Automatic picking-up method and device, terminal equipment and storage medium

Abstract

The invention provides an automatic picking-up method and device, terminal equipment and a storage medium. The method comprises the steps that an automatic driving vehicle is started based on a driving request sent by a user terminal; wherein the driving request comprises a target position where a user is located, and the user terminal establishes communication with the autonomous vehicle through the Internet of Things in advance; controlling the autonomous vehicle to call a pre-constructed high-precision map based on the driving request, performing route screening on the high-precision map based on the parking position of the autonomous vehicle and the target position, and generating a route planning result; wherein the high-precision map comprises a plurality of parking lot maps including a parking lot where the parking position is located; and controlling the automatic driving vehicle to automatically drive based on the path planning result until the automatic driving vehicle reaches the target position. The problem that a user is difficult to find a car in a parking lot of a large shopping mall can be solved, and the user experience is improved.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to an automatic pick-up method, a device terminal equipment and a storage medium. Background Art

[0002] With the acceleration of urbanization and the continuous increase in the number of cars, the construction of large parking lots for households and supermarkets is developing rapidly. These parking lots are usually large in area and complex in structure. Although they provide convenient parking spaces for car owners, they also bring many new challenges and problems. For example, users often face multiple problems in parking lots, such as difficulty in finding cars, inconvenience in parking, and the risk of scratches. These problems not only affect the user's parking experience, but may also lead to traffic congestion and increased searching time in the parking lot, thus affecting the overall operational efficiency of the parking lot. Since modern parking lots usually adopt an open design and parking spaces are often distributed in multiple areas, many car owners find it difficult to quickly find their vehicles after parking.

[0003] Traditional methods of finding a car rely on car owners to memorize parking space numbers or surrounding landmarks. However, this method has many shortcomings in practical applications: car owners may forget the specific parking space number due to environmental changes when parking, especially during peak hours, when changes in the parking lot environment (such as temporary obstacles, construction, or obstruction of signs) may render the originally marked landmarks invalid, causing trouble for car owners when finding their cars. Traditional methods of finding a car are difficult to meet the growing needs of users, and cannot effectively improve the parking experience. Summary of the invention

[0004] The present invention aims to provide an automatic pick-up method, a device terminal and a storage medium to solve the problem of users having difficulty finding a car in a parking lot of a large shopping mall and improve user experience.

[0005] In order to solve the above technical problems, the present invention provides an automatic pick-up method, comprising:

[0006] Starting the autonomous driving vehicle based on a driving request sent by a user terminal; wherein the driving request includes a target location of the user, and the user terminal and the autonomous driving vehicle establish communication in advance through the Internet of Things;

[0007] Controlling the autonomous driving vehicle to call a pre-built high-precision map based on the driving request, performing route screening on the high-precision map based on the parking position of the autonomous driving vehicle and the target position, and generating a path planning result; wherein the high-precision map includes a plurality of parking lot maps including the parking lot where the parking position is located;

[0008] Based on the path planning result, the autonomous driving vehicle is controlled to perform autonomous driving until the autonomous driving vehicle reaches the target location.

[0009] In the above scheme, the precise parking space coordinates and environmental features are provided by the pre-built high-precision map, ensuring that the autonomous driving vehicle can accurately identify the parking location. The user only needs to send a driving request through the terminal device, and the autonomous driving vehicle can automatically start and go to the designated location. There is no need for users to remember the parking space number or surrounding landmarks, which significantly reduces the user's memory burden. The autonomous driving vehicle has real-time environmental perception capabilities and can detect environmental changes such as temporary obstacles in the parking lot, road construction, or obstruction of signs. The automatic pick-up method can dynamically adjust the driving route based on the real-time perception results to ensure that the vehicle can avoid obstacles and reach the target location safely. The path planning algorithm based on the high-precision map can generate the optimal driving route in real time, adapt to environmental changes, and ensure the smoothness and safety of the driving process. The automatic pick-up method can quickly respond to the user's driving request, quickly plan the best route and execute it, significantly shortening the time to find a car and improving the efficiency of finding a car.

[0010] In one implementation, the automatic pick-up method further includes performing patrol training of the autonomous driving vehicle in different types of parking lots, and constructing the high-precision map based on patrol training results, specifically:

[0011] Controlling the autonomous driving vehicle to perform autonomous driving training in different types of parking lots; wherein the autonomous driving vehicle is equipped with sensors and a positioning system;

[0012] Based on the sensor, the parking environment data of each autonomous driving training process is collected.

[0013] Based on the positioning system, the autonomous driving vehicle is globally positioned in each autonomous driving training, and driving data of the autonomous driving vehicle is collected;

[0014] Performing performance evaluation on the driving data based on preset performance evaluation indicators, and ending the autonomous driving training for the parking lot when each of the preset performance evaluation indicators reaches a preset standard;

[0015] The parking lot environment data collected during each autonomous driving training is acquired, element annotation is performed on the parking lot environment data, and the high-precision map is constructed based on the element-annotated parking lot environment data.

[0016] In one implementation, controlling the autonomous driving vehicle to call a pre-built high-precision map based on the driving request specifically includes:

[0017] Obtaining real-time location information and real-time parking environment data of the autonomous driving vehicle;

[0018] Performing a preliminary match on the pre-built high-precision map based on the real-time parking environment data, obtaining a first high-precision map corresponding to the parking location and calling the first high-precision map;

[0019] Feature matching is performed on the first high-precision map based on the real-time parking environment data, and the parking position is generated according to the feature matching result of the real-time parking environment data in the first high-precision map.

[0020] In one implementation, the performing of route screening on the high-precision map based on the parking position of the autonomous driving vehicle and the target position to generate a path planning result specifically includes:

[0021] Preprocessing the real-time parking lot environment data to obtain first real-time parking lot environment data;

[0022] The first real-time parking lot environment data is arranged in time series to obtain a first real-time parking lot environment data sequence, and the first real-time parking lot environment data sequence is input into a pre-constructed neural network model so that the neural network model performs route planning based on the called first high-precision map to generate the path planning result; wherein the neural network model is trained based on the historical driving data of the autonomous driving vehicle and the parking lot environment data of different parking lots.

[0023] In one implementation, the neural network model is trained based on the driving data of the autonomous driving vehicle and parking environment data of different parking lots, specifically including:

[0024] Collecting historical driving data of the autonomous driving vehicle and parking environment data of different parking lots, and constructing a training data set based on the historical driving data and the parking environment data;

[0025] Constructing an initial neural network model; wherein the initial neural network model includes an input layer, a feature extraction module and an output layer;

[0026] The training data set is input into the initial neural network model for iterative training; wherein the input layer is used to transmit the received training data to the feature extraction module, so that the feature extraction module performs feature extraction on the training data and learns the behavior pattern of the autonomous driving vehicle based on the feature extraction results, and the output layer is used to output the path planning result according to the feature extraction result, and the path planning result includes the driving route and driving control instructions.

[0027] In one implementation, inputting the training data set into the initial neural network for iterative training specifically includes:

[0028] Iteratively training the initial neural network model based on a back propagation algorithm; wherein, the gradient of the initial neural network model at each layer after each training is calculated, and the weight of the initial neural network model at each layer is updated based on a preset optimization algorithm;

[0029] When the output of the initial neural network model meets the preset evaluation index, the iterative training of the initial neural network model is terminated to obtain the neural network model.

[0030] In one implementation, controlling the autonomous driving vehicle to perform autonomous driving based on the path planning result until the autonomous driving vehicle reaches the target location specifically includes:

[0031] Controlling the driving trajectory of the autonomous driving vehicle based on the driving route;

[0032] The driving control instructions are distributed to each control system of the automatic driving vehicle so that each control system adjusts the driving behavior based on the received driving control instructions.

[0033] In a second aspect, the present application also provides an automatic pick-up device, including: a request receiving module, a route planning module and a driving control module;

[0034] The request receiving module is used to start the autonomous driving vehicle based on a driving request sent by a user terminal; wherein the driving request includes a target location where the user is located, and the user terminal and the autonomous driving vehicle establish communication in advance through the Internet of Things;

[0035] The route planning module is used to control the autonomous driving vehicle to call a pre-built high-precision map based on the driving request, perform route screening on the high-precision map based on the parking position of the autonomous driving vehicle and the target position, and generate a path planning result; wherein the high-precision map includes multiple parking lot maps including the parking lot where the parking position is located;

[0036] The driving control module is used to control the autonomous driving vehicle to perform autonomous driving based on the path planning result until the autonomous driving vehicle reaches the target location.

[0037] In the above scheme, the precise parking space coordinates and environmental features are provided by the pre-built high-precision map, ensuring that the autonomous driving vehicle can accurately identify the parking location. The user only needs to send a driving request through the terminal device, and the autonomous driving vehicle can automatically start and go to the designated location. There is no need for users to remember the parking space number or surrounding landmarks, which significantly reduces the user's memory burden. The autonomous driving vehicle has real-time environmental perception capabilities and can detect environmental changes such as temporary obstacles in the parking lot, road construction, or obstruction of signs. The automatic pick-up method can dynamically adjust the driving route according to the real-time perception results to ensure that the vehicle can avoid obstacles and reach the target location safely. The path planning algorithm based on the high-precision map can generate the optimal driving route in real time, adapt to environmental changes, and ensure the smoothness and safety of the driving process. The automatic pick-up device can quickly respond to the user's driving request, quickly plan the best route and execute it, significantly shortening the time to find a car and improving the efficiency of finding a car.

[0038] In one implementation, the automatic pick-up device further includes performing patrol training of the automatic driving vehicle in different types of parking lots, and constructing the high-precision map based on patrol training results, specifically:

[0039] Controlling the autonomous driving vehicle to perform autonomous driving training in different types of parking lots; wherein the autonomous driving vehicle is equipped with sensors and a positioning system;

[0040] Based on the sensor, the parking environment data of each autonomous driving training process is collected.

[0041] Based on the positioning system, the autonomous driving vehicle is globally positioned in each autonomous driving training, and driving data of the autonomous driving vehicle is collected;

[0042] Performing performance evaluation on the driving data based on preset performance evaluation indicators, and ending the autonomous driving training for the parking lot when each of the preset performance evaluation indicators reaches a preset standard;

[0043] The parking lot environment data collected during each autonomous driving training is acquired, element annotation is performed on the parking lot environment data, and the high-precision map is constructed based on the element-annotated parking lot environment data.

[0044] In one implementation, controlling the autonomous driving vehicle to call a pre-built high-precision map based on the driving request specifically includes:

[0045] Obtaining real-time location information and real-time parking environment data of the autonomous driving vehicle;

[0046] Performing a preliminary match on the pre-built high-precision map based on the real-time parking environment data, obtaining a first high-precision map corresponding to the parking location and calling the first high-precision map;

[0047] Feature matching is performed on the first high-precision map based on the real-time parking environment data, and the parking position is generated according to the feature matching result of the real-time parking environment data in the first high-precision map.

[0048] In one implementation, the performing of route screening on the high-precision map based on the parking position of the autonomous driving vehicle and the target position to generate a path planning result specifically includes:

[0049] Preprocessing the real-time parking lot environment data to obtain first real-time parking lot environment data;

[0050] The first real-time parking lot environment data is arranged in time series to obtain a first real-time parking lot environment data sequence, and the first real-time parking lot environment data sequence is input into a pre-constructed neural network model so that the neural network model performs route planning based on the called first high-precision map to generate the path planning result; wherein the neural network model is trained based on the historical driving data of the autonomous driving vehicle and the parking lot environment data of different parking lots.

[0051] In one implementation, the neural network model is trained based on the driving data of the autonomous driving vehicle and parking environment data of different parking lots, specifically including:

[0052] Collecting historical driving data of the autonomous driving vehicle and parking environment data of different parking lots, and constructing a training data set based on the historical driving data and the parking environment data;

[0053] Constructing an initial neural network model; wherein the initial neural network model includes an input layer, a feature extraction module and an output layer;

[0054] The training data set is input into the initial neural network model for iterative training; wherein the input layer is used to transmit the received training data to the feature extraction module, so that the feature extraction module performs feature extraction on the training data and learns the behavior pattern of the autonomous driving vehicle based on the feature extraction results, and the output layer is used to output the path planning result according to the feature extraction result, and the path planning result includes the driving route and driving control instructions.

[0055] In one implementation, inputting the training data set into the initial neural network for iterative training specifically includes:

[0056] Iteratively training the initial neural network model based on a back propagation algorithm; wherein, the gradient of the initial neural network model at each layer after each training is calculated, and the weight of the initial neural network model at each layer is updated based on a preset optimization algorithm;

[0057] When the output of the initial neural network model meets the preset evaluation index, the iterative training of the initial neural network model is terminated to obtain the neural network model.

[0058] In one implementation, controlling the autonomous driving vehicle to perform autonomous driving based on the path planning result until the autonomous driving vehicle reaches the target location specifically includes:

[0059] Controlling the driving trajectory of the autonomous driving vehicle based on the driving route;

[0060] The driving control instructions are distributed to each control system of the automatic driving vehicle so that each control system adjusts the driving behavior based on the received driving control instructions.

[0061] In a third aspect, the present application also provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, the automatic pick-up method as described above is implemented.

[0062] In a fourth aspect, the present application further provides a computer-readable storage medium, the computer-readable storage medium including a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the automatic pick-up method as described above. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 A schematic diagram of a flow chart of an automatic pick-up method provided in one embodiment of the present invention;

[0064] Figure 2 The figure is a schematic diagram of a module of an automatic receiving device provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0065] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0066] The terms "first" and "second" and the like in the specification, claims and drawings of this application are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof 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 limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products or devices.

[0067] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0068] First, some terms in this application are explained to facilitate understanding by those skilled in the art.

[0069] (1) User Terminal refers to the equipment or software used by users to directly interact and access services in network communications, information processing or other technical systems. They are usually the entry point to a larger system, and users use these terminals to obtain information, transmit data, perform operations and management tasks.

[0070] (2) Inertial Navigation System (INS): It is an autonomous navigation system. Its working principle is mainly based on Newton's laws of mechanics. It measures the acceleration of the carrier in the inertial reference system, integrates it with time, and then transforms it into the navigation coordinate system to obtain information such as the carrier's speed, yaw angle and position in the navigation coordinate system.

[0071] (3) LiDAR (Light Detection and Ranging): It is an advanced remote sensing technology that accurately detects and measures the distance, position and shape of objects by emitting laser pulses and measuring their return time.

[0072] Example 1

[0073] See also Figure 1 , Figure 1 The flowchart of an automatic pick-up method provided in one embodiment of the present invention is shown in FIG. The automatic pick-up method provided in one embodiment of the present invention includes steps 101 to 103, and each step is specifically as follows:

[0074] Step 101: Starting an autonomous driving vehicle based on a driving request sent by a user terminal; wherein the driving request includes a target location of the user, and the user terminal and the autonomous driving vehicle establish communication in advance through the Internet of Things.

[0075] In an embodiment of the present invention, when a user needs to find a vehicle parked in a parking lot when returning from a parking lot, a driving request containing the user's current location can be sent to the autonomous driving vehicle through a user terminal. The user terminal can be a mobile device such as a smart phone or a tablet computer, or a wearable device such as a smart watch. It should be noted that the user terminal establishes stable communication with the autonomous driving vehicle in advance through cellular network technology or Internet of Things technology. Cellular network technology, especially 4G and 5G, has become the core means for vehicles and user terminals to communicate over long distances. Through the vehicle's built-in communication module (such as an LTE module or a 5G module), the vehicle can access the Internet and communicate in real time with the user's mobile phone APP, web page or other devices. Internet of Things (IoT) technology provides a more intelligent and automated solution for the connection between vehicles and user terminals. Through the IoT platform, both vehicles and user devices can serve as nodes in the IoT and establish stable communication connections with each other. The IoT platform can monitor the status of the vehicle in real time and ensure stable communication between the user terminal and the vehicle. For example, when a user sends a request to "remotely start the vehicle", the IoT platform ensures that the command is delivered to the vehicle through a reliable communication link and feedbacks the operation results after the vehicle executes the command.

[0076] Step 102: Control the autonomous driving vehicle to call a pre-constructed high-precision map based on the driving request, perform route screening on the high-precision map based on the parking position of the autonomous driving vehicle and the target position, and generate a path planning result; wherein the high-precision map includes multiple parking lot maps including the parking lot where the parking position is located.

[0077] In one embodiment, the automatic pick-up method also includes conducting patrol training of the autonomous driving vehicle in different types of parking lots, and constructing the high-precision map based on the patrol training results, specifically: controlling the autonomous driving vehicle to conduct autonomous driving training in different types of parking lots; wherein the autonomous driving vehicle is equipped with sensors and a positioning system; collecting parking lot environment data of each autonomous driving training process based on the sensors, performing global positioning of the autonomous driving vehicle in each autonomous driving training based on the positioning system, and collecting driving data of the autonomous driving vehicle; performing performance evaluation on the driving data based on preset performance evaluation indicators, and ending the autonomous driving training of the parking lot when each of the preset performance evaluation indicators reaches a preset standard; obtaining the parking lot environment data collected for each autonomous driving training, performing element annotation on the parking lot environment data, and constructing the high-precision map based on the parking lot environment data after element annotation.

[0078] In an embodiment of the present invention, an autonomous driving vehicle is controlled to conduct patrol-style autonomous driving training in parking lots of different types (such as open, underground, multi-story, etc.) to collect environmental data in various scenarios. Specifically, the driving of the autonomous driving vehicle is controlled by the sensors and positioning system of the autonomous driving vehicle. The autonomous driving vehicle collects environmental data in real time through a variety of sensors. For example, static and dynamic obstacles (such as walls, pillars, pedestrians, vehicles, etc.) in the parking lot are collected through a laser radar. The laser radar emits laser pulses and receives reflected signals. The distance and direction of the obstacles are calculated by measuring the time delay and angle of the laser pulses. The camera captures images and videos of the parking lot through visual features in the parking lot (such as lane lines, signs, facilities, ambient lighting, etc.), and the visual features are extracted through image processing technology. Ultrasonic sensors are used to collect close-range obstacles, such as the boundaries of parking spaces, other vehicles, etc., and ultrasonic waves are emitted and received. The reflected signal is received, and the position of the obstacle is calculated by measuring the propagation time and distance of the ultrasonic wave, which can be used for accurate detection and obstacle avoidance of parking spaces; the positioning system (such as GNSS, RTK-GPS) is used for the global position of the vehicle, and the satellite signal and the ground base station signal are combined for high-precision positioning. The global position data (latitude, longitude, altitude) is used for global positioning and navigation of the vehicle in the parking lot. Furthermore, the embodiment of the present invention also collects the acceleration and angular velocity of the vehicle through the inertial measurement unit, and calculates the posture and motion state of the vehicle by measuring the acceleration and angular velocity of the vehicle in all directions, and further obtains the driving data of the autonomous driving vehicle. The driving data in the autonomous driving training process is evaluated by the performance evaluation indicators such as positioning accuracy, path planning efficiency, and driving safety. When the evaluation indicators all reach the preset standards, the autonomous driving training of the parking lot is terminated, and the above method is used to realize the patrol training of multiple different types of parking lots. It should be noted that the standards of the performance evaluation indicators can be customized according to the performance of the autonomous driving vehicle, the user's requirements for positioning accuracy, etc., and are not limited here. After completing the patrol training of the parking lot, obtain the parking lot environment data collected during each autonomous driving training, and annotate and classify the collected parking lot environment data. Identify the locations of parking spaces, aisle widths, floor information, stairs / elevators and other elements in the parking lot. Parking space locations are used to identify available and occupied parking spaces; aisle widths are used to ensure width information that can be safely passed during path planning; floor information: used to identify the various floors of the parking lot to meet the driving needs of multi-story parking lots; stair / elevator locations are key vertical transportation facility identifiers to facilitate navigation in multi-story parking lots. Accurately annotating important elements such as parking spaces, aisles, exits, stairs and elevators can ensure that their locations and attributes in the map are clear, and a high-precision, complete parking lot map can be constructed to support subsequent navigation and path planning.

[0079] In one embodiment, the controlling of the autonomous driving vehicle to call a pre-constructed high-precision map based on the driving request specifically includes: obtaining the real-time location information and real-time parking lot environment data of the autonomous driving vehicle; performing preliminary matching on the pre-constructed high-precision map based on the real-time parking lot environment data, obtaining a first high-precision map corresponding to the parking position and calling the first high-precision map; performing feature matching on the first high-precision map based on the real-time parking lot environment data, and generating the parking position according to the feature matching result of the real-time parking lot environment data on the first high-precision map.

[0080] In the embodiment of the present invention, the real-time position information of the autonomous driving and the parking lot environment information are collected through various sensors such as GPS, inertial navigation system (INS), laser radar (LiDAR), camera, etc. of the autonomous driving vehicle. GPS can be used to provide longitude and latitude information; inertial navigation system (INS) can be used to compensate for speed and direction information when GPS signal is lost; laser radar (LiDAR) can provide high-precision three-dimensional environment perception; camera is used to capture visual features, such as lane lines, signs, etc. Using laser radar and camera, environmental data in the parking lot, such as lane lines, signs, wall contours, etc., are collected, and these data will be used for subsequent matching with data in the map. As an optimization scheme of the embodiment of the present invention, multi-source sensor data such as GPS, INS, LiDAR, camera, etc. can also be fused to reduce the error caused by a single sensor and improve the accuracy and reliability of the data. According to the real-time position (longitude and latitude, direction) of the vehicle and the map data, a preliminary position matching is performed on the pre-constructed high-precision map to determine the map area where the vehicle may be located, and the corresponding first high-precision map is obtained and called. Then, the environmental features collected by LiDAR and camera are further compared and matched with the features in the map. These features can include the location and angle of lane lines, the outlines of walls and obstacles, etc. Advanced matching algorithms such as particle filtering, Kalman filtering, HMM models, etc. are used to calculate the similarity or probability between the vehicle and the map features, select the best matching point, and generate the parking position of the autonomous driving vehicle in the first high-precision map. According to the results of map matching, the real-time position of the vehicle can be calibrated. If there is an error in the positioning of GPS or INS, the map matching result will provide a more accurate vehicle location. Based on real-time parking environment data and feature matching results, an accurate parking location is generated. Based on the above steps, the autonomous driving vehicle can achieve high-precision positioning and navigation in a complex parking environment to ensure safe and efficient completion of driving requests.

[0081] In one embodiment, the route screening is performed on the high-precision map based on the parking position of the autonomous driving vehicle and the target position to generate a path planning result, which specifically includes: preprocessing the real-time parking lot environment data to obtain first real-time parking lot environment data; arranging the first real-time parking lot environment data in time series to obtain a first real-time parking lot environment data sequence, and inputting the first real-time parking lot environment data sequence into a pre-constructed neural network model, so that the neural network model performs route planning based on the called first high-precision map to generate the path planning result; wherein the neural network model is trained based on the historical driving data of the autonomous driving vehicle and the parking lot environment data of different parking lots.

[0082] In an embodiment of the present invention, environmental data in the parking lot is collected in real time by sensors of the autonomous driving vehicle (such as lidar, camera, GPS, inertial navigation system, etc.), including lane lines, parking spaces, obstacle locations, floor information, traffic signs, etc. The collected parking lot environmental data is preprocessed, including data cleaning, formatting and normalization, and data organization. Data cleaning is used to remove noise and outliers, and format and normalize the data to facilitate data learning and processing of the neural network model. The first real-time parking lot environmental data is input into a sequence in chronological order into the neural network model. The neural network model can perform feature extraction and learn vehicle behavior patterns based on the first parking lot environmental data and the driving data of the autonomous driving vehicle, thereby outputting the path planning result of the autonomous driving vehicle.

[0083] In one embodiment, the neural network model is trained based on the driving data of the autonomous driving vehicle and the parking environment data of different parking lots, specifically including: collecting historical driving data of the autonomous driving vehicle and parking environment data of different parking lots, and constructing a training data set based on the historical driving data and the parking environment data; constructing an initial neural network model; wherein the initial neural network model includes an input layer, a feature extraction module and an output layer; inputting the training data set into the initial neural network model for iterative training; wherein the input layer is used to transmit the received training data to the feature extraction module, so that the feature extraction module extracts features from the training data and learns the behavior pattern of the autonomous driving vehicle based on the feature extraction results, and the output layer is used to output a path planning result according to the feature extraction result, and the path planning result includes a driving route and driving control instructions.

[0084] In an embodiment of the present invention, historical driving data of the autonomous driving vehicle and parking environment data in different parking lots are collected by laser radar, camera, GPS, inertial navigation system, etc. Environmental data include lane lines, parking spaces, obstacle locations, floor information, traffic signs, etc.; driving speed includes dynamic data such as driving trajectory, speed, acceleration, steering angle, etc. The control instructions of the autonomous driving vehicle are also marked, such as steering angle, speed instruction, braking instruction, etc. These instructions can be used to directly guide the driving behavior of the vehicle. A training data set is constructed by collecting historical driving data, parking environment data and vehicle control instructions. Preferably, the collected parking environment data is also pre-marked with key information such as the location of the parking space, channel width, floor conversion point, etc. Further, when the training data is input into the training data set for iterative training, it is also pre-processed by data cleaning, formatting, organizing the environmental data and vehicle driving data into a training data sequence in chronological order, etc. Then an initial neural network model is constructed. Preferably, in an embodiment of the present invention, a convolutional neural network and a recurrent neural network are combined to build a model. Convolutional neural networks can be used to process image data (such as images captured by a camera), and recurrent neural networks can be used to process time series data (such as vehicle driving trajectories). For the initial neural network model constructed, the input layer is used to receive the preprocessed training data, namely the environmental data and the vehicle driving data. The feature extraction module contains multiple convolutional layers, pooling layers and recurrent layers. The feature extraction module is used to extract the driving characteristics of the vehicle and learn the behavior pattern of the vehicle. Specifically, the convolutional layer is mainly used to process image data and extract local features in the image through convolution operations. In autonomous driving, the convolutional layer can help the model identify key elements such as lane lines and obstacles; the pooling layer is located after the convolutional layer and is used to reduce the dimension of the data while retaining important features. This helps to reduce the computational complexity of the model and improve the robustness of the features; the recurrent layer, such as the LSTM (Long Short-Term Memory Network) layer, is suitable for processing time series data. In autonomous driving, the recurrent layer can capture the temporal dependency of vehicle driving data and help the model predict future driving status. The depth and complexity of the hidden layer are adjusted according to the characteristics of the specific task and data set, and are not limited here. By stacking multiple convolutional layers, pooling layers and recurrent layers, the model can learn more abstract and advanced feature representations. The output layer is the end point of the neural network model. It is responsible for mapping the features learned by the feature extraction module to specific control instructions or path planning results. The output layer outputs the planning results of the driving route, such as a series of coordinate points or driving directions. These planning results can guide the vehicle to drive along the predetermined path. Furthermore, since the model is trained based on the historical driving data of the autonomous driving vehicle, the model can also directly output the vehicle's driving control instructions, such as steering angle, speed, etc. These instructions are directly used to control the vehicle's driving behavior.To ensure safety and accuracy, the output layer can also use a certain activation function to limit the range of output values. By carefully designing the input layer, feature extraction module, and output layer, the neural network model can effectively process complex environmental data and driving data in autonomous driving, and output accurate control instructions or path planning results, thereby achieving safe and efficient autonomous driving functions.

[0085] In one embodiment, the inputting the training data set into the initial neural network for iterative training specifically includes: iteratively training the initial neural network model based on a back propagation algorithm; wherein, the gradient of the initial neural network model at each layer after each training is calculated, and the weight of the initial neural network model at each layer is updated based on a preset optimization algorithm; when the output of the initial neural network model meets the preset evaluation index, the iterative training of the initial neural network model is terminated to obtain the neural network model.

[0086] In an embodiment of the present invention, the chain rule is used to calculate the gradient of each layer, and the gradient is back-propagated from the output layer to the input layer layer by layer. A calculation method that minimizes numerical errors is used to ensure the accuracy of gradient calculation. The weights and biases of the model are updated using optimization algorithms (such as Adam and SGD). Adam combines the advantages of momentum and adaptive learning rate, and uses first-order and second-order moment estimates to adjust the learning rate of each parameter. One sample is used for training each time, and the model parameters are updated in real time. The value of the loss function is gradually reduced by updating the model parameters through multiple iterations. As the training proceeds, the learning rate is gradually reduced so that the model can adjust the parameters more finely in the later stage of training. As an optimization scheme of the present invention, an adaptive learning rate algorithm, such as AdaGrad and RMSprop, can also be used to dynamically adjust the learning rate according to the historical information of the gradient. Further, when the output of the initial neural network model meets the requirements of evaluation indicators such as accuracy, recall, F1 score, average time, driving distance and safety, it is determined that the model has been trained and can adapt to different parking environments and driving conditions, and the training of the model is terminated and the trained neural network model is deployed on the autonomous driving vehicle. Among them, accuracy, recall rate and F1 score are used to evaluate the classification performance of the model and the accuracy of path planning; driving time is used to evaluate the average driving time of the vehicle from the starting point to the end point; driving distance is used to evaluate the driving distance of the vehicle; safety is evaluated by indicators such as collision detection and number of extreme operations.

[0087] Step 103: Based on the path planning result, control the autonomous driving vehicle to perform autonomous driving until the autonomous driving vehicle reaches the target location.

[0088] In one embodiment, controlling the autonomous driving vehicle to perform autonomous driving based on the path planning result until the autonomous driving vehicle reaches the target location specifically includes: controlling the driving trajectory of the autonomous driving vehicle based on the driving route; and distributing driving control instructions to each control system of the autonomous driving vehicle so that each control system adjusts the driving behavior based on the received driving control instructions.

[0089] In an embodiment of the present invention, the driving route in the path planning result is parsed into a series of coordinate points or driving instructions. These coordinate points or instructions include the starting point, end point, road nodes passed, driving direction, etc. of the vehicle. The driving control instructions in the path planning result are parsed into specific control parameters, such as steering angle, speed instruction, braking instruction, etc. These instructions directly guide the driving behavior of the vehicle. The parsed driving route and driving control instructions are distributed to various control systems of the vehicle, such as steering control system, speed control system, braking control system, etc. Each control system makes corresponding adjustments according to the received instructions. Exemplarily, according to the received steering angle instruction, the steering mechanism of the vehicle is adjusted to ensure that the vehicle travels along the predetermined path. The steering control system realizes precise control of the steering angle through electronic power steering (EPS) or wire steering (SBW); according to the received speed instruction, the throttle or brake system of the vehicle is adjusted to control the driving speed of the vehicle. The speed control system can use adaptive cruise control (ACC) or wire throttle (SBW) to realize precise control of speed; when it is necessary to stop or slow down, according to the received braking instruction, the braking system of the vehicle is controlled to ensure that the vehicle stops safely. The brake control system can use the electronic stability program (ESP) or brake-by-wire (SBW) to achieve precise control of the brakes. Through detailed analysis and distribution of path planning results, execution control of autonomous vehicles, environmental perception and dynamic adjustment, safety and reliability assurance, target location arrival and task completion processing, as well as continuous optimization and iteration, it can ensure that autonomous vehicles can travel safely and efficiently in various environments until they reach the target location. Through data-driven and user feedback, the performance and reliability of the system are continuously improved to provide users with a safe and convenient autonomous driving experience.

[0090] In an embodiment of the present invention, an automatic pick-up device is also provided, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the above-mentioned automatic pick-up method is implemented.

[0091] In an embodiment of the present invention, a computer-readable storage medium is further provided. The computer-readable storage medium includes a stored computer program. When the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the above-mentioned automatic pick-up method.

[0092] Exemplarily, the computer program may be divided into one or more modules, one or more modules are stored in a memory and executed by a processor to implement the present invention. One or more modules may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program in the automatic pick-up device.

[0093] The automatic pick-up device may be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The automatic pick-up device may include, but is not limited to, a processor, a memory, and a display. Those skilled in the art will appreciate that the above components are merely examples of the automatic pick-up device and do not constitute a limitation on the automatic pick-up device. The automatic pick-up device may include more or fewer components than the components, or a combination of certain components, or different components. For example, the automatic pick-up device may also include input and output devices, network access devices, buses, and the like.

[0094] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the automatic reception device, and uses various interfaces and lines to connect the various parts of the entire automatic reception device.

[0095] The memory can be used to store computer programs and / or modules. The processor realizes various functions of the automatic pick-up device by running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function (such as a sound playback function, a text conversion function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, text message data, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0096] Among them, if the module based on automatic pick-up 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 present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. Computer-readable media can include: any entity or device that can carry computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0097] The embodiment of the present invention provides an automatic pick-up method, which provides accurate parking space coordinates and environmental features through a pre-built high-precision map to ensure that the autonomous driving vehicle can accurately identify the parking location. The user only needs to send a driving request through the terminal device, and the autonomous driving vehicle can automatically start and go to the designated location. There is no need for the user to remember the parking space number or surrounding landmarks, which significantly reduces the user's memory burden. The autonomous driving vehicle has real-time environmental perception capabilities and can detect environmental changes such as temporary obstacles, road construction or obstruction of signs in the parking lot. The automatic pick-up method can dynamically adjust the driving route according to the real-time perception results to ensure that the vehicle can avoid obstacles and reach the target location safely. The path planning algorithm based on the high-precision map can generate the optimal driving route in real time, adapt to environmental changes, and ensure the smoothness and safety of the driving process. The automatic pick-up method can quickly respond to the user's driving request, quickly plan the best route and execute it, significantly shortening the time to find a car and improving the efficiency of finding a car.

[0098] Example 2

[0099] See also Figure 2 , Figure 2 Schematic diagram of a module of an automatic pick-up device provided in an embodiment of the present invention. The present invention provides an automatic pick-up device, including: a request receiving module 201, a route planning module 202 and a driving control module 203;

[0100] The request receiving module 201 is used to start the autonomous driving vehicle based on a driving request sent by a user terminal; wherein the driving request includes a target location of the user, and the user terminal and the autonomous driving vehicle establish communication in advance through the Internet of Things;

[0101] The route planning module 202 is used to control the autonomous driving vehicle to call a pre-built high-precision map based on the driving request, perform route screening on the high-precision map based on the parking position of the autonomous driving vehicle and the target position, and generate a path planning result; wherein the high-precision map includes multiple parking lot maps including the parking lot where the parking position is located;

[0102] The driving control module 203 is used to control the autonomous driving vehicle to perform autonomous driving based on the path planning result until the autonomous driving vehicle reaches the target location.

[0103] In one embodiment, the automatic pick-up device also includes patrol training of the autonomous driving vehicle in different types of parking lots, and constructing the high-precision map based on the patrol training results, specifically: controlling the autonomous driving vehicle to perform autonomous driving training in different types of parking lots; wherein the autonomous driving vehicle is equipped with sensors and a positioning system; collecting parking lot environment data of each autonomous driving training process based on the sensors, performing global positioning of the autonomous driving vehicle in each autonomous driving training based on the positioning system, and collecting driving data of the autonomous driving vehicle; performing performance evaluation on the driving data based on preset performance evaluation indicators, and ending the autonomous driving training of the parking lot when each of the preset performance evaluation indicators reaches a preset standard; obtaining the parking lot environment data collected for each autonomous driving training, performing element annotation on the parking lot environment data, and constructing the high-precision map based on the parking lot environment data after element annotation.

[0104] In one embodiment, the controlling of the autonomous driving vehicle to call a pre-constructed high-precision map based on the driving request specifically includes: obtaining the real-time location information and real-time parking lot environment data of the autonomous driving vehicle; performing preliminary matching on the pre-constructed high-precision map based on the real-time parking lot environment data, obtaining a first high-precision map corresponding to the parking position and calling the first high-precision map; performing feature matching on the first high-precision map based on the real-time parking lot environment data, and generating the parking position according to the feature matching result of the real-time parking lot environment data on the first high-precision map.

[0105] In one embodiment, the route screening is performed on the high-precision map based on the parking position of the autonomous driving vehicle and the target position to generate a path planning result, which specifically includes: preprocessing the real-time parking lot environment data to obtain first real-time parking lot environment data; arranging the first real-time parking lot environment data in time series to obtain a first real-time parking lot environment data sequence, and inputting the first real-time parking lot environment data sequence into a pre-constructed neural network model, so that the neural network model performs route planning based on the called first high-precision map to generate the path planning result; wherein the neural network model is trained based on the historical driving data of the autonomous driving vehicle and the parking lot environment data of different parking lots.

[0106] In one embodiment, the neural network model is trained based on the driving data of the autonomous driving vehicle and the parking environment data of different parking lots, specifically including: collecting historical driving data of the autonomous driving vehicle and parking environment data of different parking lots, and constructing a training data set based on the historical driving data and the parking environment data; constructing an initial neural network model; wherein the initial neural network model includes an input layer, a feature extraction module and an output layer; inputting the training data set into the initial neural network model for iterative training; wherein the input layer is used to transmit the received training data to the feature extraction module, so that the feature extraction module extracts features from the training data and learns the behavior pattern of the autonomous driving vehicle based on the feature extraction results, and the output layer is used to output a path planning result according to the feature extraction result, and the path planning result includes a driving route and driving control instructions.

[0107] In one embodiment, the inputting the training data set into the initial neural network for iterative training specifically includes: iteratively training the initial neural network model based on a back propagation algorithm; wherein, the gradient of the initial neural network model at each layer after each training is calculated, and the weight of the initial neural network model at each layer is updated based on a preset optimization algorithm; when the output of the initial neural network model meets the preset evaluation index, the iterative training of the initial neural network model is terminated to obtain the neural network model.

[0108] In one embodiment, controlling the autonomous driving vehicle to perform autonomous driving based on the path planning result until the autonomous driving vehicle reaches the target location specifically includes: controlling the driving trajectory of the autonomous driving vehicle based on the driving route; and distributing driving control instructions to each control system of the autonomous driving vehicle so that each control system adjusts the driving behavior based on the received driving control instructions.

[0109] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0110] The embodiment of the present invention provides a method of providing accurate parking space coordinates and environmental features through a pre-built high-precision map, ensuring that the autonomous driving vehicle can accurately identify the parking location. The user only needs to send a driving request through the terminal device, and the autonomous driving vehicle can automatically start and go to the designated location. There is no need for the user to remember the parking space number or surrounding landmarks, which significantly reduces the user's memory burden. The autonomous driving vehicle has real-time environmental perception capabilities and can detect environmental changes such as temporary obstacles, road construction or obstruction of signs in the parking lot. The automatic pick-up method can dynamically adjust the driving route according to the real-time perception results to ensure that the vehicle can avoid obstacles and safely reach the target location. The path planning algorithm based on the high-precision map can generate the optimal driving route in real time, adapt to environmental changes, and ensure the smoothness and safety of the driving process. The automatic pick-up device can quickly respond to the user's driving request, quickly plan the best route and execute it, significantly shortening the time to find a car and improving the efficiency of finding a car.

[0111] The above are only preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and substitutions can be made without departing from the technical principles of the present invention. These improvements and substitutions should also be regarded as the scope of protection of the present invention.

Claims

1. An automatic pick-up method, characterized in that: include: Starting the autonomous driving vehicle based on a driving request sent by a user terminal; wherein the driving request includes a target location of the user, and the user terminal and the autonomous driving vehicle establish communication in advance through the Internet of Things; Controlling the autonomous driving vehicle to call a pre-built high-precision map based on the driving request, performing route screening on the high-precision map based on the parking position of the autonomous driving vehicle and the target position, and generating a path planning result; wherein the high-precision map includes a plurality of parking lot maps including the parking lot where the parking position is located; Based on the path planning result, the autonomous driving vehicle is controlled to perform autonomous driving until the autonomous driving vehicle reaches the target location.

2. The automatic pick-up method according to claim 1, characterized in that: The automatic pick-up method further includes conducting patrol training of the autonomous driving vehicle in different types of parking lots, and constructing the high-precision map based on patrol training results, specifically: Controlling the autonomous driving vehicle to perform autonomous driving training in different types of parking lots; wherein the autonomous driving vehicle is equipped with sensors and a positioning system; Based on the sensor, the parking environment data of each autonomous driving training process is collected. Based on the positioning system, the autonomous driving vehicle is globally positioned in each autonomous driving training, and driving data of the autonomous driving vehicle is collected; Performing performance evaluation on the driving data based on preset performance evaluation indicators, and ending the autonomous driving training for the parking lot when each of the preset performance evaluation indicators reaches a preset standard; The parking lot environment data collected during each autonomous driving training is acquired, element annotation is performed on the parking lot environment data, and the high-precision map is constructed based on the element-annotated parking lot environment data.

3. An automatic pick-up method as claimed in claim 2, characterized in that: The controlling the autonomous driving vehicle to call a pre-built high-precision map based on the driving request specifically includes: Obtaining real-time location information and real-time parking environment data of the autonomous driving vehicle; Performing a preliminary match on the pre-built high-precision map based on the real-time parking environment data, obtaining a first high-precision map corresponding to the parking location and calling the first high-precision map; Feature matching is performed on the first high-precision map based on the real-time parking environment data, and the parking position is generated according to the feature matching result of the real-time parking environment data in the first high-precision map.

4. An automatic pick-up method as claimed in claim 3, characterized in that: The performing of route screening on the high-precision map based on the parking position of the autonomous driving vehicle and the target position to generate a path planning result specifically includes: Preprocessing the real-time parking lot environment data to obtain first real-time parking lot environment data; The first real-time parking lot environment data is arranged in time series to obtain a first real-time parking lot environment data sequence, and the first real-time parking lot environment data sequence is input into a pre-constructed neural network model so that the neural network model performs route planning based on the called first high-precision map to generate the path planning result; wherein the neural network model is trained based on the historical driving data of the autonomous driving vehicle and the parking lot environment data of different parking lots.

5. The automatic pick-up method according to claim 4, characterized in that: The neural network model is trained based on the driving data of the autonomous driving vehicle and parking environment data of different parking lots, and specifically includes: Collecting historical driving data of the autonomous driving vehicle and parking environment data of different parking lots, and constructing a training data set based on the historical driving data and the parking environment data; Constructing an initial neural network model; wherein the initial neural network model includes an input layer, a feature extraction module and an output layer; The training data set is input into the initial neural network model for iterative training; wherein the input layer is used to transmit the received training data to the feature extraction module, so that the feature extraction module extracts features from the training data and learns the behavior pattern of the autonomous driving vehicle based on the feature extraction results, and the output layer is used to output the path planning results according to the feature extraction results, and the path planning results include the driving route and driving control instructions.

6. An automatic pick-up method as claimed in claim 5, characterized in that: The step of inputting the training data set into the initial neural network for iterative training specifically includes: Iteratively training the initial neural network model based on a back propagation algorithm; wherein, the gradient of the initial neural network model at each layer after each training is calculated, and the weight of the initial neural network model at each layer is updated based on a preset optimization algorithm; When the output of the initial neural network model meets the preset evaluation index, the iterative training of the initial neural network model is terminated to obtain the neural network model.

7. The automatic pick-up method according to claim 5, characterized in that: The controlling the autonomous driving vehicle to perform autonomous driving based on the path planning result until the autonomous driving vehicle reaches the target location specifically includes: Controlling the driving trajectory of the autonomous driving vehicle based on the driving route; The driving control instructions are distributed to each control system of the automatic driving vehicle so that each control system adjusts the driving behavior based on the received driving control instructions.

8. An automatic reception device, characterized in that: include: Request receiving module, route planning module and driving control module; The request receiving module is used to start the autonomous driving vehicle based on a driving request sent by a user terminal; wherein the driving request includes a target location where the user is located, and the user terminal and the autonomous driving vehicle establish communication in advance through the Internet of Things; The route planning module is used to control the autonomous driving vehicle to call a pre-built high-precision map based on the driving request, perform route screening on the high-precision map based on the parking position of the autonomous driving vehicle and the target position, and generate a path planning result; wherein the high-precision map includes multiple parking lot maps including the parking lot where the parking position is located; The driving control module is used to control the autonomous driving vehicle to perform autonomous driving based on the path planning result until the autonomous driving vehicle reaches the target location.

9. A terminal device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the automatic pick-up method as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the automatic pick-up method as described in any one of claims 1 to 7.

Images