Model training method, data processing method, vehicle and storage medium

The method improves Bluetooth positioning accuracy in new energy vehicles by training a convolutional neural network model with data from a target device, addressing low precision and user experience issues, ensuring precise vehicle control and safety.

CN119067192APending Publication Date: 2024-12-03GUANGZHOU XIAOPENG MOTORS TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411054851.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-01
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

Existing Bluetooth positioning methods in new energy vehicles suffer from low accuracy and poor user experience due to meter-level precision, leading to potential misoperations and safety risks, especially in environments with long distances or signal interference.

Method used

A model training method using a target device equipped with a communication component and ultra-wideband component to collect data from multiple directions, forming a target data set that includes static and dynamic data, which is used to train an initial convolutional neural network model to improve positioning accuracy and guide vehicle control actions.

Benefits of technology

Enhances positioning accuracy and user experience by enabling centimeter-level precision and stable vehicle control, reducing the likelihood of accidental door openings or closings, thus enhancing safety and reducing property loss.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119067192A_ABST
    Figure CN119067192A_ABST
Patent Text Reader

Abstract

The invention discloses a model training method, a data processing method, a vehicle and a storage medium. The method comprises the following steps: performing data acquisition from a plurality of directions around a target vehicle by using target equipment to obtain a target data set; and training the initial convolutional neural network model based on the target data set to obtain a target convolutional neural network model. According to the invention, the technical problems of low positioning precision and poor user experience of a Bluetooth positioning mode provided in the related technology are solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of vehicle technology, and in particular to a model training method, a data processing method, a vehicle and a storage medium. Background Art

[0002] As the intelligence of cars becomes more and more popular, most new energy vehicles are now equipped with mobile phone keys, which can realize the non-sensing opening and closing of car doors, such as automatic unlocking when the mobile phone is close to the vehicle, automatic locking when away from the vehicle, and the effect of welcome lighting when the user approaches the vehicle. However, in the relevant technology, since the Bluetooth positioning accuracy can only be achieved at the meter level, in the process of developing the Bluetooth positioning algorithm, the calibration data needs to be collected manually or by machine, and then the fingerprint confirmation positioning is performed. As a result, it is impossible to accurately identify the distance between the user and the vehicle, which is easy to cause misoperation. In the case of long distance or signal interference, the Bluetooth positioning signal may be abnormal, causing the car door to open or close accidentally, thereby increasing the possibility of safety risks and property loss.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present application provide a model training method, a data processing method, a vehicle and a storage medium to at least solve the technical problems of low positioning accuracy and poor user experience existing in the Bluetooth positioning method provided in the related art.

[0005] According to one embodiment of the present application, a model training method is provided, comprising: using a target device to collect data from multiple directions around a target vehicle to obtain a target data set, wherein the target device is configured with a communication component and an ultra-wideband component, the communication component is used to obtain the signal strength between the target device and the target vehicle, the ultra-wideband component is used to obtain the distance between the target device and the target vehicle, and the target data set includes: static data and dynamic data, the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state; based on the target data set, an initial convolutional neural network model is trained to obtain a target convolutional neural network model, wherein the target convolutional neural network model is used to guide a domain controller of the target vehicle to determine a control action to be executed.

[0006] According to one embodiment of the present application, a data processing method is provided, including: obtaining a current signal strength between a target device and a target vehicle; using a target convolutional neural network model to perform distance prediction on the current signal strength to obtain a distance prediction result; determining a target control action to be executed by a domain controller of the target vehicle based on the distance prediction result; wherein the target convolutional neural network model is obtained by training an initial convolutional neural network model based on a target data set, the target data set is obtained by collecting data from multiple directions around the target vehicle using a target device, a communication component and an ultra-wideband component are configured on the target device, the communication component is used to obtain the signal strength between the target device and the target vehicle, the ultra-wideband component is used to obtain the distance between the target device and the target vehicle, the target data set includes: static data and dynamic data, the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state.

[0007] According to one embodiment of the present application, a model training device is provided, comprising: an acquisition module, for acquiring data from multiple directions around a target vehicle using a target device to obtain a target data set, wherein the target device is configured with a communication component and an ultra-wideband component, the communication component is used to obtain the signal strength between the target device and the target vehicle, the ultra-wideband component is used to obtain the distance between the target device and the target vehicle, the target data set comprises: static data and dynamic data, the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state; a training module, for training an initial convolutional neural network model based on the target data set to obtain a target convolutional neural network model, wherein the target convolutional neural network model is used to guide a domain controller of the target vehicle to determine a control action to be executed.

[0008] According to one embodiment of the present application, a data processing device is provided, including: an acquisition module, used to acquire the current signal strength between a target device and a target vehicle; a prediction module, used to use a target convolutional neural network model to perform distance prediction on the current signal strength to obtain a distance prediction result; a processing module, used to determine a target control action to be executed by a domain controller of the target vehicle based on the distance prediction result; wherein, the target convolutional neural network model is obtained by training an initial convolutional neural network model based on a target data set, the target data set is obtained by collecting data from multiple directions around the target vehicle using a target device, a communication component and an ultra-wideband component are configured on the target device, the communication component is used to acquire the signal strength between the target device and the target vehicle, the ultra-wideband component is used to acquire the distance between the target device and the target vehicle, and the target data set includes: static data and dynamic data, the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state.

[0009] According to one embodiment of the present application, a vehicle is provided, comprising: a memory storing an executable program; and a processor for running the program, wherein when the program is running, any one of the model training methods or any one of the data processing methods in the various embodiments of the present application is executed.

[0010] According to one embodiment of the present application, a computer-readable storage medium is provided, and the computer-readable storage medium includes a stored executable program, wherein when the executable program is running, the device where the storage medium is located is controlled to execute any one of the model training methods or any one of the data processing methods in the various embodiments of the present application.

[0011] According to one embodiment of the present application, a computer program product is also provided, including a computer program, which, when executed by a processor, implements any one of the model training methods or any one of the data processing methods in the various embodiments of the present application.

[0012] According to one embodiment of the present application, a computer program product is also provided, including a non-volatile computer-readable storage medium, the non-volatile computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, any model training method or any data processing method in the various embodiments of the present application is implemented.

[0013] According to one embodiment of the present application, a computer program is also provided. When the computer program is executed by a processor, it implements any one of the model training methods or any one of the data processing methods in the various embodiments of the present application.

[0014] In an embodiment of the present application, a target device equipped with a communication component and an ultra-wideband component is used to collect data from multiple directions around the target vehicle to obtain a target data set containing static data and dynamic data. Based on the target data set, an initial convolutional neural network model is trained to obtain a target convolutional neural network model, thereby guiding the domain controller of the target vehicle to determine the control action to be executed, thereby achieving the purpose of efficiently and accurately controlling the target vehicle, thereby achieving the technical effect of improving the positioning accuracy of the target device and the user experience, and thus solving the technical problems of low positioning accuracy and poor user experience in the Bluetooth positioning method provided in the relevant technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0016] Figure 1 is a flow chart of a model training method according to one embodiment of the present application;

[0017] Figure 2 is a schematic diagram of a data collection area according to one embodiment of the present application;

[0018] Figure 3 is a schematic diagram of a model training process according to one embodiment of the present application;

[0019] Figure 4 is a flow chart of a data processing method according to one embodiment of the present application;

[0020] Figure 5 is a schematic diagram of a data processing method according to one embodiment of the present application;

[0021] Figure 6 is a structural block diagram of a model training device according to one embodiment of the present application;

[0022] Figure 7 It is a structural block diagram of a data processing device according to one embodiment of the present application. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0024] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0025] According to an embodiment of the present application, a method embodiment of a model training method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0026] The method embodiment can be executed in an electronic device or a similar computing device including a memory and a processor. Taking running on a computer terminal as an example, the computer terminal may include one or more processors (processors may include but are not limited to central processing units (CPU), graphics processing units (GPU), digital signal processing (DSP) chips, microprocessors (MCU), programmable logic devices (Field Programmable Gate Array, FPGA), neural network processors (Neural-network Processor Unit, NPU), tensor processors (TensorProcessing Unit, TPU), artificial intelligence (Artificial Intelligence, AI) type processors and other processing devices) and a memory for storing data. Optionally, the above-mentioned computer terminal may also include a transmission device, an input and output device, and a display device for communication functions. It can be understood by those of ordinary skill in the art that the above-mentioned structural description is only for illustration, and it does not limit the structure of the above-mentioned computer terminal. For example, the computer terminal may also include more or fewer components than the above-mentioned structural description, or have a configuration different from the above-mentioned structural description.

[0027] The memory can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the model training method in the embodiment of the present application, and the processor executes various functional applications and data processing by running the computer program stored in the memory, that is, realizing the above-mentioned model training method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the mobile terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0028] The transmission device is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device can be a radio frequency (Radio Frequency, referred to as RF) module, which is used to communicate with the Internet wirelessly.

[0029] The display device may be, for example, a touch screen type liquid crystal display (LCD) and a touch display (also referred to as a "touch screen" or "touch display screen"). The liquid crystal display may enable a user to interact with a user interface of a mobile terminal. In some embodiments, the mobile terminal has a graphical user interface (GUI), and a user may interact with the GUI by finger contacts and / or gestures on a touch-sensitive surface, wherein the human-computer interaction functions here may optionally include the following interactions: creating web pages, drawing, word processing, making electronic documents, games, video conferencing, instant messaging, sending and receiving emails, call interfaces, playing digital videos, playing digital music and / or web browsing, etc. The executable instructions for executing the above human-computer interaction functions are configured / stored in a computer program product or readable storage medium executable by one or more processors.

[0030] According to an embodiment of the present application, a method embodiment of a model training method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0031] Figure 1 is a flow chart of a model training method according to one embodiment of the present application, such as Figure 1 As shown, the method comprises the following steps:

[0032] Step S11, using the target device to collect data from multiple directions around the target vehicle to obtain a target data set, wherein the target device is configured with a communication component and an ultra-wideband component, the communication component is used to obtain the signal strength between the target device and the target vehicle, and the ultra-wideband component is used to obtain the distance between the target device and the target vehicle, and the target data set includes: static data and dynamic data, the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state;

[0033] Step S12: Based on the target data set, the initial convolutional neural network model is trained to obtain a target convolutional neural network model, wherein the target convolutional neural network model is used to guide the domain controller of the target vehicle to determine the control action to be executed.

[0034] The target device may be a mobile device equipped with a communication component and an ultra-wideband component, including but not limited to a smartphone, a tablet computer or a smart watch, and hardware with communication and ultra-wideband (UWB) functions. The target device may collect and process data through pre-installed applications or software to obtain signal strength and distance information around the target vehicle.

[0035] The communication component on the target device is used to obtain the signal strength between the target device and the target vehicle, such as the Received Signal Strength Indication (RSSI) value. By monitoring the signal strength, the communication quality between the target device and the target vehicle can be evaluated, and the communication parameters can be adjusted as needed to ensure stable data transmission and connection performance. RSSI values ​​are usually expressed in decibels (dBm), and the larger the value, the stronger the signal strength. By real-time monitoring and analysis of RSSI values, it can help optimize the performance of the communication system and improve the stability and reliability of communication.

[0036] The ultra-wideband component on the target device is used to obtain the distance between the target device and the target vehicle. For example, data transmission is achieved by sending and receiving pulse signals with extremely short time domain width, thereby measuring the distance between the target device and the target vehicle, and achieving centimeter-level positioning accuracy.

[0037] For example, during the data collection process, the target device will periodically send signals to the target vehicle and receive signals returned by the target vehicle, obtain signal strength information through the communication component, and obtain distance information through the ultra-wideband component. Through multiple acquisitions, signal strength and distance information in different directions can be obtained to construct a target data set. The target data set can be used to locate and track the target vehicle. For example, the direction of the target vehicle relative to the target device can be determined through signal strength information, and the position of the target vehicle can be determined through distance information.

[0038] The target data set includes static data and dynamic data. The static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, thereby simulating the scenario of the user briefly staying around the target vehicle. The dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state, thereby simulating the scenario of the user approaching or moving away from the target vehicle.

[0039] Exemplarily, static data can be obtained by fixing the target device at different positions and collecting signal strength data and distance data, so that a mapping relationship between signal strength and distance at different positions can be established. Dynamic data can be obtained by collecting signal strength data and distance data in real time when the target device moves, so as to simulate the change of signal strength of the user at different distances. When collecting dynamic data, dynamic calibration can be performed with the target vehicle as the center, first moving away and then approaching. By comprehensively analyzing static data and dynamic data, the initial convolutional neural network model between the target device and the target vehicle can be trained to obtain the target convolutional neural network model, thereby realizing real-time monitoring and positioning of the user's position changes around the target vehicle.

[0040] Figure 2 is a schematic diagram of a data collection area according to one embodiment of the present application, such as Figure 2 As shown, during the data collection process, a key or tool with UWB function is used to collect data in each direction and angle around the target vehicle. Taking the target vehicle as the origin, the unlocking area and the locking area are arranged from the inside to the outside. In addition, a reserved area, a welcome area, a connection area, etc. can be set according to needs.

[0041] Based on the target data set, the initial convolutional neural network model is trained to obtain the target convolutional neural network model, which is a model after the convolution kernel is continuously optimized and confirmed, and can guide the domain controller of the target vehicle to determine the control actions to be executed. The above-mentioned control actions to be executed include but are not limited to vehicle unlocking, starting, accelerating, decelerating, steering and other actions.

[0042] During the training process, the target dataset can be used to supervise the initial convolutional neural network model to improve the accuracy and generalization ability of the model. By continuously adjusting the parameters and structure of the model, it can better adapt to the characteristics of the target dataset, thereby achieving more accurate control action prediction. After the training is completed, the target convolutional neural network model can be used to guide the domain controller of the target vehicle to determine the control actions to be performed, so as to improve the performance and safety of the vehicle in the target area.

[0043] Based on the above steps S11 to S12, by using a target device equipped with a communication component and an ultra-wideband component to collect data from multiple directions around the target vehicle, a target data set containing static data and dynamic data is obtained, and then based on the target data set, the initial convolutional neural network model is trained to obtain a target convolutional neural network model, thereby guiding the domain controller of the target vehicle to determine the control action to be executed, thereby achieving the purpose of efficiently and accurately controlling the target vehicle, thereby achieving the technical effect of improving the positioning accuracy of the target device and the user experience, and thus solving the technical problems of low positioning accuracy and poor user experience in the Bluetooth positioning method provided in the relevant technology.

[0044] The model training method in the embodiment of the present application is further introduced below.

[0045] Optionally, in step S12, based on the target data set, the initial convolutional neural network model is trained to obtain a target convolutional neural network model including:

[0046] Step S121, screening the target data set to obtain screening results;

[0047] Step S122: training the initial convolutional neural network model based on the screening results to obtain a target convolutional neural network model.

[0048] Exemplarily, data screening of the target data set can be performed according to task requirements. For example, data can be screened based on labels to select only data samples of a specific category; data can also be screened based on data quality to remove noise data, etc.

[0049] After obtaining the screening results, use the screening results to train the initial convolutional neural network model. By training the initial convolutional neural network model, the characteristics and rules in the screening results can be gradually learned, thereby improving the performance of the model on the task. The final target convolutional neural network model can better adapt to the task requirements and have better performance.

[0050] Based on the above optional embodiments, irrelevant or noisy data can be removed through data screening, thereby improving the quality of the data set, so that the model can learn the characteristics of the data more accurately. Training based on the screening results can enable the model to learn the characteristics of the data set more accurately, thereby improving the generalization ability and accuracy of the model.

[0051] Optionally, in step S121, the target data set is screened, and the screening results include:

[0052] Step S1211, determining whether the distance data collected in the target data set has matching signal strength data;

[0053] Step S1212, in response to part of the distance data in the target data set not having matching signal strength data, determining part of the distance data as dirty data;

[0054] Step S1213, filter out dirty data from the target data set to obtain a filter result.

[0055] Specifically, it is determined whether the distance data collected in the target data set has matching signal strength data. When part of the distance data in the target data set does not have matching signal strength data, that is, the real-time RSSI value is not monitored, the part of the distance data is determined as dirty data, and the dirty data is filtered out from the target data set to obtain a filtering result.

[0056] Based on the above optional embodiments, in response to the fact that some distance data in the target data set do not have matching signal strength data, some distance data are determined as dirty data, and then the dirty data are filtered out from the target data set to obtain a filtering result. By identifying and removing dirty data, errors and noise in the data set can be reduced, thereby improving the reliability and validity of the data.

[0057] Optionally, in step S122, the initial convolutional neural network model is trained based on the screening result to obtain a target convolutional neural network model including:

[0058] Step S1221, determining input data of the initial convolutional neural network model based on the screening result, wherein the input data includes: signal strength input data and distance input data determined based on the screening result;

[0059] Step S1222, using the input data to train the initial convolutional neural network model to obtain output data of the initial convolutional neural network model, wherein the output data includes: predicted distance data and predicted azimuth data;

[0060] Step S1223, adjust the initial convolutional neural network model according to the output data and the real data to obtain the target convolutional neural network model.

[0061] Specifically, the signal strength input data is the RSSI value in the screening results: R0, R1, R2, R3..., and the distance input data is the position distance in the screening results: D0, D1, D2, D3... The input data also includes the orientation information Dire1 (front), Dire2 (left front), Dire3 (left back), Dire4 (back), Dire5 (right back), Dire6 (right front) in the screening results. The input data is used to train the initial convolutional neural network model to obtain the output data of the initial convolutional neural network model, and the output data includes the predicted distance data D1', D2', D3'..., and the predicted orientation data Dire1' (front), Dire2' (left front), Dire3' (left back), Dire4' (back), Dire5' (right back), Dire6' (right front). The initial convolutional neural network model is adjusted according to the output data and the real data to obtain the target convolutional neural network model.

[0062] Based on the above optional embodiments, the input data of the initial convolutional neural network model is determined based on the screening results, and then the initial convolutional neural network model is trained using the input data to obtain the output data of the initial convolutional neural network model. Finally, the initial convolutional neural network model is adjusted according to the output data and the real data to obtain the target convolutional neural network model. By continuously adjusting and optimizing the model, the model's ability to recognize and classify input data can be improved, thereby improving the accuracy and performance of the model. The target convolutional neural network model finally obtained can better adapt to actual application needs and improve the efficiency and accuracy of data processing and analysis.

[0063] Optionally, the initial convolutional neural network model includes: a convolutional neural network region model and a convolutional neural network orientation model. In step S1222, the initial convolutional neural network model is trained using input data to obtain output data of the initial convolutional neural network model, including: training the convolutional neural network region model using signal strength input data and distance input data, and outputting predicted distance data; training the convolutional neural network orientation model using signal strength input data and distance input data, and outputting predicted orientation data.

[0064] The above-mentioned primary convolutional neural network model includes a convolutional neural network regional model and a convolutional neural network orientation model. The signal strength input data R0, R1, R2, R3... and the distance input data D0, D1, D2, D3... are used to train the convolutional neural network regional model, and the predicted distance data D1', D2', D3'... are output. The signal strength input data R0, R1, R2, R3... and the distance input data D0, D1, D2, D3... are used to train the convolutional neural network orientation model, and the predicted orientation data Dire1' (front), Dire2' (left front), Dire3' (left back), Dire4' (back), Dire5' (right back), and Dire6' (right front) are output.

[0065] Based on the above optional embodiments, by using signal strength input data and distance input data to train the convolutional neural network area model and outputting predicted distance data, and by using signal strength input data and distance input data to train the convolutional neural network orientation model and outputting predicted orientation data, feature extraction and association learning of the input data can be achieved, thereby accurately predicting the distance and orientation data.

[0066] Optionally, the real data includes: real distance data corresponding to the predicted distance data and real orientation data corresponding to the predicted orientation data. In step S1223, the initial convolutional neural network model is adjusted according to the output data and the real data to obtain a target convolutional neural network model, including: distance loss calculated based on the predicted distance data and the real distance data, and orientation loss calculated based on the predicted orientation data and the real orientation data; model parameters of the convolutional neural network regional model are adjusted according to the distance loss, and model parameters of the convolutional neural network orientation model are adjusted according to the orientation loss to obtain the target convolutional neural network model.

[0067] Figure 3 is a schematic diagram of a model training process according to one embodiment of the present application, such as Figure 3 As shown, the input data x of the initial convolutional neural network model is determined based on the screening results. i , using input data x i Train the initial convolutional neural network model to obtain the output data y of the initial convolutional neural network model i , each time the output result y i With real data Perform a curve regression. The distance loss can be calculated based on the predicted distance data and the actual distance data, or the orientation loss can be calculated based on the predicted orientation data and the actual orientation data, thereby providing input data for the next back propagation.i1 、w ij 、w ik Continuously optimize and confirm, and determine the model accuracy based on the computing power of the vehicle control unit (MCU).

[0068] Based on the above optional embodiment, the initial convolutional neural network model is optimized by combining the distance loss and the orientation loss, thereby improving the model's prediction accuracy for distance and orientation. By calculating the distance loss and the orientation loss, the performance of the model can be better evaluated, and the model parameters can be adjusted in a targeted manner to improve the prediction precision and accuracy of the model.

[0069] Optionally, the distance input data includes: multiple distance values, wherein the multiple distance values ​​are used to describe the distance between a starting position and multiple end positions, the starting position is a reference position determined based on the target vehicle, the multiple end positions are boundary positions of multiple areas, and the multiple areas are areas divided from the inside to the outside with the target vehicle as the center.

[0070] The above-mentioned multiple distance values ​​are used to describe the distances between a starting position and multiple end positions, where the starting position is a reference position determined based on the target vehicle, for example, a reference position determined based on the outline of the target vehicle.

[0071] Optionally, the multiple directions include: directly in front, left front, left rear, directly rear, right rear and right front centered on the target vehicle.

[0072] Figure 4 is a flow chart of a data processing method according to one embodiment of the present application, such as Figure 4 As shown, the method comprises the following steps:

[0073] Step S41, obtaining the current signal strength between the target device and the target vehicle;

[0074] Step S42, using the target convolutional neural network model to perform distance prediction on the current signal strength to obtain a distance prediction result;

[0075] Step S43, determining the target control action to be executed by the domain controller of the target vehicle based on the distance prediction result; wherein, the target convolutional neural network model is obtained by training the initial convolutional neural network model based on the target data set, the target data set is obtained by collecting data from multiple directions around the target vehicle using a target device, the target device is configured with a communication component and an ultra-wideband component, the communication component is used to obtain the signal strength between the target device and the target vehicle, the ultra-wideband component is used to obtain the distance between the target device and the target vehicle, the target data set includes: static data and dynamic data, the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state.

[0076] Based on the above steps S41 to S43, by acquiring the current signal strength between the target device and the target vehicle, and then using the target convolutional neural network model to predict the distance of the current signal strength, a distance prediction result is obtained, and finally the target control action to be executed by the domain controller of the target vehicle is determined based on the distance prediction result, thereby achieving the purpose of efficiently and accurately controlling the target vehicle, thereby achieving the technical effect of improving the positioning accuracy of the target device and the user experience, and thus solving the technical problems of low positioning accuracy and poor user experience in the Bluetooth positioning method provided in the relevant technology.

[0077] The data processing method in the embodiment of the present application is further introduced below.

[0078] Optionally, in step S43, determining a target control action to be performed by the domain controller of the target vehicle based on the distance prediction result includes:

[0079] Step S431, in response to determining that the target device is located in a preset area based on the distance prediction result, using the target convolutional neural network model to perform azimuth prediction on the current signal strength to obtain an azimuth prediction result;

[0080] Step S432, determining the target control action to be executed by the domain controller of the target vehicle based on the distance prediction result and the direction prediction result.

[0081] Based on the above optional embodiments, in response to determining that the target device is located in a preset area based on the distance prediction result, a target convolutional neural network model is used to predict the direction of the current signal strength to obtain a direction prediction result, and then the target control action to be executed by the domain controller of the target vehicle is determined based on the distance prediction result and the direction prediction result, thereby further improving the accuracy and precision of the target device positioning.

[0082] Optionally, in step S43, determining a target control action to be executed by the domain controller of the target vehicle based on the distance prediction result includes: in response to determining based on the distance prediction result that the target device is not located in a preset area, determining a target control action to be executed by the domain controller of the target vehicle based on the distance prediction result.

[0083] Figure 5 is a schematic diagram of a data processing method according to one embodiment of the present application, such as Figure 5 As shown, the method comprises the following steps:

[0084] Step S501, obtaining the current signal strength between the target device and the target vehicle;

[0085] Step S502, using the target convolutional neural network model to perform distance prediction on the current signal strength to obtain a distance prediction result;

[0086] Step S503, in response to determining that the target device is located in a preset area based on the distance prediction result, using the target convolutional neural network model to perform azimuth prediction on the current signal strength to obtain an azimuth prediction result;

[0087] Step S504, determining a target control action to be executed by the domain controller of the target vehicle based on the distance prediction result and the orientation prediction result;

[0088] Step S505, in response to determining based on the distance prediction result that the target device is not located in the preset area, determining a target control action to be executed by the domain controller of the target vehicle based on the distance prediction result.

[0089] Based on the above optional embodiments, by acquiring the current signal strength between the target device and the target vehicle, and then using the target convolutional neural network model to predict the distance of the current signal strength, a distance prediction result is obtained, and finally the target control action to be executed by the domain controller of the target vehicle is determined based on the distance prediction result, thereby achieving the purpose of efficiently and accurately controlling the target vehicle, thereby achieving the technical effect of improving the positioning accuracy of the target device and the user experience, and thus solving the technical problems of low positioning accuracy and poor user experience existing in the Bluetooth positioning method provided in the relevant technology.

[0090] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0091] In the embodiments of the present application, a model training device is also provided, which is used to implement the above-mentioned embodiments and preferred embodiments, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware is also possible and conceived.

[0092] Figure 6 is a structural block diagram of a model training device according to one embodiment of the present application. Figure 6 As shown, the device comprises:

[0093] The acquisition module 601 is used to acquire data from multiple directions around the target vehicle using the target device to obtain a target data set, wherein the target device is configured with a communication component and an ultra-wideband component, the communication component is used to acquire the signal strength between the target device and the target vehicle, and the ultra-wideband component is used to acquire the distance between the target device and the target vehicle, and the target data set includes: static data and dynamic data, the static data is used to determine the signal strength data and distance data acquired when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data acquired when the target device is in a moving state;

[0094] The training module 602 is used to train the initial convolutional neural network model based on the target data set to obtain a target convolutional neural network model, wherein the target convolutional neural network model is used to guide the domain controller of the target vehicle to determine the control action to be executed.

[0095] Optionally, the training module 602 is also used to: perform data screening on the target data set to obtain screening results; and train the initial convolutional neural network model based on the screening results to obtain a target convolutional neural network model.

[0096] Optionally, the training module 602 is also used to: determine whether the distance data collected in the target data set has matching signal strength data; in response to some distance data in the target data set not having matching signal strength data, determine some distance data as dirty data; filter out the dirty data from the target data set to obtain a filtering result.

[0097] Optionally, the training module 602 is also used to: determine the input data of the initial convolutional neural network model based on the screening results, wherein the input data includes: signal strength input data and distance input data determined based on the screening results; use the input data to train the initial convolutional neural network model to obtain the output data of the initial convolutional neural network model, wherein the output data includes: predicted distance data and predicted azimuth data; adjust the initial convolutional neural network model according to the output data and the real data to obtain the target convolutional neural network model.

[0098] Optionally, the training module 602 is also used to: use signal strength input data and distance input data to train the convolutional neural network area model, and output predicted distance data; use signal strength input data and distance input data to train the convolutional neural network orientation model, and output predicted orientation data.

[0099] Optionally, the training module 602 is also used to: calculate the distance loss based on the predicted distance data and the actual distance data, and calculate the orientation loss based on the predicted orientation data and the actual orientation data; adjust the model parameters of the convolutional neural network regional model according to the distance loss and adjust the model parameters of the convolutional neural network orientation model according to the orientation loss to obtain the target convolutional neural network model.

[0100] Optionally, the distance input data includes: multiple distance values, wherein the multiple distance values ​​are used to describe the distance between a starting position and multiple end positions, the starting position is a reference position determined based on the target vehicle, the multiple end positions are boundary positions of multiple areas, and the multiple areas are areas divided from the inside to the outside with the target vehicle as the center.

[0101] Optionally, the multiple directions include: directly in front, left front, left rear, directly rear, right rear and right front centered on the target vehicle.

[0102] Figure 7 is a structural block diagram of a data processing device according to one embodiment of the present application. Figure 7 As shown, the device comprises:

[0103] An acquisition module 701 is used to acquire the current signal strength between the target device and the target vehicle;

[0104] The prediction module 702 is used to use the target convolutional neural network model to perform distance prediction on the current signal strength to obtain a distance prediction result;

[0105] The processing module 703 is used to determine the target control action to be executed by the domain controller of the target vehicle based on the distance prediction result; wherein, the target convolutional neural network model is obtained by training the initial convolutional neural network model based on the target data set, and the target data set is obtained by collecting data from multiple directions around the target vehicle using a target device. The target device is configured with a communication component and an ultra-wideband component, the communication component is used to obtain the signal strength between the target device and the target vehicle, and the ultra-wideband component is used to obtain the distance between the target device and the target vehicle. The target data set includes: static data and dynamic data, the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state.

[0106] Optionally, the processing module 703 is also used to: in response to determining that the target device is located in a preset area based on the distance prediction result, use the target convolutional neural network model to predict the direction of the current signal strength to obtain a direction prediction result; determine the target control action to be executed by the domain controller of the target vehicle based on the distance prediction result and the direction prediction result.

[0107] Optionally, the processing module 703 is further used to: in response to determining that the target device is not located in a preset area based on the distance prediction result, determine a target control action to be executed by the domain controller of the target vehicle based on the distance prediction result.

[0108] It should be noted that the above modules can be implemented by software or hardware. For the latter, it can be implemented in the following ways, but not limited to: the above modules are all located in the same processor; or the above modules are located in different processors in any combination.

[0109] An embodiment of the present application also provides a vehicle, comprising: a memory storing an executable program; and a processor for running the program, wherein when the program is running, any model training method or any data processing method in the various embodiments of the present application is executed.

[0110] Optionally, in this embodiment, the processor may be configured to perform the following steps through a computer program:

[0111] S1, using a target device to collect data from multiple directions around a target vehicle to obtain a target data set, wherein the target device is configured with a communication component and an ultra-wideband component, the communication component is used to obtain the signal strength between the target device and the target vehicle, and the ultra-wideband component is used to obtain the distance between the target device and the target vehicle, and the target data set includes: static data and dynamic data, the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state;

[0112] S2, based on the target data set, training the initial convolutional neural network model to obtain a target convolutional neural network model, wherein the target convolutional neural network model is used to guide the domain controller of the target vehicle to determine the control action to be executed.

[0113] Optionally, in this embodiment, the processor may be configured to perform the following steps through a computer program:

[0114] S1, obtaining the current signal strength between the target device and the target vehicle;

[0115] S2, using the target convolutional neural network model to predict the distance of the current signal strength and obtain the distance prediction result;

[0116] S3, determining the target control action to be executed by the domain controller of the target vehicle based on the distance prediction result; wherein, the target convolutional neural network model is obtained by training the initial convolutional neural network model based on the target data set, the target data set is obtained by collecting data from multiple directions around the target vehicle using a target device, and a communication component and an ultra-wideband component are configured on the target device. The communication component is used to obtain the signal strength between the target device and the target vehicle, and the ultra-wideband component is used to obtain the distance between the target device and the target vehicle. The target data set includes: static data and dynamic data, the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state.

[0117] An embodiment of the present application also provides a computer-readable storage medium, which includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the model training methods or any one of the data processing methods in the various embodiments of the present application.

[0118] Optionally, in this embodiment, the storage medium may be configured to store a computer program for performing the following steps:

[0119] S1, using a target device to collect data from multiple directions around a target vehicle to obtain a target data set, wherein the target device is configured with a communication component and an ultra-wideband component, the communication component is used to obtain the signal strength between the target device and the target vehicle, and the ultra-wideband component is used to obtain the distance between the target device and the target vehicle, and the target data set includes: static data and dynamic data, the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state;

[0120] S2, based on the target data set, training the initial convolutional neural network model to obtain a target convolutional neural network model, wherein the target convolutional neural network model is used to guide the domain controller of the target vehicle to determine the control action to be executed.

[0121] Optionally, in this embodiment, the storage medium may be configured to store a computer program for performing the following steps:

[0122] S1, obtaining the current signal strength between the target device and the target vehicle;

[0123] S2, using the target convolutional neural network model to predict the distance of the current signal strength and obtain the distance prediction result;

[0124] S3, determining the target control action to be executed by the domain controller of the target vehicle based on the distance prediction result; wherein, the target convolutional neural network model is obtained by training the initial convolutional neural network model based on the target data set, the target data set is obtained by collecting data from multiple directions around the target vehicle using a target device, and a communication component and an ultra-wideband component are configured on the target device. The communication component is used to obtain the signal strength between the target device and the target vehicle, and the ultra-wideband component is used to obtain the distance between the target device and the target vehicle. The target data set includes: static data and dynamic data, the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state.

[0125] Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store computer programs.

[0126] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed by a processor, implements any model training method or any data processing method in the various embodiments of the present application.

[0127] An embodiment of the present application also provides a computer program product, including a non-volatile computer-readable storage medium, the non-volatile computer-readable storage medium is used to store a computer program, and when the computer program is executed by a processor, it implements any model training method or any data processing method in the various embodiments of the present application.

[0128] An embodiment of the present application also provides a computer program, which, when executed by a processor, implements any one of the model training methods or any one of the data processing methods in the above-mentioned embodiments of the present application.

[0129] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0130] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0131] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0132] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0133] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can essentially or in other words, the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a read-only memory, a random access memory, a mobile hard disk, a magnetic disk or an optical disk.

[0134] The above is only a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A model training method, characterized in that: include: A target data set is obtained by using a target device to collect data from multiple directions around a target vehicle, wherein the target device is configured with a communication component and an ultra-wideband component, the communication component is used to obtain the signal strength between the target device and the target vehicle, the ultra-wideband component is used to obtain the distance between the target device and the target vehicle, and the target data set includes: static data and dynamic data, the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state; Based on the target data set, an initial convolutional neural network model is trained to obtain a target convolutional neural network model, wherein the target convolutional neural network model is used to guide a domain controller of the target vehicle to determine a control action to be executed.

2. The model training method according to claim 1, characterized in that: Based on the target data set, the initial convolutional neural network model is trained to obtain the target convolutional neural network model, including: Performing data screening on the target data set to obtain screening results; The initial convolutional neural network model is trained based on the screening result to obtain the target convolutional neural network model.

3. The model training method according to claim 2, characterized in that: The target data set is screened to obtain the screening result, including: Determining whether the distance data collected in the target data set has matching signal strength data; In response to part of the distance data in the target data set not having matching signal strength data, determining the part of the distance data as dirty data; The dirty data is filtered out from the target data set to obtain the filtering result.

4. The model training method according to claim 2, characterized in that: Training the initial convolutional neural network model based on the screening result to obtain the target convolutional neural network model includes: Determining input data of the initial convolutional neural network model based on the screening result, wherein the input data includes: signal strength input data and distance input data determined based on the screening result; The initial convolutional neural network model is trained using the input data to obtain output data of the initial convolutional neural network model, wherein the output data includes: predicted distance data and predicted azimuth data; The initial convolutional neural network model is adjusted according to the output data and the real data to obtain the target convolutional neural network model.

5. The model training method according to claim 4, characterized in that: The initial convolutional neural network model includes: a convolutional neural network region model and a convolutional neural network orientation model. The initial convolutional neural network model is trained using the input data to obtain the output data of the initial convolutional neural network model including: Using the signal strength input data and the distance input data to train the convolutional neural network regional model, and output the predicted distance data; The convolutional neural network orientation model is trained using the signal strength input data and the distance input data, and the predicted orientation data is output.

6. The model training method according to claim 5, characterized in that: The real data includes: the real distance data corresponding to the predicted distance data and the real orientation data corresponding to the predicted orientation data. The initial convolutional neural network model is adjusted according to the output data and the real data to obtain the target convolutional neural network model, including: Calculating a distance loss based on the predicted distance data and the actual distance data, and calculating a direction loss based on the predicted direction data and the actual direction data; The model parameters of the convolutional neural network region model are adjusted according to the distance loss, and the model parameters of the convolutional neural network orientation model are adjusted according to the orientation loss to obtain the target convolutional neural network model.

7. The model training method according to claim 4, characterized in that: The distance input data includes: multiple distance values, wherein the multiple distance values ​​are used to describe the distance between a starting position and multiple end positions, the starting position is a reference position determined based on the target vehicle, and the multiple end positions are boundary positions of multiple areas, and the multiple areas are areas divided from the inside to the outside with the target vehicle as the center.

8. The model training method according to claim 5, characterized in that: The multiple directions include: directly in front, left front, left rear, directly rear, right rear and right front with the target vehicle as the center.

9. A data processing method, characterized in that: include: Get the current signal strength between the target device and the target vehicle; Using a target convolutional neural network model to perform distance prediction on the current signal strength to obtain a distance prediction result; Determining a target control action to be executed by the domain controller of the target vehicle based on the distance prediction result; Among them, the target convolutional neural network model is obtained by training the initial convolutional neural network model based on the target data set, and the target data set is obtained by collecting data from multiple directions around the target vehicle using a target device. The target device is configured with a communication component and an ultra-wideband component, and the communication component is used to obtain the signal strength between the target device and the target vehicle, and the ultra-wideband component is used to obtain the distance between the target device and the target vehicle. The target data set includes: static data and dynamic data, and the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state.

10. The data processing method according to claim 9, characterized in that: Determining the target control action to be performed by the domain controller of the target vehicle based on the distance prediction result includes: In response to determining that the target device is located in a preset area based on the distance prediction result, using the target convolutional neural network model to perform an azimuth prediction on the current signal strength to obtain an azimuth prediction result; The target control action to be executed by the domain controller of the target vehicle is determined based on the distance prediction result and the orientation prediction result.

11. The data processing method according to claim 9, characterized in that: Determining the target control action to be performed by the domain controller of the target vehicle based on the distance prediction result includes: In response to determining that the target device is not located in a preset area based on the distance prediction result, the target control action to be performed by the domain controller of the target vehicle is determined based on the distance prediction result.

12. A model training device, characterized in that: include: A collection module, used for collecting data from multiple directions around a target vehicle using a target device to obtain a target data set, wherein the target device is configured with a communication component and an ultra-wideband component, the communication component is used to obtain the signal strength between the target device and the target vehicle, the ultra-wideband component is used to obtain the distance between the target device and the target vehicle, and the target data set includes: static data and dynamic data, the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state; A training module is used to train an initial convolutional neural network model based on the target data set to obtain a target convolutional neural network model, wherein the target convolutional neural network model is used to guide the domain controller of the target vehicle to determine a control action to be executed.

13. A data processing device, characterized in that: include: An acquisition module, used to acquire the current signal strength between the target device and the target vehicle; A prediction module, used to use a target convolutional neural network model to perform distance prediction on the current signal strength to obtain a distance prediction result; A processing module, configured to determine a target control action to be executed by a domain controller of the target vehicle based on the distance prediction result; Among them, the target convolutional neural network model is obtained by training the initial convolutional neural network model based on the target data set, and the target data set is obtained by collecting data from multiple directions around the target vehicle using a target device. The target device is configured with a communication component and an ultra-wideband component, and the communication component is used to obtain the signal strength between the target device and the target vehicle, and the ultra-wideband component is used to obtain the distance between the target device and the target vehicle. The target data set includes: static data and dynamic data, and the static data is used to determine the signal strength data and distance data collected when the target device is in a stationary state, and the dynamic data is used to determine the signal strength data and distance data collected when the target device is in a moving state.

14. A vehicle, characterized in that: include: A memory storing an executable program; A processor, used to run the program, wherein when the program is run, the model training method described in any one of claims 1 to 8 or the data processing method described in any one of claims 9 to 11 is executed.

15. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored executable program, wherein, when the executable program is running, the device where the storage medium is located is controlled to execute the model training method described in any one of claims 1 to 8 or the data processing method described in any one of claims 9 to 11.

16. A computer program product, characterized in that It includes a computer program, which, when executed by a processor, implements the model training method according to any one of claims 1 to 8 or the data processing method according to any one of claims 9 to 11.