A method and apparatus for LED positioning based on visible light communication

Abstract

The application discloses an LED positioning method and device based on visible light communication, comprising: mapping the model of each LED with the position information of the position where the LED is located; obtaining the visible light signal of the LED to be positioned, and extracting the target signal feature of the visible light signal; determining the target model of the LED to be positioned based on the target signal feature; and searching for the target position information matched with the target model in the mapping relationship, wherein the target position information is the position information of the LED to be positioned. In the above process, the positioning process of the LED to be positioned is converted into the identification process of the model of the LED to be positioned, the positioning efficiency is improved, and the problem that the existing positioning mode needs to allocate an ID for each LED, the LED needs to repeatedly send the ID information while performing visible light communication, and the terminal needs to distinguish the ID information and the communication information after receiving the signal, thereby causing a long time delay and greatly affecting the communication efficiency is avoided.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an LED positioning method and apparatus based on visible light communication. Background Technology

[0002] The Industrial Internet of Things (IIoT) is an evolution of the Internet of Things (IoT). It emphasizes not only the absence of human intervention in industrial production processes but also the autonomous nature of machines. IIoT communication for intelligent manufacturing workshops is mainly reflected in real-time task assignment, end-to-end production process control, and precise management of production materials. Based on these characteristics of industrial manufacturing, the communication needs of industrial manufacturing workshops can be summarized as requiring both real-time positioning and high-speed, high-density communication. With the increasing number of intelligent devices participating in industrial production workshops, radio spectrum is becoming increasingly scarce, making it difficult to meet the high-speed, high-density communication requirements of IIoT. Visible light, due to its unlimited bandwidth, inherent safety, controllable radiation range, and customizability, has become a new choice for indoor communication and positioning in the Industrial Internet of Things.

[0003] Currently, visible light indoor communication and positioning are generally studied separately, with complex equipment and related algorithms, posing further challenges to the already complex intelligent industrial production workshops. Currently, positioning under different LED lights mainly relies on LED-ID. This positioning method requires assigning an ID to each LED, and the LED must continuously and repeatedly transmit its ID information while conducting visible light communication. The terminal, upon receiving the signal, must distinguish between the ID information and the communication information. This inevitably results in a significant time delay, greatly impacting communication efficiency. Summary of the Invention

[0004] In view of this, the present invention provides an LED positioning method and apparatus based on visible light communication to solve the problem that existing positioning methods require assigning an ID to each LED, and the LED must repeatedly transmit its ID information while performing visible light communication. After receiving the signal, the terminal needs to distinguish between the ID information and the communication information, resulting in a long delay and greatly affecting communication efficiency. The specific solution is as follows:

[0005] An LED positioning method based on visible light communication includes:

[0006] A mapping relationship is established in advance between the model number of each LED and its location information;

[0007] Acquire the visible light signal of the LED to be located, and extract the target signal features of the visible light signal;

[0008] The target model of the LED to be located is determined based on the target signal characteristics.

[0009] The target location information that matches the target model is found in the mapping relationship, wherein the target location information is the location information of the LED to be located.

[0010] Optionally, the above method involves acquiring the visible light signal of the LED to be located and extracting the target signal features of the visible light signal, including:

[0011] Acquire the visible light signal collected by the photodiode;

[0012] The visible light signal is passed to a wavelet scattering network to obtain the target signal characteristics.

[0013] Optionally, the above method involves transmitting the visible light signal to a wavelet scattering network to obtain target signal features, including:

[0014] The visible light signal is processed by wavelet convolution to obtain the first signal feature;

[0015] The first signal feature is subjected to nonlinear processing to obtain the second signal feature;

[0016] The second signal feature is averaged to obtain the target signal feature.

[0017] Optionally, the method described above, which determines the target model of the LED to be located based on the target signal characteristics, includes:

[0018] The target signal features are passed to a preset classifier to obtain the target model of the LED to be located, wherein the preset classifier is trained using the signal features as samples and the LED model as labels.

[0019] Optionally, when the preset classifier is a Long Short-Term Memory (LSTM) network model, the training process of the preset classifier includes:

[0020] Obtain a training dataset, wherein the training dataset includes: multiple training samples, each training sample including: signal features and model label of each LED;

[0021] The training dataset is divided into a test training dataset and a validation training dataset;

[0022] The long short-term memory network model is trained based on the training samples in the test training dataset to obtain the predicted model label;

[0023] If the loss values ​​of the model label and the predicted model label meet the preset loss value threshold, the long short-term memory network model is validated based on the training samples in the validation training dataset to obtain the preset classifier.

[0024] An LED positioning device based on visible light communication includes:

[0025] A module is established to pre-map the model of each LED to its location information;

[0026] The acquisition and extraction module is used to acquire the visible light signal of the LED to be located and extract the target signal features of the visible light signal;

[0027] The determination module is used to determine the target model of the LED to be located based on the target signal characteristics;

[0028] The lookup module is used to find target location information that matches the target model in the mapping relationship, wherein the target location information is the location information of the LED to be located.

[0029] Optionally, the acquisition and extraction module in the aforementioned apparatus includes:

[0030] Acquisition unit, used to acquire visible light signals collected by photodiode;

[0031] An extraction unit is used to transmit the visible light signal to a wavelet scattering network to obtain target signal features.

[0032] Optionally, in the aforementioned apparatus, the extraction unit includes:

[0033] A convolution processing subunit is used to perform wavelet convolution processing on the visible light signal to obtain the first signal feature;

[0034] A nonlinear processing subunit is used to perform nonlinear processing on the first signal feature to obtain a second signal feature;

[0035] An averaging subunit is used to perform averaging on the second signal features to obtain the target signal features.

[0036] Optionally, in the aforementioned apparatus, the determining module includes:

[0037] The transmission unit is used to transmit the target signal features to a preset classifier to obtain the target model of the LED to be located, wherein the preset classifier is trained using the signal features as samples and the LED model as labels.

[0038] Optionally, in the above-described apparatus, when the preset classifier is a Long Short-Term Memory (LSTM) network model, the training process of the preset classifier in the transmission unit includes:

[0039] Obtain a training dataset, wherein the training dataset includes: multiple training samples, each training sample including: signal features and model label of each LED;

[0040] The training dataset is divided into a test training dataset and a validation training dataset;

[0041] The long short-term memory network model is trained based on the training samples in the test training dataset to obtain the predicted model label;

[0042] If the loss values ​​of the model label and the predicted model label meet the preset loss value threshold, the long short-term memory network model is validated based on the training samples in the validation training dataset to obtain the preset classifier.

[0043] Compared with the prior art, the present invention has the following advantages:

[0044] This invention discloses an LED positioning method and apparatus based on visible light communication, comprising: pre-establishing a mapping relationship between the model of each LED and its location information; acquiring the visible light signal of the LED to be positioned and extracting the target signal features of the visible light signal; determining the target model of the LED to be positioned based on the target signal features; and searching for target location information matching the target model in the mapping relationship, wherein the target location information is the location information of the LED to be positioned. In the above process, the positioning process of the LED to be positioned is transformed into the identification process of the LED model, improving positioning efficiency and avoiding the problem of existing positioning methods requiring the assignment of an ID to each LED, and the LED repeatedly sending its ID information while performing visible light communication. The terminal, upon receiving the signal, must distinguish between the ID information and the communication information, resulting in a long delay and significantly impacting communication efficiency. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a flowchart of an LED positioning method based on visible light communication disclosed in an embodiment of the present invention;

[0047] Figure 2 This is a schematic diagram of a wavelet scattering network structure disclosed in an embodiment of the present invention;

[0048] Figure 3A schematic diagram illustrating a classification and identification method based on different LED models, provided as an embodiment of the present invention;

[0049] Figure 4 This is a schematic diagram showing the relative positions of an LED and a PD, provided in an embodiment of the present invention.

[0050] Figure 5 This is a structural block diagram of an LED positioning device based on visible light communication disclosed in an embodiment of the present invention. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] This invention discloses an LED positioning method and apparatus based on visible light communication, applied to the positioning process of LEDs. Existing technologies for positioning different LEDs mainly rely on LED-ID. This positioning method requires assigning an ID to each LED, and the LED must continuously and repeatedly transmit its ID information while conducting visible light communication. After receiving the signal, the terminal needs to distinguish between the ID information and the communication information. The methods for distinguishing ID information and communication information vary. For example, the source and terminal may reach a protocol to transmit a fixed string of encoded information as a key before transmitting ID information. Upon receiving the key, the mobile terminal knows that the next step is to receive ID information. Although the methods differ, they all affect normal communication, causing delays and significantly impacting communication efficiency. Research has found that subtle differences in the materials and processes used to manufacture different LED models result in stable, subtle differences in the emitted light signals. These differences can be used as identification information for different LED models. These differences may manifest in waveform rising edge shape, falling edge shape, trailing edge, and other shape parameters. By extracting this identification information, different LED models can be classified. In this way, in smart industrial production workshops using visible light communication, it is only necessary to deploy LEDs of different models. The different models of LEDs are identified by the light signals emitted by them, and the location information associated with the LEDs is then used to achieve visible light positioning, transforming the visible light positioning problem into the problem of identifying different LED models. Therefore, this invention provides an LED positioning method based on visible light communication, the execution flow of which is as follows: Figure 1 As shown, the steps include:

[0053] S101. Establish a mapping relationship between the model of each LED and its location information in advance;

[0054] In this embodiment of the invention, the model and its corresponding location information of each LED in the current area are obtained in advance, wherein the model and location information are known in advance, and a mapping relationship is established between the model and its corresponding location information of each LED.

[0055] S102. Acquire the visible light signal of the LED to be located, and extract the target signal features of the visible light signal;

[0056] In this embodiment of the invention, the visible light signal of the LED to be located is obtained based on a photodiode. The photodiode is used to convert the visible light signal of the LED to be located into an electrical signal. Preferably, the visible light signal is displayed and saved using an oscilloscope.

[0057] Furthermore, the visible light signal is passed to a wavelet scattering network to obtain target signal features. The wavelet scattering network is used to extract waveform features of the signal received by the mobile terminal (oscilloscope). Analogous to deep neural networks, it is an automatic feature extractor well-suited for nonlinear and non-stationary signals. Each layer consists of three parts: wavelet convolution, nonlinearity, and averaging. Unlike deep learning networks, the weights of the wavelet scattering network filters are fixed and do not require continuous adjustment based on network feedback; therefore, it does not require a large number of training samples. Simultaneously, the features extracted by the wavelet scattering network can significantly reduce the feature dimensionality of the original signal with almost no loss of information, greatly reducing the complexity of the classifier and training time. The specific processing procedure for passing the visible light signal to the wavelet scattering network to obtain target signal features is as follows:

[0058] The complex bandpass filter ψ is obtained by scaling and rotating the mother wavelet function ψ. λ ,2 j The scaling factor, θ, is the rotation factor, i.e.:

[0059] ψ λ =2 2j ψ(2 j θ -1 x), (1)

[0060] λ=2 j θj∈Z, 0≤j<J, θ=kπ / K, k=0,1,...,K-1.

[0061] The visible light signal is processed by wavelet convolution based on formula (1) to obtain the first signal feature. F is the visible light signal, and its wavelet mode coefficients are as follows:

[0062] U[λ]F=|F*ψ λ (x)|, (2)

[0063] Where U[λ]F represents the first characteristic signal.

[0064] Then, based on formula (2), the first feature signal is subjected to nonlinear processing to obtain the second feature signal, wherein,

[0065] The wavelet mode coefficients are cascaded to form the scattering propagator U[p]:

[0066]

[0067] Where p=(λ1,λ2,...λ m ) represents a frequency-decreasing path, i.e., |λ k |≥|λ k+1 |, k=1,2,...,m-1, m is the layer number, represents the intermediate value of the m-th layer, and U[p]F represents the second characteristic signal.

[0068] Finally, the second signal feature is averaged to obtain the target signal feature, wherein the scattering operator S j In a width and 2 j Spatial averaging is performed within proportional regions:

[0069]

[0070] φ J (x) represents the scaling filter function.

[0071] The network nodes at layer m correspond to all paths p = (λ1, λ2, ..., λ) of length m. m The set p) m The m-th layer stores the propagation signal {U[p]F}p∈P. m Output scattering coefficient (target characteristic signal) {S J [p]F}p∈P m .

[0072] In this embodiment of the invention, the principle block diagram of the wavelet scattering network is as follows: Figure 2 As shown, the input signal F is processed by a scaling function, a complex wavelet filter, modulo operations, and convolution operations to obtain intermediate values, coefficient vectors, and output values. Among these, the scattering coefficients and the feature matrix are equivalent.

[0073] If the wavelet scattering network has only 0 layers (m = 0), the coefficient matrix obtained by wavelet transforming the F signal is:

[0074] F*φ J (x)

[0075] If the wavelet scattering network has only one layer, i.e., m=1, the coefficient matrix obtained by wavelet transforming the F signal is:

[0076]

[0077] If the wavelet scattering network has only 2 layers, i.e., m=2, the coefficient matrix obtained by wavelet transforming the F signal is:

[0078]

[0079] Finally, m can have any number of layers, resulting in the following coefficient matrix:

[0080]

[0081] S103. Determine the target model of the LED to be located based on the target signal characteristics;

[0082] In this embodiment of the invention, a pre-selected classifier is used for training. Preferably, the pre-selected classifier is a Long Short-Term Memory (LSTM) network model, which is a 5-layer LSTM network including a sequence input layer, an LSTM layer, a fully connected layer, a softmax layer, and a classification output layer. The LSTM layer contains 300 neurons. The relevant parameters in the LSTM network are set as follows: optimizer is set to admin, maximum training epochs are 600, minimum stride is 330, gradient threshold is 1, and learning rate is 0.001. The LSTM network uses the feature matrix of the signal extracted by the wavelet scattering network to classify the emission signals of different LED models, thereby recognizing the LED. The input of the model is the target signal feature, and the output is the LED model. Based on the LED model, the specific LED can be determined, and then its location can be determined based on the previously stored location information.

[0083] The preset classifier is trained based on the above parameters. The specific training process includes: acquiring a training dataset, wherein the training dataset includes multiple training samples, each training sample including signal features and model labels for each LED, wherein the model labels and signal features are known; dividing the training dataset into a test training dataset and a validation training dataset, wherein the division ratio of the test training dataset and the validation training dataset can be determined based on experience or specific application scenarios, and is not specifically limited in this embodiment of the invention; training the Long Short-Term Memory (LSTM) network model based on the training samples in the test training dataset to obtain predicted model labels; and validating the LSM network model based on the training samples in the validation training dataset when the loss values ​​of the model labels and the predicted model labels meet a preset loss threshold, thereby obtaining the preset classifier.

[0084] To illustrate the above process, for example, an experimental system is first constructed, as shown in the example below. Figure 3As shown, the experimental apparatus mainly includes the following modules: six different types of LEDs containing the same modulation signal, an oscilloscope, and a photodiode (PD). The models and specific parameters of the experimental apparatus are shown in Tables 1 and 2. Signal transmitting end: LEDs emit optical signals for communication (LED1, LED2, LED3, LED4, LED5, and LED6). Receiving end: The PD converts the received optical signals into electrical signals, which are then input into the oscilloscope for waveform display, facilitating data storage. The waveform features extracted from the oscilloscope are input into a classifier. The classifier identifies the LEDs based on the received signal features and finally locates them based on the stored LED-related location information.

[0085] Furthermore, the relative positions of the LED and the mobile terminal or PD are as follows: Figure 4 As shown, the LED is located 1.8m directly above the center of L1, which is a circle with a radius of 0.3m. L2, L3, L4, and L5 are each annulus with a width of 0.3m. Pre-collected data serves as the training dataset. 80% of the training dataset is used as the test dataset and input into the Long Short-Term Memory network to train the model. 20% of the training dataset is used as the validation test dataset to test the classification performance of the trained model. Preferably, additional samples are collected simultaneously within the regions of L1, L2, L3, L4, and L5 for each LED model to test the model's generalization performance.

[0086] led model Power (W) LED1 NVC 12 LED2 NVC 18 LED3 NVC 24 LED4 PHILIPS 12 LED5 PHILIPS 18 LED6 PHILIPS 24

[0087] Table 1 LED Models and Power

[0088] parameter value LED bandwidth 1-20MHz LED output signal frequency 3MHZ

[0089] Table 2 Parameters of LED Experimental Modules

[0090] Furthermore, after training is completed, the target signal features are passed to the trained preset classifier to obtain the target model of the LED to be located.

[0091] S104. Search for target location information that matches the target model in the mapping relationship, wherein the target location information is the location information of the LED to be located.

[0092] In this embodiment of the invention, each mapping relationship is traversed to find target location information matching the target model, wherein the target location information is the location information of the LED to be located. Preferably, the parameters in the preset classifier are continuously optimized based on the actual positioning results.

[0093] Furthermore, during the testing phase, the real-time collected LED emission signals are input into a pre-trained classifier for testing, enabling the identification of different LED models and thus achieving real-time positioning.

[0094] This invention discloses an LED positioning method based on visible light communication, comprising: pre-establishing a mapping relationship between the model of each LED and its location information; acquiring the visible light signal of the LED to be positioned and extracting the target signal features of the visible light signal; determining the target model of the LED to be positioned based on the target signal features; and searching for target location information matching the target model in the mapping relationship, wherein the target location information is the location information of the LED to be positioned. In the above process, the positioning process of the LED to be positioned is transformed into the identification process of the LED model, improving positioning efficiency and avoiding the problem of existing positioning methods requiring the assignment of an ID to each LED. Furthermore, the LED must repeatedly send its ID information while performing visible light communication, and the terminal must distinguish between the ID information and communication information after receiving the signal, resulting in a long delay and significantly impacting communication efficiency.

[0095] In this invention, the LED positioning problem is transformed into the identification of different LED models. By classifying the optical signals received by the terminal for communication, the identification of different LED models is achieved, ultimately leading to positioning. The key technology for terminal positioning through LED model identification lies in utilizing the differences in emitted signals caused by subtle differences in materials and manufacturing processes among different LED models as their identification information. This approach, starting from the inherent characteristics of LEDs, studies the visible light positioning problem from a new perspective. This invention does not require any additional equipment and performs positioning simultaneously with visible light communication, achieving terminal positioning at a lower cost. Furthermore, a wavelet scattering and long short-term memory combined network is used to identify different LED models through the emitted optical signals. Supervised learning is performed using features continuously extracted during the training phase. The trained optimal model still maintains a high classification accuracy for different areas under the LED light, demonstrating good robustness and applicability.

[0096] Based on the above-described LED positioning method using visible light communication, this embodiment of the invention also provides an LED positioning device based on visible light communication, the structural block diagram of which is shown below. Figure 5 As shown, it includes:

[0097] The module consists of module 201 (establishment), module 202 (acquisition and extraction), module 203 (determination), and module 204 (search).

[0098] in,

[0099] The establishment module 201 is used to pre-establish a mapping relationship between the model of each LED and its location information;

[0100] The acquisition and extraction module 202 is used to acquire the visible light signal of the LED to be located and extract the target signal features of the visible light signal;

[0101] The determining module 203 is used to determine the target model of the LED to be located based on the target signal characteristics;

[0102] The search module 204 is used to search for target location information that matches the target model in the mapping relationship, wherein the target location information is the location information of the LED to be located.

[0103] This invention discloses an LED positioning device based on visible light communication, comprising: pre-establishing a mapping relationship between the model of each LED and its location information; acquiring the visible light signal of the LED to be positioned and extracting the target signal features of the visible light signal; determining the target model of the LED to be positioned based on the target signal features; and searching for target location information matching the target model in the mapping relationship, wherein the target location information is the location information of the LED to be positioned. In the above process, the positioning process of the LED to be positioned is transformed into the identification process of the LED model, improving positioning efficiency and avoiding the problem of existing positioning methods requiring the assignment of an ID to each LED, and the LED repeatedly sending its ID information while performing visible light communication. The terminal, upon receiving the signal, must distinguish between the ID information and the communication information, resulting in a long delay and significantly impacting communication efficiency.

[0104] In this embodiment of the invention, the acquisition and extraction module 202 includes:

[0105] Acquisition unit 205 and extraction unit 206.

[0106] in,

[0107] The acquisition unit 205 is used to acquire the visible light signal collected by the photodiode;

[0108] The extraction unit 206 is used to transmit the visible light signal to the wavelet scattering network to obtain the target signal features.

[0109] In this embodiment of the invention, the extraction unit 206 includes:

[0110] Convolution processing subunit 207, nonlinear processing subunit 208, and averaging processing subunit 209.

[0111] in,

[0112] The convolution processing subunit 207 is used to perform wavelet convolution processing on the visible light signal to obtain the first signal feature;

[0113] The nonlinear processing subunit 208 is used to perform nonlinear processing on the first signal feature to obtain the second signal feature;

[0114] The averaging subunit 209 is used to perform averaging processing on the second signal features to obtain the target signal features.

[0115] In this embodiment of the invention, the determining module 203 includes a transmission unit 210.

[0116] in,

[0117] The transmission unit 210 is used to transmit the target signal features to a preset classifier to obtain the target model of the LED to be located, wherein the preset classifier is trained using the signal features as samples and the LED model as a label.

[0118] In this embodiment of the invention, when the preset classifier is a Long Short-Term Memory (LSTM) network model, the training process of the preset classifier in the transmission unit 210 includes:

[0119] Obtain a training dataset, wherein the training dataset includes: multiple training samples, each training sample including: signal features and model label of each LED;

[0120] The training dataset is divided into a test training dataset and a validation training dataset;

[0121] The long short-term memory network model is trained based on the training samples in the test training dataset to obtain the predicted model label;

[0122] If the loss values ​​of the model label and the predicted model label meet the preset loss value threshold, the long short-term memory network model is validated based on the training samples in the validation training dataset to obtain the preset classifier.

[0123] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0124] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0125] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0126] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0127] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0128] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0129] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0130] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0131] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0132] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. An LED positioning method based on visible light communication, characterized in that, include: A mapping relationship is established in advance between the model number of each LED and its location information; Acquire the visible light signal of the LED to be located, and extract the target signal features of the visible light signal; The target model of the LED to be located is determined based on the target signal characteristics. The target location information that matches the target model is found in the mapping relationship, wherein the target location information is the location information of the LED to be located.

2. The method according to claim 1, characterized in that, Acquire the visible light signal of the LED to be located, and extract the target signal features of the visible light signal, including: Acquire the visible light signal collected by the photodiode; The visible light signal is passed to a wavelet scattering network to obtain the target signal characteristics.

3. The method according to claim 2, characterized in that, The visible light signal is passed to a wavelet scattering network to obtain target signal features, including: The visible light signal is processed by wavelet convolution to obtain the first signal feature; The first signal feature is subjected to nonlinear processing to obtain the second signal feature; The second signal feature is averaged to obtain the target signal feature.

4. The method according to claim 1, characterized in that, Determining the target model of the LED to be located based on the target signal characteristics includes: The target signal features are passed to a preset classifier to obtain the target model of the LED to be located, wherein the preset classifier is trained using the signal features as samples and the LED model as labels.

5. The method according to claim 4, characterized in that, When the preset classifier is a Long Short-Term Memory (LSTM) network model, the training process of the preset classifier includes: Obtain a training dataset, wherein the training dataset includes: multiple training samples, each training sample including: signal features and model label of each LED; The training dataset is divided into a test training dataset and a validation training dataset; The long short-term memory network model is trained based on the training samples in the test training dataset to obtain the predicted model label; If the loss values ​​of the model label and the predicted model label meet the preset loss value threshold, the long short-term memory network model is validated based on the training samples in the validation training dataset to obtain the preset classifier.

6. An LED positioning device based on visible light communication, characterized in that, include: A module is established to pre-map the model of each LED to its location information; The acquisition and extraction module is used to acquire the visible light signal of the LED to be located and extract the target signal features of the visible light signal; The determination module is used to determine the target model of the LED to be located based on the target signal characteristics; The lookup module is used to find target location information that matches the target model in the mapping relationship, wherein the target location information is the location information of the LED to be located.

7. The apparatus according to claim 6, characterized in that, The acquisition and extraction module includes: Acquisition unit, used to acquire visible light signals collected by photodiode; An extraction unit is used to transmit the visible light signal to a wavelet scattering network to obtain target signal features.

8. The apparatus according to claim 7, characterized in that, The extraction unit includes: A convolution processing subunit is used to perform wavelet convolution processing on the visible light signal to obtain the first signal feature; A nonlinear processing subunit is used to perform nonlinear processing on the first signal feature to obtain a second signal feature; An averaging subunit is used to perform averaging on the second signal features to obtain the target signal features.

9. The apparatus according to claim 6, characterized in that, The determining module includes: The transmission unit is used to transmit the target signal features to a preset classifier to obtain the target model of the LED to be located, wherein the preset classifier is trained using the signal features as samples and the LED model as labels.

10. The apparatus according to claim 9, characterized in that, In the case that the preset classifier is a Long Short-Term Memory (LSTM) network model, the training process of the preset classifier in the transmission unit includes: Obtain a training dataset, wherein the training dataset includes: multiple training samples, each training sample including: signal features and model label of each LED; The training dataset is divided into a test training dataset and a validation training dataset; The long short-term memory network model is trained based on the training samples in the test training dataset to obtain the predicted model label; If the loss values ​​of the model label and the predicted model label meet the preset loss value threshold, the long short-term memory network model is validated based on the training samples in the validation training dataset to obtain the preset classifier.

Images