.3.

As a result, the baseband OFDM signal is converted into a carrier band OFDM signal by the RF unit 1040.3.

Further, the high frequency processing unit 1040.3 is based on the clock frequency conversion unit 1310 for converting the reference frequency signal from the local oscillator 1030 into the reference clock signal of the corresponding frequency band, and the reference clock from the clock frequency conversion unit 1310. Then, based on the clock generation unit 1320 that generates the clock used for the modulation processing in the orthogonal demodulator 1210 and the reference clock from the clock frequency conversion unit 1310, the clock used for the up-conversion processing in the upconverter 1220 is generated. The clock generation unit 1340 and the clock generation unit 1340 are included.

That is, the reference frequency signal from the local oscillator 1030 is used as a clock signal in the conversion of such a baseband OFDM signal to a carrier band OFDM signal. More generally, even when the wireless communication method is different, the reference frequency signal from the local oscillator 1030 is basically used as a clock signal in the conversion from the baseband signal to the carrier band signal.

The channel usage status observation unit 1060 observes the usage status of each frequency band (for example, availability of each radio channel, busy probability, etc.) based on the sensing result of its own station, and the channel usage status statistic prediction unit 1070 observes each frequency. The latest usage status of the band is predicted, and the access control unit 1080 controls the transmission timing accordingly.
[Receiver configuration]
Hereinafter, the configuration of the receiving device used in the wireless communication system as described with reference to FIG. 2 will be described.

FIG. 15 is a functional block diagram for explaining the configuration of the receiving device 2000 according to the present embodiment.

With reference to FIG. 15, the receiving device 2000 has antennas 2010.1 to 2010.3 and antennas 2010.1 for receiving signals in a plurality of frequency bands (920 MHz band, 2.4 GHz band, 5 GHz band), respectively. Commonly provided for receiving units 2100.1 to 2100.3 and receiving units 2100.1 to 2100.3 for executing reception processing such as down-conversion processing and demodulation / decoding processing of signals of ~ 2010.3. The local oscillator 2020 that generates the reference frequency signal, which is the reference clock for the operation of the receiving units 2100.1 to 2100.3, and each sequence of the signals from the receiving units 2100.1 to 2100.3 are used as the transmitting side. In the reverse process, it includes a parallel / serial conversion unit 2700 for combining by parallel / serial conversion.

The output of the integrated frame from the parallel / serial (P / S) conversion unit 2700 is passed to the upper layer.

The receiving device 2000 detects the frequency offset of the local oscillator 2020 from the preamble signal of the received signal, generates a signal (oscillation frequency control signal) for controlling the oscillation frequency of the local oscillator 2020, and performs carrier frequency synchronization processing. Also includes a synchronization processing unit 2600 that generates a signal (synchronization timing signal) for timing synchronization in digital signal processing from the received signal.

Upon receiving the signal from the antenna 2010.1, the receiving unit 2100.1 receives low noise amplification processing, down-conversion processing, demodulation processing for a predetermined modulation method (for example, orthogonal demodulation processing for a predetermined multi-value modulation method), and analog. High frequency processing unit (RF unit) 2400.1 for executing digital conversion processing, etc., and baseband for executing baseband processing such as demodulation / decoding processing for digital signals from RF unit 2400.1. Includes processing unit 2500.1.

The receiving unit 2100.2 also includes a high frequency processing unit (RF unit) 2400.2 and a baseband processing unit 2500.2 for performing similar processing for the corresponding frequency band. Further, the receiving unit 2100.3 also includes a high frequency processing unit (RF unit) 2400.3 and a baseband processing unit 2500.3 for performing the same processing for the corresponding frequency band.

The baseband processing unit 2500.1 to 2500.3 and the parallel / serial (P / S) conversion unit 2700 are collectively referred to as a digital signal processing unit 2800.

FIG. 16 is a functional block diagram for explaining an example of a more detailed configuration of the receiving device 2000 shown in FIG.

The functional block diagram shown in FIG. 16 also shows, as an example, a configuration of a receiving device that follows the same wireless communication method as the wireless communication standard 802.11a.

Therefore, the configuration of the receiving device corresponds to the configuration of the transmitting device shown in FIG.

FIG. 16 also exemplifies the configuration of the receiver 2100.3 in the 5 GHz band.

With reference to FIG. 16, the RF unit 2400.3 of the receiving unit 2100.3 is used to frequency-convert the output of the low noise amplifier 3010 and the low noise amplifier 3010 for amplifying the received signal from the antenna 2010.3. Down converter 3020, automatic gain controller 3030 for controlling the output of the down converter 3020 so as to have a predetermined amplitude, an orthogonal demodulator 3040 for demodulating a predetermined multi-valued modulated signal, and an orthogonal demodulator. It includes an analog-to-digital converter (ADC) 3050 for converting the I component output and the Q component output of the 3040 into digital signals, respectively.

The RF unit 2400.3 is further based on a clock frequency conversion unit 3060 for converting a reference frequency signal from the local oscillator 2020 into a reference clock signal in the corresponding frequency band, and a reference clock from the clock frequency conversion unit 3060. Based on the clock generation unit 3070 that generates the clock used for the down-conversion process in the down converter 3020 and the reference clock from the clock frequency converter 3060, the clock used for the demodulation process in the orthogonal demodulator 3040 is generated. It includes a clock generation unit 3080.

Since the same wireless communication method as the wireless communication standard 802.11a is assumed, the transmitted signal is OFDM (Orthogonal Frequency Division Multiplexing) modulated. As a result, the carrier band OFDM signal is converted into a baseband OFDM signal by the RF unit 2400.3.

The reference frequency signal from the local oscillator 2020 is then used for carrier frequency synchronization in such conversion of the carrier band OFDM signal to the baseband OFDM signal. More generally, even if the wireless communication method is different, the reference frequency signal from the local oscillator 2020 is basically used for the carrier frequency synchronization in the conversion from the carrier band signal to the baseband signal.

The baseband processing unit 2500.3 receives the signal from the ADC 3050 and performs the GI removal unit 4010 for removing the guard interval portion and the fast Fourier transform for the signal from which the guard interval is removed. It includes an FFT unit 4020 and a demapping unit 4032 for executing a demapping process on the output of the FFT unit 4020.

After performing guard interval removal, FFT processing, and demapping processing in the baseband processing units 2500.1 to 2500.3, the received data is combined with the signals of each frequency band by the P / S conversion unit 2700. The deinterleave processing by the deinterleavement unit 4042 and the error correction processing by the error correction unit 4040 are executed.

Here, the synchronization timing signal output from the synchronization processing unit 2600 is used for symbol timing synchronization for detecting the start of the OFDM symbol.

More generally, even if the wireless communication method is different, basically, the synchronization timing signal output from the synchronization processing unit 2600 is used as the synchronization signal in the baseband processing.

With the configuration as described above, it is possible to efficiently execute multi-channel simultaneous sensing when communicating in parallel in a plurality of frequency bands separated from each other. Further, it is possible to map each transmission data to a plurality of frequency bands and adjust the transmission timing to perform data transmission.

The embodiments disclosed this time are examples of configurations for concretely implementing the present invention, and do not limit the technical scope of the present invention. The technical scope of the present invention is shown by the scope of claims, not the description of embodiments, and includes modifications within the wording scope of the claims and the scope of equal meaning. Is intended.

1000 Transmitter, 1010 S / P converter, 1020.1 to 1020.3 wireless frame generator, 1030 local oscillator, 1040.1 to 1040.3 RF, 1050.1 to 1050.3 antenna, 1060 channel usage status Observation unit, 1070 channel usage statistics prediction unit, 1080 access control unit, 1110 error correction coding unit, 1112 interleaving unit, 2000 receiver, 2010.1 to 2010.3 antenna, 2020 local oscillator, 2100.1 to 2100 .3 Receiver unit, 2400.1 to 2400.3 RF unit, 2500.1 to 2500.3 baseband processing unit, 2600 synchronization processing unit, 2700 P / S conversion unit, 2800 digital signal processing unit, 4040 error correction unit, 4042 Deinterleaved part.

Claims (5)

It is a channel state prediction device for predicting the duration of the busy state and idle state of a radio channel that performs random access control in the target frequency band.
A channel usage status observation unit that observes the usage status of the radio channel at a plurality of measurement points in the frequency band,
With a channel usage status prediction unit that predicts the channel usage status at the next time point and generates prediction information by the autoregressive model for the plurality of measurement points according to the observed usage status at the plurality of measurement points. Equipped with
The channel usage prediction unit is
A busy / idle duration acquisition unit that determines the busy / idle state of the wireless channel and measures the duration of the busy / idle state.
Includes a duration prediction error probability acquisition unit that predicts the next busy / idle occurrence time from the acquired busy / idle duration and calculates the probability of occurrence of a prediction error with respect to the value measured by the channel usage observation unit. The duration prediction error probability acquisition unit learns the parameters of the prediction formula of the self-return model from the acquired busy / idle duration, and predicts the next busy / idle occurrence time using the learned parameters. , Calculate the probability of occurrence of a prediction error for the value measured by the channel usage observation unit,
A channel state predictor further comprising a duration occurrence probability predictor that predicts the probability of occurrence of the next busy / idle duration based on the probability of occurrence of the prediction error . It is a channel state prediction method for predicting the duration of the busy state and idle state of a radio channel that performs random access control in the target frequency band.
A channel usage status observation step for observing the usage status of the radio channel in the frequency band at a plurality of measurement time points,
It is provided with a step of predicting the channel usage status at the next time point and generating prediction information by the autoregressive model for the plurality of measurement time points according to the usage situation at the plurality of measurement time points observed.
The step of generating the prediction information is
A busy / idle duration acquisition step that determines the busy / idle state of the wireless channel and measures the duration of the busy / idle state.
Includes a duration prediction error probability acquisition step that predicts the next busy / idle occurrence time from the acquired busy / idle duration and calculates the probability of occurrence of a prediction error with respect to the measured value in the channel usage observation step . In the duration prediction error probability acquisition step, the parameters of the prediction formula of the self-return model are learned from the acquired busy / idle duration, and the next busy / idle occurrence time is predicted using the learned parameters. Then, the probability of occurrence of a prediction error with respect to the measured value in the channel usage observation step is calculated.
A channel state prediction method further comprising a duration occurrence probability prediction step for predicting the occurrence probability of the next busy / idle duration based on the occurrence probability of the prediction error. A wireless communication device for transmitting signals using a plurality of radio channels that perform random access control in each of a plurality of frequency bands separated from each other.
A digital signal processing unit for dividing transmission data into a plurality of partial data corresponding to each of the plurality of frequency bands and generating a digital signal of a transmission packet for each of the plurality of frequency bands.
A plurality of high-frequency signal processing units provided for each of the frequency bands and for converting the digital signal into a high-frequency signal for each corresponding frequency band.
A channel usage status observation unit that observes the usage status of the radio channel at a plurality of measurement points in the plurality of frequency bands.
With a channel usage status prediction unit that predicts the channel usage status at the next time point and generates prediction information by the autoregressive model for the plurality of measurement points according to the observed usage status at the plurality of measurement points. Equipped with
The channel usage prediction unit
A busy / idle duration acquisition unit that determines the busy / idle state of the wireless channel and measures the duration of the busy / idle state.
Includes a duration prediction error probability acquisition unit that predicts the next busy / idle occurrence time from the acquired busy / idle duration and calculates the probability of occurrence of a prediction error with respect to the value measured by the channel usage observation unit. The duration prediction error probability acquisition unit learns the parameters of the prediction formula of the self-return model from the acquired busy / idle duration, and predicts the next busy / idle occurrence time using the learned parameters. , Calculate the probability of occurrence of a prediction error for the value measured by the channel usage observation unit,
It further includes a duration occurrence probability prediction unit that predicts the occurrence probability of the next busy / idle duration based on the occurrence probability of the prediction error.
Based on the prediction information, the digital signal processing unit and the high frequency signal processing unit are controlled, and each of the partial data is synchronized and the same timing as a packet for each of the plurality of frequency bands by the plurality of radio channels. A wireless communication device including an access control unit for transmission by. The wireless communication device according to claim 3 , wherein the duration prediction error probability acquisition unit executes a process of updating the parameter of the prediction formula according to a traffic situation. It is a wireless communication method for transmitting signals by using a plurality of wireless channels that perform random access control in each of a plurality of frequency bands separated from each other.
A step of dividing the transmission data into a plurality of partial data corresponding to each of the plurality of frequency bands and generating a digital signal of a transmission packet for each of the plurality of frequency bands.
For each of the frequency bands, a step of converting the digital signal into a corresponding high frequency signal for each frequency band,
A channel usage status observation step for observing the usage status of the radio channel at a plurality of measurement time points in the plurality of frequency bands,
It is provided with a step of predicting the channel usage status at the next time point and generating prediction information by the autoregressive model for the plurality of measurement time points according to the usage situation at the plurality of measurement time points observed.
The step of generating the prediction information is
A busy / idle duration acquisition step that determines the busy / idle state of the wireless channel and measures the duration of the busy / idle state.
Includes a duration prediction error probability acquisition step that predicts the next busy / idle occurrence time from the acquired busy / idle duration and calculates the probability of occurrence of a prediction error with respect to the measured value in the channel usage observation step . In the duration prediction error probability acquisition step, the parameters of the prediction formula of the self-return model are learned from the acquired busy / idle duration, and the next busy / idle occurrence time is predicted using the learned parameters. Then, the probability of occurrence of a prediction error with respect to the measured value in the channel usage observation step is calculated.
It further includes a duration occurrence probability prediction step that predicts the occurrence probability of the next busy / idle duration based on the occurrence probability of the prediction error.
A wireless communication method comprising a step of synchronously transmitting each of the partial data as a packet for each of the plurality of frequency bands at the same timing by the plurality of radio channels based on the prediction information.

Images