WO2021184424A1

WO2021184424A1 - Indoor visible light communication positioning method and system based on machine learning and ofdm

Info

Publication number: WO2021184424A1
Application number: PCT/CN2020/082658
Authority: WO
Inventors: 游善红; 倪珅晟; 胡剑凌; 刘武; 王峰; 韩淑莹
Original assignee: 苏州大学
Priority date: 2020-03-18
Filing date: 2020-04-01
Publication date: 2021-09-23
Also published as: CN111404607A; CN111404607B

Abstract

The present invention relates to an indoor visible light communication positioning method and system based on machine learning and OFDM. The method comprises: dividing a medium-large indoor plane into multiple connected subareas, and distributing multiple LED devices on each subarea, the multiple LED devices being respectively fixed to the vertex of the subarea, and the adjacent subareas sharing the LED devices, wherein identity information is set in the LED devices; performing quadrature amplitude modulation on binary data streams required to be sent, then respectively placing communication data and identity information data on the respectively distributed subcarriers, and performing STBC coding on the communication data; performing direct-current bias optical orthogonal frequency division multiplexing modulation on the processed data; and sending respective data on the multiple LED devices in the same subarea in the form of optical signals, sending the optical signals, after being subjected to free space transmission, to a receiver, and receiving the optical signals by means of the receiver. The present invention is simple, and high in reliability.

Description

Indoor visible light communication positioning method and system based on machine learning and OFDM

Technical field

The present invention relates to the technical field of optical communication, in particular to an indoor visible light communication positioning method and system based on machine learning and OFDM.

Background technique

With the continuous development of economy and modern technology, people's indoor activity space is becoming larger and more complex, and the demand for navigation and positioning services is also increasing. Therefore, positioning technology has received extensive attention. At present, the most commonly used outdoor positioning technologies are the Global Positioning System and Beidou Navigation and Positioning System. These positioning technologies have better positioning results outdoors, but because the satellite positioning signals are difficult to penetrate large buildings, the indoor positioning effects are poor. In order to achieve high-precision indoor positioning, some indoor positioning technologies have been proposed, such as infrared, ultrasonic, radio frequency identification (RFID), wireless local area network (WLAN), Bluetooth (Bluetooth) and ultra-wideband (UWB). However, when the above-mentioned technical means are used to achieve positioning, a complex positioning facility environment needs to be built, which not only has high cost, but has limited positioning accuracy, and safety cannot be effectively guaranteed. In recent years, with the rapid development of solid-state lighting technology, a new generation of light-emitting diodes (Light-Emitting Diodes, LEDs) are replacing traditional incandescent lamps on a large scale with their advantages of high brightness, long life, short response time, and low cost. And energy-saving lamps to provide lighting. In addition, after research by scientific researchers, it is found that short-distance high-speed wireless communication can be realized indoors by using the high-speed flicker characteristics of LEDs. As an emerging wireless communication method, LED-based indoor visible light communication has obvious advantages compared with radio frequency wireless communication in terms of electromagnetic radiation, use environment, and safety. It can realize the dual functions of lighting and communication at the same time. Due to the many advantages of Visible Light Communication (VLC), the VLC system has received extensive attention and research. In order to further develop and improve the functions of the VLC system, related technologies that use visible light communication to achieve indoor positioning have subsequently appeared, which are similar to the above-mentioned traditional positioning technologies. In contrast, the positioning technology based on visible light communication has the advantages of high positioning accuracy, good confidentiality, and low cost. For these reasons, the use of visible light communication to achieve indoor positioning is considered an efficient solution.

In the VLC system, the signal generated by the signal generator passes through the LED drive circuit and is loaded on the LED lighting equipment, and the signal transmission is realized by controlling the brightness of the LED. In VLC, the commonly used signal modulation techniques include On-Off Keying (OOK), Pulse Positioning Modulation (PPM), Carrier-less Amplitude and Phase (CAP) and Quadrature Orthogonal Frequency Division Multiplexing (OFDM). Among them, both CAP and OFDM modulation technologies can achieve high-speed transmission with high spectrum efficiency under limited bandwidth conditions, and OFDM is widely used because it can effectively combat frequency selective fading.

At present, the methods for realizing indoor visible light positioning based on white light LEDs are mainly divided into four categories: geometric measurement method, scene analysis method, approximate perception method and image sensor imaging method. The geometric measurement method and the approximate perception method are relatively simple to implement, but the positioning accuracy is low. The scene analysis method is also called the fingerprint positioning method, which requires the establishment of a fingerprint database. The richer the samples in the fingerprint database, the higher the positioning accuracy. However, the process of finding matching data in this method is time-consuming and the algorithm has poor adaptability. Once the indoor scene changes, its positioning accuracy will be greatly affected. The image sensor imaging law uses an optical camera for positioning, which has relatively high accuracy, but its algorithm is more complex, takes a long time to calculate, and uses a camera to achieve communication speed and reliability is low. In the case of a large indoor environment, when the required positioning area is large, only one set of LEDs cannot complete the positioning task of the entire positioning area. Therefore, it is necessary to deploy more LED light-emitting devices in large-scale positioning scenarios, and use the layout of LEDs to divide the entire positioning area into multiple sub-location areas. When the receiver reaches a certain sub-area, use the LED devices that belong to the sub-area. The data sent is used for communication and positioning.

In recent years, with the rapid improvement of computer performance and the widespread popularity of smart devices, machine learning (ML) has been widely used in various disciplines. Some products that rely on machine learning have successfully entered our lives, such as human faces. Recognition, autonomous driving and smart medical equipment, etc. Artificial Neural Network (ANN) is a supervised learning method in machine learning. Because of its strong robustness and fault tolerance, it can obtain optimized solutions to complex problems very quickly, so it is widely used. Support Vector Machine (SVM) is also a commonly used supervised machine learning algorithm. SVM can implement classification and regression tasks. SVM can perform non-linear classification and non-linear regression by introducing Kernel Method. Both types of machine learning methods need to be trained with sufficient data samples before they can be used. During training, the data is usually divided into training set and test set. The training set is used to train the machine learning model, and the test set is used to test the performance of the machine learning model. According to the test results, the structure and parameters of the model can be adjusted appropriately, and better performance can be obtained after retraining. A well-trained machine learning model can fully approximate complex nonlinear relationships, and the machine learning model has a self-learning function, which can make the system more adaptable. Therefore, if machine learning algorithms can be applied to visible light positioning, the positioning accuracy can be further improved.

In the current schemes that use machine learning algorithms to achieve indoor visible light communication positioning, most of the single-carrier modulation formats are used, and some schemes use OFDM modulation technology. Using OFDM technology combined with time division multiplexing for indoor positioning, the use of time division multiplexing will increase the complexity of the system and reduce the communication rate, and the area of the system positioning area is limited. Considering the case of a single receiver, the VLC system can be regarded as a multiple input single output (Multiple Input Single Output, MISO) communication system, so what the receiver captures is a superimposed signal. If the transmitted signal is modulated by OFDM, the receiver cannot directly demodulate the signal. The use of Space Time Block Code (STBC) coding technology can solve this problem, but STBC requires that the original data sent by all senders applying this code be the same, and then the receiver can recover the communication data well. , But at this time, each LED device cannot send its own positioning information, so that the system can only achieve positioning in one area, which makes the system unable to support positioning area division and cannot meet the positioning needs in medium and large indoor positioning scenarios.

Summary of the invention

To this end, the technical problem to be solved by the present invention is to overcome the problem that the prior art cannot meet the positioning requirements in medium and large indoor positioning scenarios, thereby providing a machine learning and OFDM-based method that can meet the positioning requirements in medium and large indoor positioning scenarios. Indoor visible light communication positioning method and system.

In order to solve the above technical problems, an indoor visible light communication positioning method based on machine learning and OFDM of the present invention includes: dividing a medium and large indoor plane into a plurality of connected sub-areas, and distributing a plurality of LED devices on each sub-areas , And a plurality of LED devices are respectively fixed on the vertices of the sub-regions, and adjacent sub-regions share LED devices, wherein the LED devices are provided with identity information; the binary data stream to be sent is subjected to quadrature amplitude modulation, Then place the communication data and identity information data on their assigned subcarriers, and perform STBC encoding on the communication data; perform DC bias optical orthogonal frequency division multiplexing modulation on the data after the above processing; The respective data on the multiple LED devices in the area are sent in the form of optical signals, and the optical signals are sent to the receiver after being transmitted in free space, and received by the receiver.

In an embodiment of the present invention, the size of the sub-region is adjusted according to the power of the LED device, the viewing angle of the receiver, the illumination requirement, and the size of the indoor space and other physical conditions.

In an embodiment of the present invention, when STBC encoding is performed on the complex signal, the original data content sent by each LED device in the sub-area must be the same.

In an embodiment of the present invention, before the STBC encoding of the complex signal, the communication data and the identity information data are allocated to different frequency bands, where the data used for communication has a larger proportion in the data packet, and it is placed On a continuous segment of carrier.

In an embodiment of the present invention, the frequency bands occupied by the identity information subcarriers allocated by the plurality of LED devices are different from each other.

The present invention also provides an indoor visible light communication positioning method based on machine learning and OFDM, which includes the following steps: the receiving end processes the received data; separates the communication data and the identity information data, and for the identity information data, according to the restoration of the identity Information data determines the sub-area where the receiver is located; for communication data, extract the pilot data and calculate the channel estimation matrix, after extracting the eigenvalues, use the machine learning model to locate, obtain the relative coordinates of the receiver, and use the channel estimation matrix for STBC decoding , Get the communication data; combine the obtained relative coordinates with the identity information data obtained by the receiver to calculate the final coordinates of the receiver.

In an embodiment of the present invention, the method for the receiving end to process the received data is: through DC bias optical orthogonal frequency division multiplexing demodulation, the time domain signal is converted into the frequency domain signal, and then used The frequency domain channel equalization technology performs channel estimation, equalizes the data of the identity information, decodes the communication data, and outputs a binary data stream after the identity information data and the communication data are restored and undergoes quadrature amplitude demodulation.

In an embodiment of the present invention, the method for positioning using a machine learning model is: in the offline training process, collecting training samples; extracting pilot data, calculating the channel estimation matrix, and extracting the eigenvalues. As the input parameters of the machine learning model, machine learning model training is used to obtain machine learning model positioning.

In an embodiment of the present invention, after the machine learning model training is completed, it is stored in the local memory of the receiver or in the cloud. When positioning, the input parameters of the network are uploaded to the cloud, and the cloud server is used to calculate the coordinates. , The cloud server calculates the coordinates and then downloads them to the lower computer.

The present invention also provides an indoor visible light communication positioning system based on machine learning and OFDM, including a dividing module for dividing a medium and large indoor plane into multiple connected sub-areas, and multiple LED devices are allocated on each sub-areas , And multiple LED devices are respectively fixed on the vertices of the sub-regions, and adjacent sub-regions share LED devices, wherein the LED devices are provided with identity information; the processing module is used to process the binary data stream that needs to be sent Quadrature amplitude modulation, and then place the communication data and identity information data on their assigned sub-carriers, and perform STBC encoding on the communication data; the modulation module performs DC offset optical orthogonal frequency division on the data after the above processing Multiplexing modulation; sending module, used to send the respective data on multiple LED devices in the same sub-area in the form of optical signals, and the optical signals are transmitted to the receiver after being transmitted in free space, and then received by the receiver.

Compared with the prior art, the above-mentioned technical solution of the present invention has the following advantages:

The indoor visible light communication positioning method and system based on machine learning and OFDM of the present invention can enable the indoor visible light communication positioning method based on machine learning algorithm and OFDM modulation technology to support positioning area division, so that this type of system can be suitable for medium and large indoors Scenes. The proposed method does not need to increase additional hardware cost and algorithm complexity, and the reliability of the system is high.

Description of the drawings

In order to make the content of the present invention easier to be understood clearly, the following further describes the present invention in detail according to specific embodiments of the present invention and in conjunction with the accompanying drawings.

Figure 1 is a flowchart of the first embodiment of the present invention;

Figure 2 is a schematic diagram of the LED distribution structure;

Fig. 3 is a schematic diagram of a spatial model of a single sub-region of the present invention;

Figure 4 is a schematic diagram of data sub-carrier allocation in the method of the present invention;

Fig. 5 is a flowchart of the second embodiment of the present invention.

Detailed ways

Example one

As shown in Figure 1, this embodiment provides an indoor visible light communication positioning method based on machine learning and OFDM, which includes: Step S1: Divide a medium and large indoor plane into a plurality of connected sub-areas, and allocate on each sub-areas Multiple LED devices, and multiple LED devices are respectively fixed on the vertices of the sub-regions, and adjacent sub-regions share LED devices, wherein the LED devices are provided with identity information; Step S2: Binary data to be sent The stream undergoes quadrature amplitude modulation, and then the communication data and identity information data are respectively placed on their assigned sub-carriers, and the communication data is encoded by STBC; Step S3: DC bias optical quadrature is performed on the data after the above processing Frequency division multiplexing modulation; Step S4: Send the respective data on the multiple LED devices in the same sub-area in the form of optical signals, and the optical signals are transmitted to the receiver after being transmitted in free space, and received by the receiver .

The indoor visible light communication positioning method based on machine learning and OFDM in this embodiment, for the transmitting end, in the step S1, the medium and large indoor plane is divided into multiple connected sub-areas, and multiple sub-areas are allocated to each sub-areas. LED devices, and multiple LED devices are respectively fixed on the vertices of the sub-regions, and adjacent sub-regions share LED devices, wherein the LED devices are provided with identity information, which is beneficial for the receiver to determine its specifics based on the positioning information. The position is conducive to the completion of positioning; in the step S2, since the positioning function is based on data communication, at the transmitting end, the binary data stream to be sent is subjected to quadrature amplitude modulation, and after quadrature amplitude modulation The signal is a complex signal, and then the communication data and identity information data are respectively placed on their assigned subcarriers, and the communication data is STBC encoded, that is, the complex signal is STBC encoded, and the identity information data is not STBC encoded , And after quadrature amplitude modulation, the identity information data is placed on the allocated sub-carriers, so as to help ensure that the receiver can receive the identity information sent by the multiple LED devices; in step S3, The processed data is subjected to DC bias optical orthogonal frequency division multiplexing modulation. After the signal is modulated, the electrical signal can be converted into an optical signal by hardware, which is conducive to the LED sending data; in the step S4, it will be in the same sub-region The respective data on the multiple LED devices in the LED devices are sent in the form of optical signals, and the optical signals are sent to the receiver after being transmitted in free space, and received by the receiver, which is beneficial to the receiver according to its relative coordinates and all data. The final coordinates are calculated from the position of the sub-area. This application can enable the indoor visible light communication positioning method based on machine learning algorithms and OFDM modulation technology to support positioning area division, so that it can be applied to medium and large indoor scenes without adding additional hardware costs and The algorithm complexity, so the reliability is high.

As shown in Figures 2 and 3, in this embodiment, the medium and large indoor plane is divided into a plurality of connected rectangular sub-areas, each sub-area is allocated four LED devices, and the four LED devices are fixed in the rectangular sub-areas. On the four vertices, two adjacent sub-areas share two LED devices. In addition, each LED device is assigned a unique identification (ID) information, and the data packet sent by each LED device contains its ID information (where the ID information includes the coordinates, number, status and other information of the LED device). The size of each sub-area can be appropriately adjusted according to the size of the indoor space, the power of the LED device, the viewing angle of the receiver, and the illumination requirements. However, it must be ensured that the receiver can receive the signals sent from the four LED devices in the sub-region in any sub-region, and the interference signals from the LED devices in other sub-regions are relatively small. After dividing the area, each LED device needs to be assigned a unique number, and each group of LED numbers corresponds to a sub-area number. The receiver can know which sub-area it is in according to the positioning information, which is very important for the receiver to complete the final positioning.

The communication data sent by the four LED devices in the same sub-area are the same, and they are sent simultaneously after STBC coding and OFDM modulation. In order to ensure that the receiver shown can receive the ID information sent by the four LED devices in the sub-area where it is located, this application proposes to use most of the data sub-carriers (Sub-Carrier, SC) for the transmission of communication data, and a small part of the data sub-carriers. To transmit the ID information of the LED device. The data on these sub-carriers for transmitting ID information does not need to be STBC encoded, but it is necessary to ensure that the frequency bands occupied by the ID information sub-carriers allocated by each LED device are different.

As shown in Figure 4, when STBC encoding is performed on a complex signal, the original data content sent by each LED device in the sub-region must be the same. In order to facilitate the sending and receiving of ID information, the communication data and ID information data need to be allocated to different frequency bands. Among them, the data used for communication has a larger proportion in the data packet, so more subcarriers need to be allocated, so it is placed on a continuous segment of carrier and STBC encoding is performed. In order to enable the receiver to receive the ID information sent by each LED device, each LED device in the same sub-area is allocated a carrier to send its own ID information. Only a small amount of sub-carriers used to transmit ID information are required. Assuming that each LED device is assigned an ID information carrier, the four ID information are placed on the four carriers in order. And this carrier is only used by one LED device in the sub-area, and the other three LED devices do not send any data on this carrier. The total frequency band length occupied by the LED devices in each sub-area for sending ID information is the same, but the sub-carriers allocated to each LED device need to be allocated appropriately to prevent the occurrence of the same sub-carriers allocated to different LED devices in the same sub-area of. The data on the four carriers used to transmit ID information does not need to be STBC encoded, just put it behind the communication data carrier. After all data is loaded on the carrier according to this principle, OFDM modulation is performed.

In this embodiment, the indoor visible light communication positioning function is based on data communication. Therefore, at the sending end, the binary data stream to be sent is first subjected to quadrature amplitude modulation (Quadrature Amplitude Modulation, QAM). The application takes 4-QAM as an example, and other quadrature amplitude modulation formats can also be selected; the signal after quadrature amplitude modulation is a complex signal, and then STBC encoding is performed on the complex signal. It is worth noting that the data in the ID information part does not need to be encoded by STBC. After quadrature amplitude modulation, it can be placed on the assigned sub-carrier; then the combined data is subjected to DC offset optical orthogonal frequency division. Multiplexing modulation (Direct Current Biased Optical Orthogonal Frequency Division Multiplexing, DCO-OFDM); Finally, the four LED devices in the same sub-area all send their data in the form of optical signals, and the optical signals are sent to The receiver.

Example two

As shown in Figure 5, this embodiment provides an indoor visible light communication positioning method based on machine learning and OFDM, which includes the following steps: Step S1: The receiving end processes the received data; Step S2: Separate communication data and identity Information data, for identity information data, determine the sub-area where the receiver is located according to the recovered identity information data; for communication data, extract pilot data and calculate the channel estimation matrix, after extracting the eigenvalues, use the machine learning model to locate and get the receiver’s Relative coordinates, while using the channel estimation matrix to perform STBC decoding to obtain communication data; combine the obtained relative coordinates with the identity information data obtained by the receiver to calculate the final coordinates of the receiver.

According to the indoor visible light communication positioning method based on machine learning and OFDM in this embodiment, for the receiving end, in the step S1, the receiving end processes the received data, thereby facilitating obtaining the final coordinates; in the step S2 Separate communication data and identity information data. For identity information data, determine the sub-area where the receiver is located according to the recovered identity information data; for communication data, extract pilot data and calculate channel estimation matrix, after extracting eigenvalues, use machine learning The model is located, the received relative coordinates are obtained, and the channel estimation matrix is calculated at the same time, and then STBC is used to decode the communication data to obtain the final coordinates of the receiver, which is beneficial to the calculation of the final coordinates. This application can be based on machine learning algorithms and The indoor visible light communication positioning method of OFDM modulation technology supports the division of positioning areas, which makes it applicable to medium and large indoor scenes, and does not need to increase the additional hardware cost and algorithm complexity, so the reliability is high.

The method for the receiving end to process the received data is: converting time-domain signals into frequency-domain signals through DC offset optical orthogonal frequency division multiplexing demodulation, and then using frequency-domain channel equalization technology to perform channel estimation, The data of the identity information is equalized, the communication data is decoded, and after the identity information data and the communication data are restored, the binary data stream is output after quadrature amplitude demodulation. Specifically, at the receiving end, the receiver uses the pilot information (this part of the pilot information requires STBC encoding) to calculate four channel estimation matrices, and then uses the channel estimation matrix and the decoder to obtain communication data, and the ID information is also passed The pilot information is restored after channel equalization, but since the ID information is not encoded by STBC, this part of the pilot information does not need to be encoded. Therefore, the receiver can obtain the positioning information sent by each LED device while receiving the communication data. The receiver can know which sub-area it is in according to the positioning information, which is very important for the receiver to complete the final positioning.

Specifically, in the online positioning process, firstly, data processing is also performed to obtain the channel estimation matrix, and the input parameters of the model are further obtained. At this time, input the obtained characteristic parameters into the trained model to obtain the relative three-dimensional coordinates of the receiver. The relative coordinates are the coordinates of the receiver in the current sub-region as the coordinate system, not the indoor space. The coordinates of the coordinate system, the ID information is first extracted from the data packet, and the ID information data can be recovered after another channel estimation. The receiver can determine which sub-area it is currently in based on the four ID information; finally, The receiver further calculates the final coordinates according to the relative coordinates and the position of the sub-area.

The method of positioning using the machine learning model is: in the offline training process, collecting training samples; extracting pilot data, calculating the channel estimation matrix, extracting the eigenvalues, using the eigenvalues as the input parameters of the machine learning model, using Machine learning model training obtains machine learning model positioning. Specifically, in order to implement positioning using a machine learning algorithm, it is first necessary to select a suitable machine learning model (such as ANN and SVM, etc.), and then further specify the machine learning model according to the physical model. In the offline training process, sufficient training samples need to be collected first to ensure a good training effect. When recording each set of samples, it is necessary to record the relative three-dimensional coordinates of the receiver and the corresponding channel estimation matrix under the current coordinates (the channel estimation matrix can be obtained during the data processing at the receiving end), and then use the channel estimation matrix to extract the eigenvalues. The feature values will be used as the input parameters of the machine learning model; after collecting a large number of samples, you can start training. During the training process, the structure and various parameters of the model can be adjusted appropriately to achieve better performance, and the trained machine learning The model can be stored in the local storage of the receiver; it can also be stored in the cloud. When positioning, the input parameters of the network are uploaded to the cloud, the cloud server is used to calculate the coordinates, and the server calculates the coordinates before downloading to the lower computer.

Example three

Based on the same inventive concept, this embodiment provides an indoor visible light communication positioning system based on machine learning and OFDM. The principle of solving the problem is similar to the indoor visible light communication positioning method based on machine learning and OFDM described in the first embodiment. I won't repeat it here.

The indoor visible light communication positioning system based on machine learning and OFDM in this embodiment includes:

The dividing module is used to divide the medium and large indoor plane into multiple connected sub-areas, and multiple LED devices are allocated to each sub-region, and the multiple LED devices are respectively fixed on the vertices of the sub-regions and are adjacent to each other. Share LED equipment in the sub-areas of, where the LED equipment is provided with identity information;

The processing module is used to perform quadrature amplitude modulation on the binary data stream that needs to be sent, and then place the communication data and the identity information data on their assigned subcarriers respectively, and perform STBC encoding on the communication data;

The modulation module performs DC bias optical orthogonal frequency division multiplexing modulation on the data after the above processing;

The sending module is used to send the respective data on multiple LED devices in the same sub-area in the form of optical signals, and the optical signals are sent to the receiver after being transmitted in free space, and received by the receiver.

Example four

Based on the same inventive concept, this embodiment provides a second indoor visible light communication positioning system based on machine learning and OFDM. The principle of solving the problem is similar to the indoor visible light communication positioning method based on machine learning and OFDM described in the second embodiment, repeating I won't repeat it here.

The data processing module is used for the receiving end to process the received data;

The coordinate confirmation module is used to separate communication data and identity information data. For identity information data, determine the subarea where the receiver is located according to the recovered identity information data; for communication data, extract pilot data and calculate channel estimation matrix to extract feature values Then, use the machine learning model to locate, obtain the relative coordinates of the receiver, and use the channel estimation matrix for STBC decoding to obtain the communication data; combine the obtained relative coordinates with the identity information data obtained by the receiver to calculate the final coordinates of the receiver .

Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, this application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

This application is described with reference to flowcharts and/or block diagrams of methods, devices (systems), and computer program products according to embodiments of this application. It should be understood that each process and/or block in the flowchart and/or block diagram, and the combination of processes and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing equipment to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing equipment are used to generate It is a device that realizes the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.

These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing equipment to work in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction device. The device implements the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.

These computer program instructions can also be loaded on a computer or other programmable data processing equipment, so that a series of operation steps are executed on the computer or other programmable equipment to produce computer-implemented processing, so as to execute on the computer or other programmable equipment. The instructions provide steps for implementing the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.

Obviously, the above-mentioned embodiments are merely examples for clear description, and do not limit the implementation manners. For those of ordinary skill in the art, other changes or changes in different forms can be made on the basis of the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or changes derived from this are still within the protection scope created by the present invention.

Claims

An indoor visible light communication positioning method based on machine learning and OFDM is characterized in that it includes the following steps:

Step S1: Divide the medium and large indoor plane into multiple connected sub-areas, and allocate multiple LED devices to each sub-area, and the multiple LED devices are respectively fixed on the vertices of the sub-areas, and adjacent sub-regions Regional shared LED equipment, where the LED equipment is provided with identity information;

Step S2: Perform quadrature amplitude modulation on the binary data stream that needs to be sent, and then place the communication data and the identity information data on their respective allocated subcarriers, and perform STBC encoding on the communication data;

Step S3: Perform DC bias optical orthogonal frequency division multiplexing modulation on the data after the above processing;

Step S4: Send the respective data on the multiple LED devices in the same sub-area in the form of optical signals, and the optical signals are transmitted to the receiver after being transmitted in free space, and received by the receiver.
The indoor visible light communication positioning method based on machine learning and OFDM according to claim 1, wherein the size of the sub-region is based on the power of the LED device, the viewing angle of the receiver, the illuminance requirement, and the size of the indoor space. Wait for physical conditions to make adjustments.
The indoor visible light communication positioning method based on machine learning and OFDM according to claim 1, characterized in that: when the complex signal is subjected to STBC encoding, the original data content sent by each LED device in the sub-area must be the same.
The indoor visible light communication positioning method based on machine learning and OFDM according to claim 1 or 3, characterized in that: before STBC encoding the complex signal, communication data and identity information data are allocated to different frequency bands, wherein Because the communication data has a larger proportion in the data packet, it is placed on a continuous segment of carrier.
The indoor visible light communication positioning method based on machine learning and OFDM according to claim 1, wherein the frequency bands occupied by the identity information sub-carriers allocated by the plurality of LED devices are different.
An indoor visible light communication positioning method based on machine learning and OFDM is characterized in that it includes the following steps:

Step S1: The receiving end processes the received data;

Step S2: Separate the communication data and the identity information data. For the identity information data, determine the subarea where the receiver is located according to the recovered identity information data; for the communication data, extract the pilot data and calculate the channel estimation matrix, after extracting the eigenvalues, use Machine learning model positioning, the receiver's relative coordinates are obtained, and the channel estimation matrix is used for STBC decoding to obtain communication data; the obtained relative coordinates and the identity information data obtained by the receiver are combined to calculate the receiver's final coordinates.
The indoor visible light communication positioning method based on machine learning and OFDM according to claim 6, wherein the method for the receiving end to process the received data is: DC bias optical orthogonal frequency division multiplexing The time domain signal is converted into a frequency domain signal, and then the frequency domain channel equalization technology is used for channel estimation, the data of the identity information is equalized, and the communication data is decoded. After the amplitude demodulation, the binary data stream is output.
The indoor visible light communication positioning method based on machine learning and OFDM according to claim 6, characterized in that: the method of positioning using a machine learning model is: collecting training samples during offline training; extracting pilot data and calculating In the channel estimation matrix, after extracting the eigenvalues, the eigenvalues are used as input parameters of the machine learning model, and machine learning model training is used to obtain machine learning model positioning.
The indoor visible light communication positioning method based on machine learning and OFDM according to claim 8, characterized in that: after the machine learning model training is completed, it is stored in the local memory of the receiver or in the cloud, and the network data is used for positioning. The input parameters are uploaded to the cloud, the cloud server is used to calculate the coordinates, and the cloud server calculates the coordinates and then downloads them to the lower computer.
An indoor visible light communication positioning system based on machine learning and OFDM, which is characterized in that it includes:

The dividing module is used to divide the medium and large indoor plane into multiple connected sub-areas, and multiple LED devices are allocated to each sub-region, and the multiple LED devices are respectively fixed on the vertices of the sub-regions and are adjacent to each other. Share LED equipment in the sub-areas of, where the LED equipment is provided with identity information;

The processing module is used to perform quadrature amplitude modulation on the binary data stream that needs to be sent, and then place the communication data and the identity information data on their assigned subcarriers respectively, and perform STBC encoding on the communication data;

The modulation module performs DC bias optical orthogonal frequency division multiplexing modulation on the data after the above processing;

The sending module is used to send respective data on multiple LED devices in the same sub-area in the form of optical signals, and the optical signals are sent to the receiver after being transmitted in free space, and then received by the receiver.