A Parking Path Planning Method and System Based on Deep Learning

Abstract

The present invention discloses a parking path planning method and system based on deep learning, which relates to the technical field of deep learning and improves the accuracy and safety of parking. The present invention matches a number of candidate parking spaces, sets a number of candidate parking driving paths for each candidate parking space, obtains the comprehensive scores of each candidate parking driving path, compares the comprehensive scores of each candidate parking driving path, selects the target parking space and the corresponding parking driving path according to the comparison result, discretizes the parking driving path to obtain a number of path nodes, and then determines whether the demand vehicle deviates from the parking driving path according to the real-time driving angle and the current driving speed of the demand vehicle when passing through each path node. When the demand vehicle reaches the target parking space, a multimodal interaction instruction is generated and executed until it is determined that the demand vehicle has completed parking.

Description

Technical Field

[0001] The present invention relates to the technical field of deep learning, and specifically to a parking path planning method and system based on deep learning. Background Art

[0002] With the acceleration of the urbanization process and the continuous growth of the car ownership, the problem of parking difficulty has become increasingly prominent. Traditional parking methods mainly rely on the experience and skills of drivers. When looking for a suitable parking space and performing parking operations, it often takes a lot of time and effort. Especially in areas with concentrated parking demands such as large parking lots and commercial centers, this problem is particularly prominent.

[0003] In addition, some existing parking assistance systems have relatively single functions. Most of them can only provide simple reverse images or distance prompts, and cannot comprehensively and intelligently solve many problems in the parking process. For example, they cannot accurately match a suitable parking space for the driver according to the real-time position and scene information of the vehicle, and it is also difficult to plan an efficient and safe parking driving path. Moreover, there is a lack of real-time monitoring and adjustment of the vehicle driving state during the parking process, which easily leads to deviation of the parking path, increasing the difficulty and time cost of parking. Therefore, a parking path planning method and system based on deep learning are provided. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of the present invention is to provide a parking path planning method and system based on deep learning.

[0005] In order to achieve the above purpose, the present invention provides the following technical solutions:

[0006] A parking path planning method based on deep learning, comprising the following steps:

[0007] Step S1: Collect multi-source scene data of the parking scene through a multi-source data collection device, and establish a visual parking scene model according to the multi-source scene data;

[0008] Step S2: Obtain the real-time position information and vehicle information of the required vehicle and input them into the visual parking scene model, then match several candidate parking spaces, set several candidate parking driving paths for each candidate parking space, obtain the comprehensive scores of each candidate parking driving path, compare the comprehensive scores of each candidate parking driving path, and select multiple candidate parking spaces and corresponding candidate parking driving paths according to the comparison result and send them to the required vehicle;

[0009] Step S3: The demand vehicle selects a target parking space and the corresponding parking driving path, discretizes the parking driving path to obtain a number of path nodes, and sets an expected arrival time, an allowable speed range, and a pre-steering angle range for each path node. Then, according to the real-time driving angle and the current driving speed of the demand vehicle when passing through each path node, it is determined whether the demand vehicle deviates from the parking driving path;

[0010] Step S4: When the demand vehicle arrives at the target parking space, a multi-modal interaction instruction is generated and executed until it is determined that the demand vehicle has completed parking.

[0011] Furthermore, several multi-source data collection devices and an MEC server are installed in the parking scenario. The multi-source data collection devices include cameras, lidar, ultrasonic sensors, and millimeter-wave radars, and a parking space detection device is set for each parking space. The parking space detection device includes a GPS positioning device, a camera, and a wireless communication device.

[0012] Furthermore, the collection process of the multi-source scenario data in the parking scenario includes:

[0013] A scenario data collection period is set for each multi-source data collection device. Whenever a data collection period starts, the camera in each multi-source data collection device is used to capture the optical image data of the parking scenario, the lidar generates the three-dimensional point cloud data of the parking scenario, the ultrasonic sensor detects the distance information of the obstacles in the parking scenario, and the millimeter-wave radar obtains the motion trajectory of the dynamic target in real time.

[0014] Furthermore, the establishment process of the visual parking scenario model includes:

[0015] A three-dimensional parking scenario model is established based on the optical image data of the parking scenario, and the three-dimensional parking scenario model and the three-dimensional point cloud data are split into several data display cubes of the same size. Then, each data display cube is matched with each other, and the three-dimensional parking scenario model and the three-dimensional point cloud data are fused according to the matching result to obtain the visual parking scenario model;

[0016] The distance information of the obstacles, the motion trajectory of the dynamic target, and the position information of the parking space are mapped onto the visual parking scenario model, and several scenario semantic objects are segmented in the visual parking scenario model through a deep learning algorithm.

[0017] Furthermore, the selection process of the candidate parking space includes:

[0018] A GPS positioning device, an inertial measurement unit, and a wireless communication device are installed on the required vehicle. When the required vehicle enters the parking scenario, the MEC server automatically communicates with the wireless communication device of the required vehicle, and at the same time, the vehicle information, real-time position information, and real-time heading angle of the required vehicle are automatically uploaded through the GPS positioning device and the inertial measurement unit.

[0019] According to the vehicle information, real-time position information, and real-time heading angle of the required vehicle, a corresponding scene semantic object is automatically generated in the visual parking scenario model, and the information of the required vehicle is marked on the scene semantic object.

[0020] According to the currently idle parking spaces in the visual parking scenario model and their corresponding length and width values, and in combination with the size information of the required vehicle, the parking spaces that meet the condition of being larger than the length and width of the required vehicle are screened through the greedy algorithm, and the screened parking spaces are recorded as candidate parking spaces.

[0021] Further, the process of setting the pending parking driving path and obtaining the corresponding comprehensive score includes:

[0022] Based on the spatial position of the candidate parking spaces in the visual parking scenario model and the real-time position information of the required vehicle, several pending parking driving paths are preset for each candidate parking space.

[0023] Set the unit driving speed, and then simulate the driving of each pending parking driving path and the parking driving path being executed in the parking scenario according to the unit driving speed.

[0024] Select the pending parking driving paths without collision according to the driving simulation results, and record them as the candidate parking driving paths of each candidate parking space.

[0025] Obtain the comprehensive scores of each candidate parking driving path of each candidate parking space.

[0026] Sort the candidate parking driving paths of all candidate parking spaces in descending order according to the comprehensive scores, and then the MEC server selects the ten candidate parking spaces with the highest comprehensive scores and the corresponding candidate parking driving paths and sends them to the required vehicle.

[0027] Further, the calculation formula of the comprehensive score is:

[0028] ;

[0029] Where 、 、 are the efficiency calculation weights, and , P and T represent the number of obstacles encountered in the candidate parking driving path and the estimated time required to travel at the unit driving speed, and n, θ, and L respectively represent the number of turns, the average turning angle at each turning point, and the total length of the candidate parking driving path.

[0030] Further, the process of determining whether the demand vehicle deviates from the parking driving path includes:

[0031] The demand vehicle selects a target parking space and the corresponding parking driving path from the candidate parking spaces and the corresponding candidate parking driving paths sent by the MEC server;

[0032] Set the unit detection distance, divide the parking driving path into multiple path nodes according to the unit detection distance, and set the expected arrival time, allowable speed range, and pre-steering angle range for each path node. It should be noted that the allowable speed range and the pre-steering angle range of different path nodes are different, and there are path nodes without a pre-steering angle range;

[0033] During the process that the demand vehicle passes through each path node in sequence according to the parking driving path, the MEC server synchronously updates the visual parking scene model. If it is determined that a new obstacle appears or the vehicle deviates from the predicted path in any current un-traveled path node of the demand vehicle, then plan a parking driving path for the corresponding demand vehicle, otherwise do nothing;

[0034] At the same time, when the demand vehicle passes through each path node, obtain its real-time driving angle and current driving speed through the inertial measurement unit on the demand vehicle, and then determine whether the current driving speed of the demand vehicle at each path node is within the allowable speed range, and whether the real-time driving angle at the turning point of the path node is within the pre-steering angle range;

[0035] If it is determined that either the real-time driving angle or the current driving speed is not within the corresponding range, it is determined that the corresponding demand vehicle is driving abnormally, and a corresponding speed or steering angle repair decision is generated through the PID control algorithm until it is determined that both the real-time driving angle and the current driving speed are within the corresponding ranges, otherwise do nothing;

[0036] Repeat the above process of determining whether the demand vehicle deviates from the parking driving path until the demand vehicle reaches the target parking space.

[0037] Further, after the demand vehicle reaches the target parking space, superimpose the demand vehicle and the corresponding scene semantic objects of the target parking space in the visual parking scene model, and then generate a multi-modal interaction instruction to guide the demand vehicle to park;

[0038] When it is determined that the distances between both sides of the demand vehicle and the parking space boundary lines of the target parking space are both between 0.2 and 0.5 m, and the extension angle between the head or tail of the demand vehicle and the tail line of the target parking space does not exceed plus or minus three to five degrees, and the vertical distance is between 0.3 and 0.5 m, it is determined that the demand vehicle has completed parking; otherwise, the multimodal interaction instruction is continuously executed.

[0039] A parking path planning system based on deep learning includes a scene data acquisition module, a path planning module, a driving supervision module, and a parking guidance module.

[0040] The scene data acquisition module is used to collect multi-source scene data of the parking scene by setting up multi-source data acquisition devices, and establish a visual parking scene model based on the multi-source scene data.

[0041] The path planning module is used to obtain the real-time position information and vehicle information of the demand vehicle, input them into the visual parking scene model to match candidate parking spaces, set several candidate parking driving paths for each candidate parking space, and obtain the comprehensive scores of each candidate parking driving path.

[0042] The driving supervision module is used to select the target parking space and the corresponding parking driving path, discretize the parking driving path to obtain several path nodes, and set the expected arrival time, allowable speed range, and pre-steering angle range for each path node respectively. Then, according to the real-time driving angle and current driving speed of the demand vehicle when passing through each path node, it is determined whether the demand vehicle deviates from the parking driving path.

[0043] The parking guidance module is used to generate and execute multimodal interaction instructions, and determine in real time whether the demand vehicle has completed parking.

[0044] Compared with the prior art, the beneficial effects of the present invention are:

[0045] By setting the expected arrival time, allowable speed range, and pre-steering angle range of each path node, and determining whether the path deviates according to the real-time driving angle and speed of the vehicle. Furthermore, while promptly detecting and correcting the deviation of the vehicle during parking and ensuring that the vehicle parks accurately according to the planned path, the accuracy and safety of parking are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention.

[0047] Figure 1 It is a flowchart of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0048] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be described in detail below. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other implementation manners obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope protected by the present invention.

[0049] As Figure 1 shown, a parking path planning method based on deep learning includes the following steps:

[0050] Step S1: Collect multi-source scenario data of the parking scenario through a multi-source data collection device, and establish a visual parking scenario model based on the multi-source scenario data;

[0051] Step S2: Obtain the real-time position information and vehicle information of the required vehicle and input them into the visual parking scenario model, then match a number of candidate parking spaces, set a number of candidate parking driving paths for each candidate parking space, obtain the comprehensive scores of each candidate parking driving path, compare the comprehensive scores of each candidate parking driving path, and select multiple candidate parking spaces and corresponding candidate parking driving paths according to the comparison result and send them to the required vehicle;

[0052] Step S3: The required vehicle selects a target parking space and the corresponding parking driving path, discretizes the parking driving path to obtain a number of path nodes, and sets an expected arrival time, an allowable speed range, and a pre-steering angle range for each path node, and then determines whether the required vehicle deviates from the parking driving path according to the real-time driving angle and the current driving speed when the required vehicle passes through each path node;

[0053] Step S4: Generate and execute a multi-modal interaction instruction when the required vehicle reaches the target parking space until it is determined that the required vehicle has completed parking.

[0054] The above Step S1 is implemented through the following process:

[0055] Install a number of multi-source data collection devices and an MEC server in the parking scenario. The multi-source data collection device includes a camera, a lidar, an ultrasonic sensor, and a millimeter-wave radar, and a parking space detection device is set for each parking space. The parking space detection device includes a GPS positioning device, a camera, and a wireless communication device;

[0056] The MEC server is built with a local area network, and the signal coverage range includes the entire parking scenario;

[0057] It should be noted that the data acquisition ranges of the respective multi-source data acquisition devices are the same, and there is an overlap in the data acquisition ranges of the multi-source data acquisition devices at adjacent spatial positions;

[0058] Set a scenario data acquisition period for each multi-source data acquisition device, where the duration of the scenario data acquisition period is generally between 50 and 100 ms;

[0059] Whenever a data acquisition period starts, the cameras in the respective multi-source data acquisition devices are used to capture optical image data of the parking scenario, the lidar generates three-dimensional point cloud data of the parking scenario, the ultrasonic sensors detect the distance information of obstacles in the parking scenario, and the millimeter-wave radar obtains the motion trajectories of dynamic targets in real time;

[0060] Set up a scenario fusion base based on the floor area of the parking scenario, and establish a three-dimensional parking scenario model according to the optical image data of the parking scenario. Then, first map the three-dimensional parking scenario model and the three-dimensional point cloud data to the scenario fusion base. Since both the three-dimensional parking scenario model and the three-dimensional point cloud data are generated corresponding to the parking scenario, there is a one-to-one correspondence between the scenario information contained in the two;

[0061] Furthermore, split the three-dimensional parking scenario model and the three-dimensional point cloud data into several data display cubes of the same size, and then match the respective data display cubes with each other. According to the matching results, fuse the three-dimensional parking scenario model and the three-dimensional point cloud data to obtain a visualized parking scenario model;

[0062] Map the distance information of obstacles, the motion trajectories of dynamic targets, and the position information of parking spaces to the visualized parking scenario model, and segment several scenario semantic objects in the visualized parking scenario model through a deep learning algorithm;

[0063] The scenario semantic object represents a local scenario model with semantic marks in the visualized parking scenario model, and the semantic marks include obstacle names, dynamic target identifiers, and parking space identifiers;

[0064] It should be noted that for the scenario semantic object corresponding to the parking space, it is marked with whether the parking space is currently idle and the length and width values.

[0065] The step S2 is implemented through the following process:

[0066] A GPS positioning device, an inertial measurement unit, and a wireless communication device are installed on the demand vehicle. When the demand vehicle enters the parking scenario, the MEC server automatically communicates and connects with the wireless communication device of the demand vehicle, and at the same time automatically uploads the vehicle information, real-time position information, and real-time heading angle of the demand vehicle through the GPS positioning device and the inertial measurement unit;

[0067] The vehicle information includes vehicle size, minimum turning radius and current load status;

[0068] According to the vehicle information, real-time location information and real-time heading angle of the required vehicle, a corresponding scene semantic object is automatically generated in the visual parking scene model, and various information of the required vehicle is marked on the scene semantic object;

[0069] According to the currently idle parking spaces and the corresponding length and width values ​​in the visual parking scene model, combined with the size information of the required vehicle, a greedy algorithm is used to select parking spaces that are larger than the length and width of the required vehicle, and the selected parking spaces are recorded as candidate parking spaces;

[0070] Based on the spatial position of the candidate parking space in the visualized parking scene model and the real-time position information of the required vehicle, a number of pending parking driving paths are preset for each candidate parking space;

[0071] A unit driving speed is set, and then each pending parking driving path and the parking driving path being executed in the parking scene are simulated according to the unit driving speed;

[0072] Selecting a pending parking driving path without collision according to the driving simulation result and recording it as a candidate parking driving path for each candidate parking space;

[0073] The comprehensive score of each candidate parking driving path of each candidate parking space is calculated, and the calculation formula of the comprehensive score is:

[0074] ;

[0075] in , , Calculate the weights for efficiency, and , P and T represent the number of obstacles in the candidate parking path and the estimated time required for driving at a unit speed, n, θ and L represent the number of turns, the average turning angle of each turn and the total length of the candidate parking path respectively;

[0076] The candidate parking paths of all candidate parking spaces are sorted in descending order according to their comprehensive scores, and then the MEC server selects the ten candidate parking spaces with the highest comprehensive scores and the corresponding candidate parking paths and sends them to the required vehicle.

[0077] The step S3 is implemented by the following process:

[0078] The demand vehicle selects the target parking space and the corresponding parking driving path from the candidate parking spaces and the corresponding candidate parking driving paths sent by the MEC server;

[0079] Set the unit detection distance, divide the parking driving path into multiple path nodes according to the unit detection distance, and set the expected arrival time, allowable speed range, and pre-steering angle range for each path node. It should be noted that the allowable speed range and pre-steering angle range of different path nodes are different, and there are path nodes without a pre-steering angle range;

[0080] During the process that the demand vehicle passes through each path node in sequence according to the parking driving path, the MEC server synchronously updates the visual parking scene model. If it is judged that a new obstacle appears or the vehicle deviates from the predicted path at the current un-traveled path node of any demand vehicle, then plan a parking driving path for the corresponding demand vehicle, otherwise do nothing;

[0081] At the same time, when the demand vehicle passes through each path node, obtain its real-time driving angle and current driving speed through the inertial measurement unit on the demand vehicle, and then judge whether the current driving speed of the demand vehicle at each path node is within the allowable speed range, and whether the real-time driving angle at the path node corner is within the pre-steering angle range;

[0082] If it is judged that either the real-time driving angle or the current driving speed is not within the corresponding range, then judge that the corresponding demand vehicle is driving abnormally, and generate a corresponding speed or steering angle repair decision through the PID control algorithm until it is judged that both the real-time driving angle and the current driving speed are within the corresponding range, otherwise do nothing;

[0083] Repeat the above process of judging whether the demand vehicle deviates from the parking driving path until the demand vehicle reaches the target parking space.

[0084] The step S4 is realized through the following process:

[0085] After the demand vehicle reaches the target parking space, superimpose the demand vehicle and the corresponding scene semantic objects of the target parking space in the visual parking scene model, and then generate a multi-modal interaction instruction to guide the demand vehicle for parking;

[0086] The multi-modal interaction instruction includes:

[0087] Visual prompt: Real-time obtain the image data of the demand vehicle relative to the target parking space through the camera in the parking space detection device of the target parking space, and display its relative position with the target parking space on the on-vehicle display screen of the demand vehicle, and mark the recommended steering wheel steering angle;

[0088] Voice prompt: Broadcast real-time operation instructions, such as "fine-tune 15 degrees to the left" or "slow down to 2 km / h";

[0089] Automatic control instruction: If the vehicle in demand is equipped with a steer-by-wire chassis, directly control the steering motor and the drive motor to perform the parking action; otherwise, automatically skip the instruction.

[0090] When it is judged that the distances between both sides of the vehicle in demand and the parking space boundary lines of the target parking space are both between 0.2 m and 0.5 m, and the extension angle between the head or tail of the vehicle in demand and the tail line of the target parking space does not exceed plus or minus three to five degrees, and the vertical distance is between 0.3 m and 0.5 m, it is judged that the vehicle in demand has completed parking; otherwise, continuously execute the multi-modal interaction instruction.

[0091] The present invention also discloses a parking path planning system based on deep learning, including a scene data acquisition module, a path planning module, a driving supervision module, and a parking guidance module.

[0092] The scene data acquisition module is used to collect multi-source scene data of the parking scene by setting a multi-source data acquisition device, and establish a visual parking scene model according to the multi-source scene data.

[0093] The path planning module is used to obtain the real-time position information and vehicle information of the vehicle in demand, input them into the visual parking scene model to match candidate parking spaces, set several candidate parking driving paths for each candidate parking space, and obtain the comprehensive scores of each candidate parking driving path.

[0094] The driving supervision module is used to select a target parking space and the corresponding parking driving path, discretize the parking driving path to obtain several path nodes, and respectively set the expected arrival time, allowable speed range, and pre-steering angle range for each path node. Furthermore, according to the real-time driving angle and the current driving speed of the vehicle in demand when passing through each path node, it is judged whether the vehicle in demand deviates from the parking driving path.

[0095] The parking guidance module is used to generate and execute multi-modal interaction instructions, and judge in real time whether the vehicle in demand has completed parking.

[0096] The above content is only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A parking path planning method based on deep learning, characterized in that: The following steps are involved: Step S1: collecting multi-source scene data of a parking scene by setting a multi-source data collection device, and establishing a visual parking scene model according to the multi-source scene data; Step S2: obtaining the real-time location information and vehicle information of the required vehicle and inputting them into the visual parking scene model, thereby matching a number of candidate parking spaces, and setting a number of candidate parking driving paths for each candidate parking space, obtaining a comprehensive score of each candidate parking driving path, and comparing the comprehensive scores of each candidate parking driving path, selecting a number of candidate parking spaces and corresponding candidate parking driving paths according to the comparison results and sending them to the required vehicle; Step S3: the target parking space and the corresponding parking driving path are selected by the demand vehicle, the parking driving path is discretized to obtain a number of path nodes, and the expected arrival time, the allowed speed range and the pre-steering angle range are respectively set for each path node, and then according to the real-time driving angle and the current driving speed of the demand vehicle when passing each path node, it is determined whether the demand vehicle deviates from the parking driving path; Step S4: when the required vehicle arrives at the target parking space, a multimodal interaction instruction is generated and executed until it is determined that the required vehicle has completed parking; The process of establishing the visual parking scene model includes: A 3D parking scene model is established based on the optical image data of the parking scene, and the 3D parking scene model and the 3D point cloud data are split into a number of data display cubes of the same size, and then the data display cubes are matched with each other, and the 3D parking scene model and the 3D point cloud data are fused according to the matching results to obtain a visualized parking scene model; The distance information of obstacles, the motion trajectory of dynamic targets, and the location information of parking spaces are mapped to the visual parking scene model, and a number of scene semantic objects are segmented in the visual parking scene model through a deep learning algorithm.

2. The parking path planning method based on deep learning according to claim 1, characterized in that: The parking scene is equipped with several multi-source data acquisition devices and a MEC server.

3. The parking path planning method based on deep learning according to claim 2, characterized in that: The process of collecting multi-source scene data for parking scenes includes: A scene data acquisition cycle is set for each multi-source data acquisition device. When a data acquisition cycle starts, optical image data, three-dimensional point cloud data, distance information of obstacles, motion trajectory of dynamic targets and location information of parking spaces of the parking scene are collected.

4. The parking path planning method based on deep learning according to claim 3, characterized in that: The process of selecting the candidate parking space includes: The required vehicle is equipped with a GPS positioning device, an inertial measurement unit and a wireless communication device. When the required vehicle enters the parking scene, the MEC server automatically communicates with the wireless communication device of the required vehicle, and automatically uploads the vehicle information, real-time location information and real-time heading angle of the required vehicle through the GPS positioning device and the inertial measurement unit; According to the vehicle information, real-time location information and real-time heading angle of the required vehicle, a scene semantic object is automatically generated in the visual parking scene model, and various information of the required vehicle is marked on the scene semantic object; According to the currently idle parking spaces and the corresponding length and width values ​​in the visual parking scene model, combined with the size information of the required vehicle, a greedy algorithm is used to select parking spaces that are larger than the length and width of the required vehicle, and the selected parking spaces are recorded as candidate parking spaces.

5. The parking path planning method based on deep learning according to claim 4, characterized in that: The process of setting a pending parking driving route and obtaining the corresponding comprehensive score includes: Based on the spatial position of the candidate parking space in the visualized parking scene model and the real-time position information of the required vehicle, a number of pending parking driving paths are preset for each candidate parking space; A unit driving speed is set, and then each pending parking driving path and the parking driving path being executed in the parking scene are simulated according to the unit driving speed; Selecting a pending parking driving path without collision according to the driving simulation result and recording it as a candidate parking driving path for each candidate parking space; The comprehensive scores of the candidate parking paths of the candidate parking spaces are obtained, and the candidate parking paths of all the candidate parking spaces are sorted from high to low according to the comprehensive scores, and the ten candidate parking spaces with the highest comprehensive scores and the corresponding candidate parking paths are selected and sent to the demand vehicle.

6. The parking path planning method based on deep learning according to claim 5, characterized in that: The process of determining whether the required vehicle deviates from the parking path includes: The vehicle in need selects the target parking space and the corresponding parking driving path, sets the unit detection distance, divides the parking driving path into multiple path nodes according to the unit detection distance, and sets the expected arrival time, allowed speed range and pre-steering angle range for each path node; When the demand vehicle passes through each path node in sequence according to the parking driving path, the MEC server synchronously updates the visual parking scene model. If it is determined that there are new obstacles or deviations from the predicted path at the current untraveled path node of any demand vehicle, the parking driving path is planned for the corresponding demand vehicle, otherwise no operation is performed; At the same time, when the required vehicle passes each path node, its real-time driving angle and current driving speed are obtained through the inertial measurement unit on the required vehicle, so as to judge whether the current driving speed of the required vehicle at each path node is within the allowed speed range, and whether the real-time driving angle at the corner of the path node is within the pre-steering angle range; If it is determined that either the real-time driving angle or the current driving speed is not within the corresponding range, the corresponding vehicle is judged to be driving abnormally, otherwise no operation is performed.

7. The parking path planning method based on deep learning according to claim 6, characterized in that: When the required vehicle reaches the target parking space, the scene semantic objects corresponding to the required vehicle and the target parking space are superimposed in the visual parking scene model, and then multimodal interactive instructions are generated to guide the required vehicle in parking until it is determined that the required vehicle has completed parking.

8. A parking path planning system based on deep learning, used to implement a parking path planning method based on deep learning according to any one of claims 1 to 7, characterized in that: It includes scene data collection module, path planning module, driving supervision module and parking guidance module; The scene data acquisition module is used to collect multi-source scene data of the parking scene by setting a multi-source data acquisition device, and establish a visual parking scene model according to the multi-source scene data; The path planning module is used to obtain the real-time location information and vehicle information of the required vehicle, and input them into the visual parking scene model to match the candidate parking spaces, set a number of candidate parking driving paths for each candidate parking space, and obtain the comprehensive score of each candidate parking driving path; The driving supervision module is used to select a target parking space and a corresponding parking driving path, discretize the parking driving path to obtain a number of path nodes, and set an expected arrival time, an allowable speed range, and a pre-steering angle range for each path node, and then judge whether the required vehicle deviates from the parking driving path according to the real-time driving angle and current driving speed of the required vehicle when passing each path node; The parking guidance module is used to generate and execute multimodal interactive instructions, and to determine in real time whether the required vehicle has completed parking.

Images