Setting method of autonomous parking system for large electric trucks based on field-side collaboration

Abstract

The present invention discloses a method for setting up an autonomous parking system for a large electric truck based on field-side collaboration, and the method includes the following steps: the first step is to carry a vehicle networking platform; the second step is to establish a parking lot management module; the third step is to build an autonomous parking control system; the fourth step is to perform trajectory tracking control; the fifth step is to build a real-time vehicle-side collision avoidance system; the sixth step is to establish an information processing system; the beneficial effect is to improve the efficiency of traffic operation. The system is suitable for the transition stage between semi-intelligent driving and intelligent driving, and is a great leap in technology. The application of this system can not only save the driver's parking time, but also improve people's driving experience. It can effectively alleviate the contradiction between parking supply and demand, realize the maximum utilization of resources, and help people save time. It can reduce the difficulty of driving, improve the industrial structure, and promote the information level of the entire country.

Description

Technical Field

[0001] The present invention relates to a method for setting up an autonomous parking system, and in particular to a method for setting up an autonomous parking system for a large electric truck based on field-side collaboration. Background Art

[0002] At present, autonomous driving vehicles are an intelligent vehicle system that integrates environmental perception, path planning, trajectory tracking, control execution and other functions, also known as unmanned driving. It is an inevitable trend in the development of the current automobile industry and a major area of ​​technological innovation. For large trucks, it has a high degree of operating difficulty and an increasingly high frequency of use. At present, there are more and more intelligent parking lots for large trucks at home and abroad. The path in the large intelligent parking lot scene has the characteristics of low speed, fewer traffic participants, and controllable safety. It is considered to be one of the key scenes for the first realization of autonomous driving industrialization. In the large parking lot scene, the autonomous driving vehicle can carry out autonomous path planning with the information provided by the parking lot, and then complete autonomous parking according to its own perception system. How to reasonably complete efficient parking for large trucks is of great significance to improving the vehicle operation efficiency and management and scheduling of parking lots. However, intelligent parking requires higher power for the vehicle, and the information collection signals are all transmitted in the form of electrical signals, so this solution is currently only for electric trucks.

[0003] Based on the parameters of the electric vehicle itself, it is necessary to control its motion, cooperate with the upper-level machine vision and trajectory planning functions, obtain real-time information, and control the vehicle to achieve reasonable trajectory tracking control. Trajectory tracking control means that in the inertial coordinate system, the mobile unmanned vehicle must start from a given initial state and reach and follow a given target trajectory. The initial point of the unmanned vehicle can be on this trajectory or not. This is an important issue in the field of unmanned vehicle intelligent driving research. At present, the research objects of trajectory tracking problems are mostly wheeled mobile robots, unmanned vehicles and drones.

[0004] The Chinese invention patent application "Unmanned vehicle parking method, device and electronic equipment based on temporary events" (application number: CN 202110810680.3, application publication date: 2021.08.17) discloses an unmanned vehicle parking method based on temporary events, which can trigger the unmanned vehicle parking function based on temporary events, and dynamically adjust the parking position of the unmanned vehicle through comprehensive detection of road environment, traffic rules, obstacles and other constraints through on-board sensors and controllers, so as to improve the interaction ability and environmental response ability between the unmanned vehicle and external users; this patent application mainly solves the parking decision-making problem of unmanned vehicles when temporary events intervene, so as to determine whether to respond to temporary events, whether to stop, how to stop and where to park.

[0005] The Chinese invention patent "An automatic parking method based on unmanned driving and vehicle networking" (patent number ZL201910892986.0) invented an automatic parking method for unmanned vehicles. Through the combination of vehicle networking and unmanned driving technology, the owner can input parking parameter information and select the parking mode on the mobile terminal to realize the automatic parking action of the vehicle; this patent is mainly applicable to manned self-driving vehicles, targeting specific parking lots and vehicles with automatic parking functions, to solve the pain point of customers' difficulty in parking.

[0006] The Chinese invention patent application "An automatic driving parking system for new energy vehicles based on artificial intelligence" (application number: 202011368423.0, application publication number: 2021.02.19) discloses an automatic driving parking system for new energy vehicles based on artificial intelligence. The remaining power detection module is used to detect the remaining power of the new energy vehicle after it arrives at the parking lot, and judge whether the car can reach the destination with the remaining power based on historical data, and select charging and non-charging empty parking spaces based on the judgment results; after finding a suitable empty parking space, the empty space scanning module is used to scan and confirm whether there are obstacles in the empty parking space, and further decide whether to park; this patent application is mainly based on on-board sensors and controllers, which are used to solve the problems of whether the vehicle is charged and actively selecting parking spaces.

[0007] Regarding the above-mentioned patents, most of the existing parking systems for autonomous vehicles are applied to parking scenarios, and do not consider the path navigation of the autonomous vehicle from the location where the parking application is issued to the parking space; most of the existing parking systems for autonomous vehicles are applied to small family cars, which have relatively simple operation conditions and fewer considerations for implementation results, and are not suitable for large electric trucks. Most of the existing parking systems for autonomous vehicles are based on single-vehicle intelligence, that is, they rely on the vehicle's own sensors to locate and identify parking spaces, and do not integrate the information interaction between "vehicle-road-cloud-parking lot", and cannot perform unified parking space scheduling and path planning well. Summary of the invention

[0008] The purpose of the present invention is to solve the problem that most of the existing parking systems for autonomous driving vehicles are based on single-vehicle intelligence, that is, they rely on the vehicle's own sensors to locate and identify parking spaces, and do not integrate the information interaction between "vehicle-road-cloud-parking lot", and cannot perform unified parking space scheduling and path planning well. Instead, a method for setting up an autonomous parking system for large electric trucks based on field-side collaboration is provided.

[0009] The method for setting up an autonomous parking system for a large electric truck based on field-side collaboration provided by the present invention comprises the following steps:

[0010] The first step is to install the Internet of Vehicles platform, which includes an information receiving module, an information sorting module, a high-precision map platform and an information sending module. The information receiving module and the information sending module are connected to the field system and the on-board computer through the network to realize real-time transmission of information and application and acceptance of commands. The information sorting module includes a parking space storage unit and a parking management unit. The information sorting module will process and summarize the parking information of vehicles in the parking lot according to the received information. The parking space storage unit uses the detection of the field support system to obtain the number and location of the vacant parking space. When the vacant parking space is identified and located, the information processing The module stores the parking space information in the parking space database and waits for the vehicle to apply. When the autonomous vehicle issues a parking application command, the parking space recommendation unit uses a sorting algorithm that approximates the ideal solution to obtain the optimal sorting of free parking spaces and recommends the optimal parking space to the autonomous vehicle. The Internet of Vehicles platform will update the parking space database and no longer release the recommended parking spaces to other vehicles. The high-precision map platform is responsible for monitoring the map data of the parking lot. The high-precision map platform locates the parking space uploaded to the Internet of Vehicles platform and the positioning coordinates of the applying vehicle. The Internet of Vehicles platform will send the relevant information to the corresponding vehicle by accessing the pre-stored map.

[0011] The second step is to establish a parking lot management module, which includes a parking space detection unit at the parking lot and a roadside data unit. The parking space detection unit at the parking lot detects the vacancy status and parking space number of the parking space through the parking space detection sensor and pre-processes the parking space signal. The roadside data unit is responsible for receiving the vehicle parking space application command issued by the Internet of Vehicles platform and the parking space information issued by the platform.

[0012] The third step is to build an autonomous parking control system. The autonomous parking control system includes building a vehicle model, performing path planning, and transmitting driving information to the motion execution part. It is the control center of the entire autonomous parking system. The specific process is as follows:

[0013] The vehicle's autonomous parking control system first determines the minimum turning radius of the car based on the vehicle's own properties, such as the vehicle's wheelbase L and the maximum steering angle of the front wheels θmax. Then, based on the radius and the speed and heading angle of the vehicle during movement and the relevant position coordinates, a real-time model of the vehicle is established to control the vehicle. Then, the vehicle plans the path based on its own model and the map information and target parking space information released by the Internet of Vehicles platform. At the same time, collision avoidance needs to be considered. The Hybrid A* algorithm is used. It will gradually perform heuristic search in a continuous coordinate system, expand subnodes, and perform collision detection on the path based on the existing obstacle map. The nodes corresponding to the collision-free path will be added to the expansion tree to generate a large number of possible paths. Finally, the loss function is designed to construct the Voronoi diagram and the conjugate gradient method is used for non-parametric interpolation to improve the quality of the generated path in order to obtain a target path that can effectively avoid obstacles, is smooth and efficient.

[0014] The fourth step is to carry out trajectory tracking control. The vehicle's trajectory tracking control system controls the vehicle to have a suitable speed and steering angle according to relevant calculations, and controls the vehicle to travel along the corresponding path;

[0015] The vehicle's trajectory tracking control system regards the vehicle as a bicycle model with the rear axle as the center under ideal conditions. Then, based on the Pure Pursuit trajectory tracking control method, the sensors collect the vehicle's geometric center position coordinates, the angle between the speed and the horizontal line, the front wheel angle, the angle between the vehicle body and the target point, the forward sight distance and the movement speed at a specific moment in real time, so that the vehicle has an accurate driving route when moving. Finally, the trajectory tracking control system calculates the target values ​​of the current vehicle actuators of throttle, brake and steering according to the input target trajectory, and transmits the target data to the vehicle actuator ECU through CAN. The actuator controls the specific actions of the actuator according to the target value.

[0016] Step 5: Build a real-time vehicle-side collision avoidance system. When the vehicle-side collision avoidance system finds an unavoidable obstacle, it will automatically sound an alarm and take braking or evasive measures to avoid a collision.

[0017] The real-time vehicle-side collision avoidance system is a method for vehicles to avoid collisions or a solution when encountering unavoidable situations. The real-time vehicle-side collision avoidance system includes a signal acquisition module, a data processing module and an execution module. The execution module includes several decision-making execution methods such as starting the alarm, closing the windows and emergency braking. The signal acquisition module collects the real-time status information of the vehicle itself, surrounding vehicles and obstacles and the current road information in real time based on its own vehicle sensors, and continuously transmits the real-time data to the data processing module of the real-time vehicle-side collision avoidance system. The data processing module calculates the critical safety distance for collision avoidance based on its own vehicle speed and the real-time distance between the vehicle and the obstacle, completes the collision avoidance decision based on the current longitudinal distance between the vehicle and the obstacle, and plans the vehicle's decision-making method. If the distance allows braking to avoid collision, the vehicle implements appropriate deceleration to slow down. If the distance is too short to avoid collision, the vehicle will be forced to trigger several decisions such as alarms and emergency braking.

[0018] Step 6: Establishment of the information processing system, which is responsible for collecting relevant information of the vehicle during driving and sending it to various systems for sharing;

[0019] The information processing system consists of three parts: information collection layer, transmission layer and application layer. The collection layer is responsible for collecting the real-time operating parameters of the vehicle, road environment parameters and prediction parameters and converting them into electrical signals. The transmission layer is responsible for connecting various systems and providing information transmission services, that is, realizing the aggregation, analysis, processing and transmission of input and output data. The application layer refers to the integration of various information, the close integration of business logic and resources, and the presentation of information processing results to users in the form of rich multimedia.

[0020] The autonomous driving vehicle in the above steps is equipped with an on-board computer, which is also the control center of the autonomous driving vehicle. The on-board computer has built-in codes for the autonomous parking control system and trajectory tracking control system and the vehicle-side collision avoidance system that are input in advance. The on-board computer is connected to the Internet of Vehicles platform through the network, and the parking lot management module and the Internet of Vehicles platform are also connected through the network. The field sensors inside the parking lot collect vehicle driving information and parking space information in real time and send them to the Internet of Vehicles platform for everyone to see and use. At the same time, the high-precision map platform in the Internet of Vehicles platform will also share the detailed map of the parking lot for vehicles to use. The specific process is as follows:

[0021] After the on-board computer receives the map information and parking space information published by the Internet of Vehicles platform, it shares it with the downlink system. First, the autonomous parking control system plans a reasonable driving path through the pre-established vehicle motion model and the received map information and parking space information. The path is transmitted to the trajectory tracking control system in the computer. The trajectory tracking control system starts the vehicle, and the vehicle-side sensors start running. The data from the combined effect of the on-board camera and the field-side sensors is published on the Internet of Vehicles platform and sent to the on-board computer at the same time to calculate the real-time distance between the vehicle and the obstacle. After the vehicle moves, the vehicle-side sensors monitor the vehicle and collect the The vehicle speed information and steering angle information are converted into a trajectory tracking control system that inputs electrical signals. The trajectory tracking control system is adjusted according to the driving path and position information, and the speed and steering angle at the next moment are calculated, and input to the vehicle through electrical signals. At the same time, the vehicle speed information, steering angle information, the distance between the vehicle and the obstacle, and the location map information are shared with the vehicle-side collision avoidance system. The vehicle-side collision avoidance system runs in real time to calculate whether it is a safe distance. If everything is normal, it continues to drive. If it is less than the safe distance, the vehicle-side collision avoidance system will directly transmit electrical signals to the braking system, motor and alarm switch to force braking, emergency stop of the motor and alarm.

[0022] The work content of the Internet of Vehicles platform is as follows:

[0023] The Internet of Vehicles platform is connected via the network and is used to accept parking instructions from car owners. The information receiving module and information sending module of the Internet of Vehicles platform are simultaneously connected to the unmanned vehicle and the parking lot management module via the network, realizing the connection between the field-side sensors and the vehicle-side sensors, and is used to monitor the vehicle's driving path in all directions and count available parking spaces, obtain information about the vehicle's surrounding environment, and help guide the vehicle to avoid other vehicles and obstacles and drive into the parking space. The Internet of Vehicles platform is used to obtain available parking spaces from the parking lot management module and send them to the vehicle, and is also used to control the vehicles parked in the parking lot in real time;

[0024] The information sorting module in the Internet of Vehicles platform is mainly responsible for the management of vehicles and parking spaces and the optimal recommendation of parking spaces. It includes three parts: parking space identification / positioning unit, parking space recommendation unit and parking management unit. The parking space identification and positioning unit uses artificial intelligence technology to identify parking spaces and obtain the number and location of vacant parking spaces. After the vacant parking spaces are identified and located, the Internet of Vehicles platform stores the parking space information in the server and waits for parking space application commands. When the autonomous driving vehicle issues a parking application command, it is transmitted to the Internet of Vehicles platform through the network layer. The parking space recommendation unit in the Internet of Vehicles platform uses a sorting algorithm that approximates the ideal solution to obtain the optimal sorting of vacant parking spaces and sends the optimal parking space to the autonomous driving vehicle through the network. At the same time, the Internet of Vehicles platform will update the parking space database and will no longer release the recommended parking spaces to other vehicles.

[0025] The high-precision map platform belongs to the downstream part of the Internet of Vehicles platform. It can output lane-level parking lot map data to support low-latency, high-precision path planning. The Internet of Vehicles platform transmits the vehicle's parking application command to the information receiving module, and then the high-precision map platform sends the parking lot's traffic information to the on-board navigation system in real time through the network. The on-board navigation system determines the feasible parking path based on the parking position and the real-time traffic information of the parking lot. Based on the parking space uploaded to the Internet of Vehicles platform and the positioning coordinates of the applying vehicle, the Internet of Vehicles platform will guide the vehicle to automatically drive to the parking space and complete parking by accessing the pre-stored map.

[0026] The parking lot management module built works as follows:

[0027] The parking lot that meets the above steps is a parking lot with the function of guiding and managing unmanned vehicles. In order to rationally utilize parking spaces in the parking lot and reduce waste of resources, the parking lot management system is provided with a field-end parking space detection unit. The field-end parking space detection unit is a parking space geomagnetic detection device. The parking space geomagnetic detection device is composed of a geomagnetic sensor, a power supply, a single-chip microcomputer and a communication unit. The field-end parking space detection unit determines the parking status by detecting the disturbance of the earth's magnetic field near the parking space when the vehicle arrives at the parking space, and can upload the real-time parking space usage information to the parking lot management module. The field-end parking space detection unit is responsible for counting and reporting the information of parked vehicles in the parking lot, including license plate number, parking time, vehicle location and related information of autonomously parked vehicles when driving. After the field-end parking space detection unit collects the vehicle-related information, it transmits it to the Internet of Vehicles platform in real time through the network. The roadside data unit is responsible for monitoring the vehicle driving information on the mobile phone and transmitting it to the Internet of Vehicles platform for use.

[0028] The working process of the autonomous parking control system includes the following steps:

[0029] Step 1: Determine the minimum turning radius: The minimum turning radius of a car refers to the radius of the circle that the center plane of the outer steering wheel of the car travels on the supporting plane when the front wheel steering angle reaches the maximum and the car makes a turning motion at the lowest stable speed. R A Indicates the minimum turning radius of the car, and its calculation formula is as follows:

[0030]

[0031] Where: L represents the vehicle wheelbase, θ max Indicates the maximum steering angle of the front wheels;

[0032] Step 2: Establish a vehicle motion model: Take the center point of the rear axle of the vehicle as the reference point and establish the vehicle model based on the vehicle's own information. The bicycle model is established based on the following assumptions:

[0033] 1) The movement of the vehicle in the vertical direction, i.e. the Z-axis direction, is not considered, that is, the movement of the vehicle is assumed to be a movement on a two-dimensional plane;

[0034] 2) Assume that the left and right tires of the vehicle have the same steering angle and speed at any time; in this way, the movement of the left and right tires of the vehicle can be combined into one tire to describe;

[0035] 3) Assume that the vehicle speed changes slowly and ignore the transfer of front and rear axle loads;

[0036] 4) Assume that the vehicle body and suspension system are both rigid systems;

[0037] 5) Assume that the vehicle's movement and steering are driven by the front wheels;

[0038] In the geodetic coordinate system OXY, since the speed of the center point of the rear axle of the vehicle is slightly different from the speed at the center of mass of the vehicle and is always consistent with the direction of the vehicle body, the motion trajectory equation is easy to express. Therefore, the center point of the rear axle of the vehicle (x, y) is used as the reference point, the speed of the center point of the rear axle is v, and the coordinates of the center point of the front axle of the vehicle are (x f ,y f ), the heading angle of the vehicle is θ, the wheelbase of the vehicle is L, the extension line of the rear axle of the vehicle and the extension line perpendicular to the front wheel turning angle intersect at a point P, which is the instantaneous steering center of the vehicle, and the distance between point Р and the center point of the rear axle is the turning radius R.

[0039] The vehicle kinematic model is expressed as follows:

[0040] δ is the equivalent turning angle of the front wheel. According to the geometric relationship, the angle corresponding to the instantaneous turning center of the vehicle is also δ. Therefore, the relationship between the equivalent turning angle of the front wheel and the wheelbase and turning radius can be expressed as:

[0041]

[0042] The relationship between the coordinates of the center point of the front axle and the center point of the rear axle of the vehicle:

[0043] x f =x+L·cosθ

[0044] y f =y+L·sinθ

[0045] The component speeds of the center point of the vehicle's rear axle on the horizontal and vertical axes in the geodetic coordinate system are:

[0046]

[0047] Since the vehicle does not skid, the vertical speed of the front and rear wheels is zero;

[0048]

[0049]

[0050] From the above calculations, we can conclude that:

[0051]

[0052] In summary, the kinematic equation of the vehicle is:

[0053]

[0054] In this formula, [x, y, θ] T is the state quantity of the vehicle, indicating the position and posture of the vehicle, [v,δ] T To control the amount, the table makes the vehicle's speed and front wheel angle;

[0055] Where (x, y) represents the coordinates of the vehicle's geometric center position; (x f ,y f ) represents the coordinates of the center position of the front wheel of the vehicle, δ is the equivalent turning angle of the front wheel; v represents the vehicle speed; L represents the wheelbase; θ represents the heading angle of the vehicle; R represents the radius of the circle followed by the rear axle at a given turning angle;

[0056] Step 3: Path planning: The Hybrid A* algorithm converts the initial posture point (x 1 ,y 1 ,θ 1 ) is added to the open list, and the node with the highest priority in the open list is added to the closed list to determine whether the current node is less than the threshold. If it is less than the threshold, the Reeds-Shepp curve or Dubins curve is used to compare with the target posture point (x 2 ,y 2 ,θ 2 ) and perform collision detection. If there is no collision with the obstacle, the curve is retained and a path is generated. If there is a collision with the obstacle, the curve is abandoned and nodes are searched again from the open list and expanded again. Finally, a feasible path from the starting state to the target state that meets the vehicle kinematic constraints is generated. The specific process is as follows:

[0057] (1) First, the vehicle path planning system in the on-board computer makes a judgment based on the relevant map provided by the terminal, the current positioning information, and the target storage location. The control end converts the map information provided by the terminal into a two-dimensional grid map that can be used by the vehicle and obtains the initial posture of the vehicle (x 1 ,y 1 ,θ 1 ) and target posture (x 2 ,y 2 ,θ 2 );

[0058] (2) The Hybrid A* algorithm selects the maximum of the two as the final heuristic information, H(n) = max(Reeds-Shepp, A*). The first type of "considering constraints but not obstacles" is to prevent the robot from reaching the target from the wrong direction, and the second type of "considering obstacles but not constraints" is to prevent wasting time in dead ends or U-shaped obstacles.

[0059] (3) The Reeds-Shepp curve used in (2) above can not only explore the heuristic information of the vehicle, but also provide the interval path between two nodes that conforms to the vehicle kinematic model in the subsequent node expansion;

[0060] In the path of the Reeds-Shepp curve, cusps are allowed to exist. In order to indicate the cusps, superscripts need to be added to the letters in the field to indicate the direction of movement, as shown below:

[0061] symbol meaning Around the unit circle <![CDATA[L + ]]> Turn left forward Counterclockwise <![CDATA[L - ]]> Turn left Clockwise <![CDATA[R + ]]> Turn left forward Counterclockwise <![CDATA[R - ]]> Turn left Clockwise <![CDATA[S + ]]> Go straight ahead / <![CDATA[S - ]]> Go straight back /

[0062] The differential equation according to the vehicle kinematic model is as follows:

[0063]

[0064]

[0065]

[0066] Its integral form is as follows:

[0067]

[0068]

[0069]

[0070] Points can be obtained:

[0071]

[0072]

[0073] ψ(t)=t

[0074] The corresponding paths are as follows:

[0075]

[0076]

[0077]

[0078]

[0079]

[0080]

[0081] In order to be more precise, The somewhat redundant convention is used in that the sign of t matches the direction;

[0082] If t>0: the vehicle moves forward, the symbol is +, if t<0: the vehicle moves backward, the symbol is -.

[0083] The Reeds and Shepp curve has a total of 48 fields. The "|" in the phrase indicates that the vehicle's movement direction changes from forward to reverse or from reverse to forward. Some of these 48 fields have two formulas for the actual path of the field type. For any given situation, there are at most 68 formulas.

[0084] The following table lists the 48 fields in the full set:

[0085]

[0086] The unification of position and posture is as follows:

[0087] Before calculation, the vehicle's posture must be normalized. Assume that the vehicle's initial posture is (x 1 ,y 1 ,θ 1 ), the target posture is (x 2 ,y 2 ,θ 2 ), the radius of the vehicle is r = ρ, the normalization process is the process of vector translation and rotation, so that the transformed starting position is (0, 0, 0), the target position is (x, y, φ), and the turning radius of the vehicle is r = 1;

[0088] x=((x 2 -x 1 )*cos(θ 1 )+(y 2 -y 1 )*sin(θ 1 )) / ρ

[0089] y=(-(x 2 -x 1 )*sin(θ 1 )+(y 2 -y 1 )*cos(θ 1 )) / ρ

[0090] φ=θ 2 -θ 1

[0091] The symmetric relationship is used to reduce the complexity of the solution, as follows:

[0092] The principle of solving the Reeds-Shepp curve is that after changing the target point, it is equivalent to the original target point A being symmetrical about the y-axis and the x-axis, obtaining a new target point B, generating a set of new paths t, u, v, and using timefilp symmetry, that is, symbol symmetry and reflect symmetry, that is, RL symmetry, the following table relationship can be obtained:

[0093] (x, y, θ) (-x, y, -θ) (-x, -y, θ) (x, -y, -θ) <![CDATA[L + ]]> <![CDATA[L - ]]> <![CDATA[R - ]]> <![CDATA[R + ]]> <![CDATA[R + ]]> <![CDATA[R - ]]> <![CDATA[L - ]]> <![CDATA[L + ]]> <![CDATA[S + ]]> <![CDATA[S - ]]> <![CDATA[S - ]]> <![CDATA[S + ]]> <![CDATA[L - ]]> <![CDATA[L + ]]> <![CDATA[R + ]]> <![CDATA[R - ]]> <![CDATA[R - ]]> <![CDATA[R + ]]> <![CDATA[L + ]]> <![CDATA[L - ]]> <![CDATA[S - ]]> <![CDATA[S + ]]> <![CDATA[S + ]]> <![CDATA[S - ]]>

[0094] This simplifies the calculation;

[0095] Step 4, collision detection: replace the vehicle body with three overlapping circles. For each state, calculate whether the distance from the obstacle grid point to the center of the circle is greater than the radius to perform collision detection. The union of the three circles contains the vehicle body. The circle approximation is effective because the collision detection of the circle is very fast. It only needs to calculate whether the distance from the obstacle point in the grid to the center of the circle is less than the radius. If it is less than the radius, it means that a collision occurs, otherwise no collision occurs. The calculation amount is small and the accuracy is relatively high.

[0096] The specific working process of the trajectory tracking control system is as follows:

[0097] According to the given path, it is calculated as a reference trajectory with continuous curvature. The front and rear axles of the vehicle are both used as control points. The lateral error is controlled by the angle tracking error with the center of the front axle of the vehicle as the control point, and the longitudinal error is controlled by the speed tracking error with the center of the rear axle of the vehicle as the control point. The controller changes the motor power by controlling the motor voltage to achieve the control of the speed and the front wheel angle, decoupling the control of the vehicle angle and vehicle speed to the dual-axis control with the center of the front and rear axles as the control points, achieving the separation control of the vehicle speed and the angular speed, reducing the mutual influence of the two controls, making full use of the front and rear axle degrees of freedom of the vehicle, and having greater front and rear wheel degrees of freedom and higher tracking accuracy.

[0098] The process includes:

[0099] Step 1: According to the real-time posture of the vehicle and the posture of the parking point, a vehicle kinematic model with speed and angular velocity constraints is combined to obtain a reference posture sequence of the vehicle with the rear axle as the geometric center;

[0100] Step 2: Convert the geometric center point reference pose sequence into the vehicle's rear axle center point reference trajectory through the pose transformation relationship;

[0101] Step 3: The vehicle kinematic model is expressed as follows:

[0102] y(t+Δt)=y(t)+v t sin(δ t )Δt

[0103] x(t+Δt)=x(t)+v t cos(δ t )Δt

[0104]

[0105]

[0106] Where (x, y) represents the coordinates of the vehicle's geometric center position; δ represents the angle between the speed and the horizontal line; β represents the front wheel turning angle; α represents the angle between the vehicle body and the target point; t represents a certain moment; Δt represents the time interval; L represents the wheelbase; L f represents the foresight distance; v t represents the speed at time t; R represents the radius of the circle followed by the rear axle at a given turning angle;

[0107] Step 4: The speed and corner speed constraints are expressed as follows:

[0108] V(t+Δt)=V(t)+aΔt

[0109] V≤V max

[0110] β≤β max

[0111] V represents the speed value; V max Assume the maximum speed; β max The maximum front wheel turning angle is set;

[0112] Step 5: The turning radius and foresight distance constraints are expressed as follows:

[0113] R≤R max

[0114] L f =kv c

[0115] Among them, R max is the maximum turning radius of the vehicle; k represents the proportionality coefficient between the forward sight distance and the speed;

[0116] Step 6: The posture transformation relationship is expressed by the following formula:

[0117]

[0118]

[0119]

[0120]

[0121] Among them, x f With y f Indicates the position coordinates of the center point of the front axle; x r With y r Indicates the position coordinates of the center point of the rear axle;

[0122] According to the reference trajectory of the rear axle center, the preview method of feedback control is adopted to obtain the rear wheel angle control value;

[0123] The process includes:

[0124] Calculate the distance d from the center point of the rear axle to the closest point A on the reference trajectory of the center point of the rear axle;

[0125] Calculate the next point B whose distance from this point satisfies the foresight distance Lf;

[0126] Take point B of the reference trajectory of the vehicle's rear axle center point as the rear wheel preview point P;

[0127] The front wheel turning angle β of the vehicle is calculated based on the preset rear wheel preview coefficient k, the distance d, and the position coordinates of the rear wheel preview point P;

[0128] Finally, the vehicle updates its own status information in real time, updates the current speed and steering angle, obtains the speed control value of the controlled object through the PID controller, and finally drives into the parking space.

[0129] The working process of the vehicle-side collision avoidance system is as follows:

[0130] The vehicle-side collision avoidance system consists of three parts: signal acquisition module, data processing module and execution module. The working process of the specific part includes the following steps:

[0131] Step 1: The signal acquisition module works: The signal acquisition module uses radar and camera technology to automatically measure the distance between the vehicle and the obstacle;

[0132] Step 2: Work of the data processing module: The data processing module processes the instantaneous speed and determines the safe distance. If it is less than the safe distance, it issues a command to alarm or brake. The specific process is as follows:

[0133] The coordinate system is established with the center of mass of the vehicle as the reference point; the vehicle collision time TTC is introduced to reflect the degree of danger of the driving condition. -1 , combined with the driving information and the critical indicator of the vehicle's danger level, the coordinated collision avoidance braking deceleration threshold a is calculated and output by the following formula x-max

[0134]

[0135] In the above formula, v m is the initial speed of the vehicle, v n is the speed of the obstacle. Here, the obstacle is considered as a fixed object with no speed, so v n =0, S 0 is the longitudinal distance between the vehicle and the obstacle, and 0.8 is the TTC of the vehicle in danger level. -1 critical indicators;

[0136] Based on the ideal braking deceleration a′ xand braking deceleration threshold a x-max , after comparison, the expected braking deceleration a for cooperative collision avoidance under the current road conditions is output x-fine : when |a′ x |≤|a x-max |, let a x-fine =a′ x ; when |a′ x |>|a x-max |, let a x-fine =a x-max ;

[0137] Based on the expected braking deceleration ax-fine of cooperative collision avoidance, the cooperative collision avoidance trajectory planning is completed, which is expressed as:

[0138]

[0139] In the above formula, x′(t) and y′(t) are the coordinates of the vehicle at time t, v m is the initial speed of the vehicle, t′ m is the completion time of collaborative collision avoidance, X(t′ m ) is the longitudinal displacement of the center of mass of the ego vehicle when the coordinated collision avoidance is completed;

[0140] According to the planned trajectory, considering that the ego vehicle smoothly and effectively completes the collision avoidance, the critical collision scenario in the collision avoidance process is analyzed, and the critical safety distance D of collision avoidance is calculated;

[0141] Based on the critical safety distance D of cooperative collision avoidance, the longitudinal distance S between the vehicle and the obstacle collected in real time is compared. 0 The relationship between the critical safety distance D of cooperative collision avoidance and the two;

[0142] Step 3: Work of the execution module: The execution module receives the instructions of the data processing module and implements them, issuing instructions for alarm, closing windows and automatic braking;

[0143] If S 0 <D coordination, collision is unavoidable, control the vehicle to brake as hard as possible to reduce collision damage; execute the module and issue corresponding alarm;

[0144] If S 0 >D coordination, the collision can be avoided and the vehicle continues to move.

[0145] The working process of the information processing system is as follows:

[0146] The information processing system is mainly composed of three hierarchical structures, which are the collection layer, transmission layer and application layer from low to high. The working steps of each layer are as follows:

[0147] Step 1, the work of the collection layer: the collection layer is responsible for collecting the vehicle speed, steering angle, positioning information, mileage, engine speed and temperature inside the vehicle. The collection layer is completed by vehicle sensors. The sensors on the transmission include speed sensors, temperature sensors, shaft speed sensors and pressure sensors. The sensors on the steering gear include angle sensors, torque sensors and hydraulic sensors. The sensors on the suspension include speed sensors, acceleration sensors, body height sensors, roll angle sensors and angle sensors.

[0148] The vehicle sensor is the input device of the vehicle computer system. It converts various working condition information of the vehicle, vehicle speed, temperature of various media and engine operating conditions into electrical signals and inputs them into the vehicle computer. The collected data will also be uploaded to the backend server for unified processing and analysis to obtain the business data required by the user, providing reliable data support for the vehicle to complete subsequent work;

[0149] Step 2: Transport layer work: The transport layer is composed of network servers and WEB services. From the perspective of communication and information processing, the transport layer provides communication services to the application layer above. The transport layer is located above the network layer. It provides logical communication between processes running on different hosts, while the network layer provides logical communication between hosts. Even if the network layer protocol is unreliable, the transport layer can still provide reliable services for applications. The transport layer provides logical communication between application processes, that is, end-to-end communication. Logical communication means that the communication between the transport layers is to transmit data in the horizontal direction, but in fact there is no horizontal physical connection between the two transport layers.

[0150] When the transport layer uses connection-oriented TCP, this logical communication channel is equivalent to a full-duplex reliable channel. The server manages the data statistics and provides corresponding services for each vehicle. At the same time, the data can be jointly analyzed to become a local Internet of Vehicles service business, providing efficient, accurate and timely data services to the user group.

[0151] Step 3: Work of the application layer: The application layer removes the inherent format of all the transmission data of the network transmission layer, that is, the application of the remaining Data part after the SF, FF, CF flag bits, Len and padding bits, and gives the Data part practical meaning;

[0152] The network layer is responsible for data transmission, while the application layer is concerned with obtaining information from the data. The request in the application layer refers to the information sent from the client to the server; the response in the application layer refers to the information sent from the server to the client. The client usually refers to the diagnostic instrument, host computer and test tool, and the server usually refers to the BCM, GW and PEPS body electronic components. The standard application layer service refers to ISO 15765-3 and ISO 14229-1 documents and is used in the scenarios of testing, monitoring, diagnosis and vehicle online refresh.

[0153] Beneficial effects of the present invention:

[0154] The setting method of the autonomous parking system of a large electric truck based on field-side collaboration provided by the present invention utilizes the on-board equipment on the vehicle through wireless communication technology to effectively utilize all static and dynamic information in the Internet of Vehicles platform, and provide different functional services during vehicle operation. The Internet of Vehicles exhibits the following characteristics: the Internet of Vehicles can provide a guarantee for the distance between the vehicle and obstacles, reducing the probability of vehicle collision accidents; the Internet of Vehicles can help car owners navigate in real time or realize autonomous driving, and improve the efficiency of traffic operation through communication with other vehicles or network systems. The present invention proposes a practical autonomous parking system to facilitate people's lives. The system is suitable for the transition stage between semi-intelligent driving and intelligent driving, and is a great leap in technology. The application of this system can not only save the driver's parking time, but also improve people's driving experience.

[0155] This system uses big data analysis to rationally allocate parking resources, thereby effectively alleviating the contradiction between parking supply and demand, achieving the maximum utilization of resources, and helping people save time. In addition, it can reduce the difficulty of driving. For most people, parking is a difficult process. The implementation of this system can help us reduce driving pressure and have a better mentality to drive the vehicle. Secondly, vehicle intelligence has become one of the important symbols to measure a country's social civilization and scientific and technological progress. As a new interdisciplinary field, the research on intelligent vehicles integrates the latest achievements of information science and artificial intelligence technology, has a wide range of disciplinary development and theoretical value, and is a concentrated reflection of the comprehensive strength of the country's scientific research level. The research results will also promote the rapid development of related disciplines, drive the development of the entire national economy, improve the industrial structure, and promote the information level of the entire country. BRIEF DESCRIPTION OF THE DRAWINGS

[0156] Figure 1 Schematic diagram of the minimum turning radius of the automatic parking vehicle according to the present invention.

[0157] Figure 2 This is a schematic diagram of the Ackerman steering geometry described in the present invention.

[0158] Figure 3The figure is a schematic diagram of a bicycle model according to the present invention.

[0159] Figure 4 This is a schematic diagram of the collision detection described in the present invention.

[0160] Figure 5 Schematic diagram of the overall process of the autonomous parking system of the present invention.

[0161] Figure 6 The figure is a schematic diagram of the information exchange process between the parking lot management module and the Internet of Vehicles platform described in the present invention.

[0162] Figure 7 Schematic diagram of the information exchange process between the autonomous parking vehicle and the Internet of Vehicles platform according to the present invention.

[0163] Figure 8 The figure is a schematic diagram of the path planning process described in the present invention.

[0164] Fig. 9 It is a schematic diagram of the flow of the trajectory tracking control system described in the present invention.

[0165] Fig.10 The figure is a schematic diagram of the process flow of the vehicle-end collision avoidance system of the present invention. DETAILED DESCRIPTION

[0166] See also Figures 1 to 10 As shown:

[0167] The method for setting up an autonomous parking system for a large electric truck based on field-side collaboration provided by the present invention comprises the following steps:

[0168] The first step is to install the Internet of Vehicles platform, which includes an information receiving module, an information sorting module, a high-precision map platform and an information sending module. The information receiving module and the information sending module are connected to the field system and the on-board computer through the network to realize real-time transmission of information and application and acceptance of commands. The information sorting module includes a parking space storage unit and a parking management unit. The information sorting module will process and summarize the parking information of vehicles in the parking lot according to the received information. The parking space storage unit uses the detection of the field support system to obtain the number and location of the vacant parking space. When the vacant parking space is identified and located, the information processing The module stores the parking space information in the parking space database and waits for the vehicle to apply. When the autonomous vehicle issues a parking application command, the parking space recommendation unit uses a sorting algorithm that approximates the ideal solution to obtain the optimal sorting of free parking spaces and recommends the optimal parking space to the autonomous vehicle. The Internet of Vehicles platform will update the parking space database and no longer release the recommended parking spaces to other vehicles. The high-precision map platform is responsible for monitoring the map data of the parking lot. The high-precision map platform locates the parking space uploaded to the Internet of Vehicles platform and the positioning coordinates of the applying vehicle. The Internet of Vehicles platform will send the relevant information to the corresponding vehicle by accessing the pre-stored map.

[0169] The second step is to establish a parking lot management module, which includes a parking space detection unit at the parking lot and a roadside data unit. The parking space detection unit at the parking lot detects the vacancy status and parking space number of the parking space through the parking space detection sensor and pre-processes the parking space signal. The roadside data unit is responsible for receiving the vehicle parking space application command issued by the Internet of Vehicles platform and the parking space information issued by the platform.

[0170] The third step is to build an autonomous parking control system. The autonomous parking control system includes building a vehicle model, performing path planning, and transmitting driving information to the motion execution part. It is the control center of the entire autonomous parking system. The specific process is as follows:

[0171] The vehicle's autonomous parking control system first determines the minimum turning radius of the car based on the vehicle's own properties, such as the vehicle's wheelbase L and the maximum steering angle of the front wheels θmax. Then, based on the radius and the speed and heading angle of the vehicle during movement and the relevant position coordinates, a real-time model of the vehicle is established to control the vehicle. Then, the vehicle plans the path based on its own model and the map information and target parking space information released by the Internet of Vehicles platform. At the same time, collision avoidance needs to be considered. The Hybrid A* algorithm is used. It will gradually perform heuristic search in a continuous coordinate system, expand subnodes, and perform collision detection on the path based on the existing obstacle map. The nodes corresponding to the collision-free path will be added to the expansion tree to generate a large number of possible paths. Finally, the loss function is designed to construct the Voronoi diagram and the conjugate gradient method is used for non-parametric interpolation to improve the quality of the generated path in order to obtain a target path that can effectively avoid obstacles, is smooth and efficient.

[0172] The fourth step is to carry out trajectory tracking control. The vehicle's trajectory tracking control system controls the vehicle to have a suitable speed and steering angle according to relevant calculations, and controls the vehicle to travel along the corresponding path;

[0173] The vehicle's trajectory tracking control system regards the vehicle as a bicycle model with the rear axle as the center under ideal conditions. Then, based on the Pure Pursuit trajectory tracking control method, the sensors collect the vehicle's geometric center position coordinates, the angle between the speed and the horizontal line, the front wheel angle, the angle between the vehicle body and the target point, the forward sight distance and the movement speed at a specific moment in real time, so that the vehicle has an accurate driving route when moving. Finally, the trajectory tracking control system calculates the target values ​​of the current vehicle actuators of throttle, brake and steering according to the input target trajectory, and transmits the target data to the vehicle actuator ECU through CAN. The actuator controls the specific actions of the actuator according to the target value.

[0174] Step 5: Build a real-time vehicle-side collision avoidance system. When the vehicle-side collision avoidance system finds an unavoidable obstacle, it will automatically sound an alarm and take braking or evasive measures to avoid a collision.

[0175] The real-time vehicle-side collision avoidance system is a method for vehicles to avoid collisions or a solution when encountering unavoidable situations. The real-time vehicle-side collision avoidance system includes a signal acquisition module, a data processing module and an execution module. The execution module includes several decision-making execution methods such as starting the alarm, closing the windows and emergency braking. The signal acquisition module collects the real-time status information of the vehicle itself, surrounding vehicles and obstacles and the current road information in real time based on its own vehicle sensors, and continuously transmits the real-time data to the data processing module of the real-time vehicle-side collision avoidance system. The data processing module calculates the critical safety distance for collision avoidance based on its own vehicle speed and the real-time distance between the vehicle and the obstacle, completes the collision avoidance decision based on the current longitudinal distance between the vehicle and the obstacle, and plans the vehicle's decision-making method. If the distance allows braking to avoid collision, the vehicle implements appropriate deceleration to slow down. If the distance is too short to avoid collision, the vehicle will be forced to trigger several decisions such as alarms and emergency braking.

[0176] Step 6: Establishment of the information processing system, which is responsible for collecting relevant information of the vehicle during driving and sending it to various systems for sharing;

[0177] The information processing system consists of three parts: information collection layer, transmission layer and application layer. The collection layer is responsible for collecting the real-time operating parameters of the vehicle, road environment parameters and prediction parameters and converting them into electrical signals. The transmission layer is responsible for connecting various systems and providing information transmission services, that is, realizing the aggregation, analysis, processing and transmission of input and output data. The application layer refers to the integration of various information, the close integration of business logic and resources, and the presentation of information processing results to users in the form of rich multimedia.

[0178] The autonomous driving vehicle in the above steps is equipped with an on-board computer, which is also the control center of the autonomous driving vehicle. The on-board computer has built-in codes for the autonomous parking control system and trajectory tracking control system and the vehicle-side collision avoidance system that are input in advance. The on-board computer is connected to the Internet of Vehicles platform through the network, and the parking lot management module and the Internet of Vehicles platform are also connected through the network. The field sensors inside the parking lot collect vehicle driving information and parking space information in real time and send them to the Internet of Vehicles platform for everyone to see and use. At the same time, the high-precision map platform in the Internet of Vehicles platform will also share the detailed map of the parking lot for vehicles to use. The specific process is as follows:

[0179] After the on-board computer receives the map information and parking space information published by the Internet of Vehicles platform, it shares it with the downlink system. First, the autonomous parking control system plans a reasonable driving path through the pre-established vehicle motion model and the received map information and parking space information. The path is transmitted to the trajectory tracking control system in the computer. The trajectory tracking control system starts the vehicle, and the vehicle-side sensors start running. The data from the combined effect of the on-board camera and the field-side sensors is published on the Internet of Vehicles platform and sent to the on-board computer at the same time to calculate the real-time distance between the vehicle and the obstacle. After the vehicle moves, the vehicle-side sensors monitor the vehicle and collect the The vehicle speed information and steering angle information are converted into a trajectory tracking control system that inputs electrical signals. The trajectory tracking control system is adjusted according to the driving path and position information, and the speed and steering angle at the next moment are calculated, and input to the vehicle through electrical signals. At the same time, the vehicle speed information, steering angle information, the distance between the vehicle and the obstacle, and the location map information are shared with the vehicle-side collision avoidance system. The vehicle-side collision avoidance system runs in real time to calculate whether it is a safe distance. If everything is normal, it continues to drive. If it is less than the safe distance, the vehicle-side collision avoidance system will directly transmit electrical signals to the braking system, motor and alarm switch to force braking, emergency stop of the motor and alarm.

[0180] The work content of the Internet of Vehicles platform is as follows:

[0181] The Internet of Vehicles platform is connected via the network and is used to accept parking instructions from car owners. The information receiving module and information sending module of the Internet of Vehicles platform are simultaneously connected to the unmanned vehicle and the parking lot management module via the network, realizing the connection between the field-side sensors and the vehicle-side sensors, and is used to monitor the vehicle's driving path in all directions and count available parking spaces, obtain information about the vehicle's surrounding environment, and help guide the vehicle to avoid other vehicles and obstacles and drive into the parking space. The Internet of Vehicles platform is used to obtain available parking spaces from the parking lot management module and send them to the vehicle, and is also used to control the vehicles parked in the parking lot in real time;

[0182] The information sorting module in the Internet of Vehicles platform is mainly responsible for the management of vehicles and parking spaces and the optimal recommendation of parking spaces. It includes three parts: parking space identification / positioning unit, parking space recommendation unit and parking management unit. The parking space identification and positioning unit uses artificial intelligence technology to identify parking spaces and obtain the number and location of vacant parking spaces. After the vacant parking spaces are identified and located, the Internet of Vehicles platform stores the parking space information in the server and waits for parking space application commands. When the autonomous driving vehicle issues a parking application command, it is transmitted to the Internet of Vehicles platform through the network layer. The parking space recommendation unit in the Internet of Vehicles platform uses a sorting algorithm that approximates the ideal solution to obtain the optimal sorting of vacant parking spaces and sends the optimal parking space to the autonomous driving vehicle through the network. At the same time, the Internet of Vehicles platform will update the parking space database and will no longer release the recommended parking spaces to other vehicles.

[0183] The high-precision map platform belongs to the downstream part of the Internet of Vehicles platform. It can output lane-level parking lot map data to support low-latency, high-precision path planning. The Internet of Vehicles platform transmits the vehicle's parking application command to the information receiving module, and then the high-precision map platform sends the parking lot's traffic information to the on-board navigation system in real time through the network. The on-board navigation system determines the feasible parking path based on the parking position and the real-time traffic information of the parking lot. Based on the parking space uploaded to the Internet of Vehicles platform and the positioning coordinates of the applying vehicle, the Internet of Vehicles platform will guide the vehicle to automatically drive to the parking space and complete parking by accessing the pre-stored map.

[0184] The parking lot management module built works as follows:

[0185] The parking lot that meets the above steps is a parking lot with the function of guiding and managing unmanned vehicles. In order to rationally utilize parking spaces in the parking lot and reduce waste of resources, the parking lot management system is provided with a field-end parking space detection unit. The field-end parking space detection unit is a parking space geomagnetic detection device. The parking space geomagnetic detection device is composed of a geomagnetic sensor, a power supply, a single-chip microcomputer and a communication unit. The field-end parking space detection unit determines the parking status by detecting the disturbance of the earth's magnetic field near the parking space when the vehicle arrives at the parking space, and can upload the real-time parking space usage information to the parking lot management module. The field-end parking space detection unit is responsible for counting and reporting the information of parked vehicles in the parking lot, including license plate number, parking time, vehicle location and related information of autonomously parked vehicles when driving. After the field-end parking space detection unit collects the vehicle-related information, it transmits it to the Internet of Vehicles platform in real time through the network. The roadside data unit is responsible for monitoring the vehicle driving information on the mobile phone and transmitting it to the Internet of Vehicles platform for use.

[0186] The working process of the autonomous parking control system includes the following steps:

[0187] Step 1: Determine the minimum turning radius: The minimum turning radius of a car refers to the radius of the circle that the center plane of the outer steering wheel of the car travels on the supporting plane when the front wheel steering angle reaches the maximum and the car makes a turning motion at the lowest stable speed. R A Indicates the minimum turning radius of the car, and its calculation formula is as follows:

[0188]

[0189] Where: L represents the vehicle wheelbase, θ max Indicates the maximum steering angle of the front wheels;

[0190] Step 2: Establish a vehicle motion model: Take the center point of the rear axle of the vehicle as the reference point and establish the vehicle model based on the vehicle's own information. The bicycle model is established based on the following assumptions:

[0191] 1) The movement of the vehicle in the vertical direction, i.e. the Z-axis direction, is not considered, that is, the movement of the vehicle is assumed to be a movement on a two-dimensional plane;

[0192] 2) Assume that the left and right tires of the vehicle have the same steering angle and speed at any time; in this way, the movement of the left and right tires of the vehicle can be combined into one tire to describe;

[0193] 3) Assume that the vehicle speed changes slowly and ignore the transfer of front and rear axle loads;

[0194] 4) Assume that the vehicle body and suspension system are both rigid systems;

[0195] 5) Assume that the vehicle's movement and steering are driven by the front wheels;

[0196] In the geodetic coordinate system OXY, since the speed of the center point of the rear axle of the vehicle is slightly different from the speed at the center of mass of the vehicle and is always consistent with the direction of the vehicle body, the motion trajectory equation is easy to express. Therefore, the center point of the rear axle of the vehicle (x, y) is used as the reference point, the speed of the center point of the rear axle is v, and the coordinates of the center point of the front axle of the vehicle are (x f ,y f ), the heading angle of the vehicle is θ, the wheelbase of the vehicle is L, the extension line of the rear axle of the vehicle and the extension line perpendicular to the front wheel turning angle intersect at a point P, which is the instantaneous steering center of the vehicle, and the distance between point Р and the center point of the rear axle is the turning radius R.

[0197] The vehicle kinematic model is expressed as follows:

[0198] δ is the equivalent turning angle of the front wheel. According to the geometric relationship, the angle corresponding to the instantaneous turning center of the vehicle is also δ. Therefore, the relationship between the equivalent turning angle of the front wheel and the wheelbase and turning radius can be expressed as:

[0199]

[0200] The relationship between the coordinates of the center point of the front axle and the center point of the rear axle of the vehicle:

[0201] x f =x+L·cosθ

[0202] y f =y+L·sinθ

[0203] The component speeds of the center point of the vehicle's rear axle on the horizontal and vertical axes in the geodetic coordinate system are:

[0204]

[0205] Since the vehicle does not skid, the vertical speed of the front and rear wheels is zero;

[0206]

[0207]

[0208] From the above calculations, we can conclude that:

[0209]

[0210] In summary, the kinematic equation of the vehicle is:

[0211]

[0212] In this formula, [x, y, θ] T is the state quantity of the vehicle, indicating the position and posture of the vehicle, [v,δ] T To control the amount, the table makes the vehicle's speed and front wheel angle;

[0213] Where (x, y) represents the coordinates of the vehicle's geometric center position; (x f ,y f ) represents the coordinates of the center position of the front wheel of the vehicle, δ is the equivalent turning angle of the front wheel; v represents the vehicle speed; L represents the wheelbase; θ represents the heading angle of the vehicle; R represents the radius of the circle followed by the rear axle at a given turning angle;

[0214] Step 3: Path planning: The Hybrid A* algorithm converts the initial posture point (x 1 ,y 1 ,θ 1 ) is added to the open list, and the node with the highest priority in the open list is added to the closed list to determine whether the current node is less than the threshold. If it is less than the threshold, the Reeds-Shepp curve or Dubins curve is used to compare with the target posture point (x 2 ,y 2 ,θ 2 ) and perform collision detection. If there is no collision with the obstacle, the curve is retained and a path is generated. If there is a collision with the obstacle, the curve is abandoned and nodes are searched again from the open list and expanded again. Finally, a feasible path from the starting state to the target state that meets the vehicle kinematic constraints is generated. The specific process is as follows:

[0215] (1) First, the vehicle path planning system in the on-board computer makes a judgment based on the relevant map provided by the terminal, the current positioning information, and the target storage location. The control end converts the map information provided by the terminal into a two-dimensional grid map that can be used by the vehicle and obtains the initial posture of the vehicle (x 1 ,y 1 ,θ 1 ) and target posture (x 2 ,y 2 ,θ 2 );

[0216] (2) The Hybrid A* algorithm selects the maximum of the two as the final heuristic information, H(n) = max(Reeds-Shepp, A*). The first type of "considering constraints but not obstacles" is to prevent the robot from reaching the target from the wrong direction, and the second type of "considering obstacles but not constraints" is to prevent wasting time in dead ends or U-shaped obstacles.

[0217] (3) The Reeds-Shepp curve used in (2) above can not only explore the heuristic information of the vehicle, but also provide the interval path between two nodes that conforms to the vehicle kinematic model in the subsequent node expansion;

[0218] In the path of the Reeds-Shepp curve, cusps are allowed to exist. In order to indicate the cusps, superscripts need to be added to the letters in the field to indicate the direction of movement, as shown below:

[0219] symbol meaning Around the unit circle <![CDATA[L + ]]> Turn left forward Counterclockwise <![CDATA[L - ]]> Turn left Clockwise <![CDATA[R + ]]> Turn left forward Counterclockwise <![CDATA[R - ]]> Turn left Clockwise <![CDATA[S + ]]> Go straight ahead / <![CDATA[S - ]]> Go straight back /

[0220] The differential equation according to the vehicle kinematic model is as follows:

[0221]

[0222]

[0223]

[0224] Its integral form is as follows:

[0225]

[0226]

[0227]

[0228] Points can be obtained:

[0229]

[0230]

[0231] ψ(t)=t

[0232] The corresponding paths are as follows:

[0233]

[0234]

[0235]

[0236]

[0237]

[0238]

[0239] In order to be more precise, The somewhat redundant convention is used in that the sign of t matches the direction;

[0240] If t>0: the vehicle moves forward, the symbol is +, if t<0: the vehicle moves backward, the symbol is -.

[0241] The Reeds and Shepp curve has 48 fields in total. The “|” in the phrase indicates that the vehicle’s direction of motion changes from forward to reverse or from reverse to forward. Some of these 48 fields have two formulas for the actual path of the field type. For any given situation, there are at most 68 formulas.

[0242] The following table lists the 48 fields in the full set:

[0243]

[0244] The unification of position and posture is as follows:

[0245] Before calculation, the vehicle's posture must be normalized. Assume that the vehicle's initial posture is (x 1 ,y 1 ,θ 1 ), the target posture is (x 2 ,y 2 ,θ 2 ), the radius of the vehicle is r = ρ, and the normalization process is the process of vector translation and rotation, so that the transformed starting position is (0, 0, 0) and the target position is The turning radius of the vehicle is r = 1;

[0246] x=((x 2 -x 1 )*cos(θ 1 )+(y 2 -y 1 )*sin(θ 1 )) / ρ

[0247] y=(-(x 2 -x 1 )*sin(θ 1 )+(y 2 -y 1 )*cos(θ 1 )) / ρ

[0248] φ=θ 2 -θ 1

[0249] The symmetric relationship is used to reduce the complexity of the solution, as follows:

[0250] The principle of solving the Reeds-Shepp curve is that after changing the target point, it is equivalent to the original target point A being symmetrical about the y-axis and the x-axis, obtaining a new target point B, generating a set of new paths t, u, v, and using timefilp symmetry, that is, symbol symmetry and reflect symmetry, that is, RL symmetry, the following table relationship can be obtained:

[0251] (x, y, θ) (-x, y, -θ) (-x, -y, θ) (x, -y, -θ) <![CDATA[L + ]]> <![CDATA[L - ]]> <![CDATA[R - ]]> <![CDATA[R + ]]> <![CDATA[R + ]]> <![CDATA[R - ]]> <![CDATA[L - ]]> <![CDATA[L + ]]> <![CDATA[S + ]]> <![CDATA[S - ]]> <![CDATA[S - ]]> <![CDATA[S + ]]> <![CDATA[L - ]]> <![CDATA[L + ]]> <![CDATA[R + ]]> <![CDATA[R - ]]> <![CDATA[R - ]]> <![CDATA[R + ]]> <![CDATA[L + ]]> <![CDATA[L - ]]> <![CDATA[S - ]]> <![CDATA[S + ]]> <![CDATA[S + ]]> <![CDATA[S - ]]>

[0252] This simplifies the calculation;

[0253] Step 4, collision detection: replace the vehicle body with three overlapping circles. For each state, calculate whether the distance from the obstacle grid point to the center of the circle is greater than the radius to perform collision detection. The union of the three circles contains the vehicle body. The circle approximation is effective because the collision detection of the circle is very fast. It only needs to calculate whether the distance from the obstacle point in the grid to the center of the circle is less than the radius. If it is less than the radius, it means that a collision occurs, otherwise no collision occurs. The calculation amount is small and the accuracy is relatively high.

[0254] The specific working process of the trajectory tracking control system is as follows:

[0255] According to the given path, it is calculated as a reference trajectory with continuous curvature. The front and rear axles of the vehicle are both used as control points. The lateral error is controlled by the angle tracking error with the center of the front axle of the vehicle as the control point, and the longitudinal error is controlled by the speed tracking error with the center of the rear axle of the vehicle as the control point. The controller changes the motor power by controlling the motor voltage to achieve the control of the speed and the front wheel angle, decoupling the control of the vehicle angle and vehicle speed to the dual-axis control with the center of the front and rear axles as the control points, achieving the separation control of the vehicle speed and the angular speed, reducing the mutual influence of the two controls, making full use of the front and rear axle degrees of freedom of the vehicle, and having greater front and rear wheel degrees of freedom and higher tracking accuracy.

[0256] The process includes:

[0257] Step 1: According to the real-time posture of the vehicle and the posture of the parking point, a vehicle kinematic model with speed and angular velocity constraints is combined to obtain a reference posture sequence of the vehicle with the rear axle as the geometric center;

[0258] Step 2: Convert the geometric center point reference pose sequence into the vehicle's rear axle center point reference trajectory through the pose transformation relationship;

[0259] Step 3: The vehicle kinematic model is expressed as follows:

[0260] y(t+Δt)=y(t)+v t sin(δ t )Δt

[0261] x(t+Δt)=x(t)+v t cos(δ t )Δt

[0262]

[0263]

[0264] Where (x, y) represents the coordinates of the vehicle's geometric center position; δ represents the angle between the speed and the horizontal line; β represents the front wheel turning angle; α represents the angle between the vehicle body and the target point; t represents a certain moment; Δt represents the time interval; L represents the wheelbase; L f represents the foresight distance; v t represents the speed at time t; R represents the radius of the circle followed by the rear axle at a given turning angle;

[0265] Step 4: The speed and corner speed constraints are expressed as follows:

[0266] V(t+Δt)=V(t)+aΔt

[0267] V≤V max

[0268] β≤β max

[0269] V represents the speed value; V max Assume the maximum speed; β max The maximum front wheel turning angle is set;

[0270] Step 5: The turning radius and foresight distance constraints are expressed as follows:

[0271] R≤R max

[0272] L f =kv c

[0273] Among them, R max is the maximum turning radius of the vehicle; k represents the proportionality coefficient between the forward sight distance and the speed;

[0274] Step 6: The posture transformation relationship is expressed by the following formula:

[0275]

[0276]

[0277]

[0278]

[0279] Among them, x f With y f Indicates the position coordinates of the center point of the front axle; x r With y r Indicates the position coordinates of the center point of the rear axle;

[0280] According to the reference trajectory of the rear axle center, the preview method of feedback control is adopted to obtain the rear wheel angle control value;

[0281] The process includes:

[0282] Calculate the distance d from the center point of the rear axle to the closest point A on the reference trajectory of the center point of the rear axle;

[0283] Calculate the next point B whose distance from this point satisfies the foresight distance Lf;

[0284] Take point B of the reference trajectory of the vehicle's rear axle center point as the rear wheel preview point P;

[0285] The front wheel turning angle β of the vehicle is calculated based on the preset rear wheel preview coefficient k, the distance d, and the position coordinates of the rear wheel preview point P;

[0286] Finally, the vehicle updates its own status information in real time, updates the current speed and steering angle, obtains the speed control value of the controlled object through the PID controller, and finally drives into the parking space.

[0287] The working process of the vehicle-side collision avoidance system is as follows:

[0288] The vehicle-side collision avoidance system consists of three parts: signal acquisition module, data processing module and execution module. The working process of the specific part includes the following steps:

[0289] Step 1: The signal acquisition module works: The signal acquisition module uses radar and camera technology to automatically measure the vehicle speed and the distance to the obstacle;

[0290] Step 2: Work of the data processing module: The data processing module processes the instantaneous speed and determines the safe distance. If it is less than the safe distance, it issues a command to alarm or brake. The specific process is as follows:

[0291] The coordinate system is established with the center of mass of the vehicle as the reference point; the vehicle collision time TTC is introduced to reflect the degree of danger of the driving condition. -1 , combined with the driving information and the critical indicator of the vehicle's danger level, the coordinated collision avoidance braking deceleration threshold a is calculated and output by the following formula x-max

[0292]

[0293] In the above formula, v m is the initial speed of the vehicle, v n is the speed of the obstacle. Here, the obstacle is considered as a fixed object with no speed, so v n =0, S 0 is the longitudinal distance between the vehicle and the obstacle, and 0.8 is the TTC of the vehicle in danger level. -1 critical indicators;

[0294] Based on the ideal braking deceleration a′ xand braking deceleration threshold a x-max , after comparison, the expected braking deceleration a for cooperative collision avoidance under the current road conditions is output x-fine : when |a′ x |≤|a x-max |, let a x-fine =a′ x ; when |a′ x |>|a x-max |, let a x-fine =a x-max ;

[0295] Based on the expected braking deceleration ax-fine of cooperative collision avoidance, the cooperative collision avoidance trajectory planning is completed, which is expressed as:

[0296]

[0297] In the above formula, x′(t) and y′(t) are the coordinates of the vehicle at time t, v m is the initial speed of the vehicle, t′ m is the completion time of collaborative collision avoidance, X(t′ m ) is the longitudinal displacement of the center of mass of the ego vehicle when the coordinated collision avoidance is completed;

[0298] According to the planned trajectory, considering that the ego vehicle smoothly and effectively completes the collision avoidance, the critical collision scenario in the collision avoidance process is analyzed, and the critical safety distance D of collision avoidance is calculated;

[0299] Based on the critical safety distance D of cooperative collision avoidance, the longitudinal distance S between the vehicle and the obstacle collected in real time is compared. 0 The relationship between the critical safety distance D of cooperative collision avoidance and the two;

[0300] Step 3: Work of the execution module: The execution module receives the instructions of the data processing module and implements them, issuing instructions for alarm, closing windows and automatic braking;

[0301] If S 0 <D coordination, collision is unavoidable, control the vehicle to brake as hard as possible to reduce collision damage; execute the module and issue corresponding alarm;

[0302] If S 0 >D coordination, the collision can be avoided and the vehicle continues to move.

[0303] The working process of the information processing system is as follows:

[0304] The information processing system is mainly composed of three hierarchical structures, which are the collection layer, transmission layer and application layer from low to high. The working steps of each layer are as follows:

[0305] Step 1, the work of the collection layer: the collection layer is responsible for collecting the vehicle speed, steering angle, positioning information, mileage, engine speed and temperature inside the vehicle. The collection layer is completed by vehicle sensors. The sensors on the transmission include speed sensors, temperature sensors, shaft speed sensors and pressure sensors. The sensors on the steering gear include angle sensors, torque sensors and hydraulic sensors. The sensors on the suspension include speed sensors, acceleration sensors, body height sensors, roll angle sensors and angle sensors.

[0306] The vehicle sensor is the input device of the vehicle computer system. It converts various working condition information of the vehicle, vehicle speed, temperature of various media and engine operating conditions into electrical signals and inputs them into the vehicle computer. The collected data will also be uploaded to the backend server for unified processing and analysis to obtain the business data required by the user, providing reliable data support for the vehicle to complete subsequent work;

[0307] Step 2: Transport layer work: The transport layer is composed of network servers and WEB services. From the perspective of communication and information processing, the transport layer provides communication services to the application layer above. The transport layer is located above the network layer. It provides logical communication between processes running on different hosts, while the network layer provides logical communication between hosts. Even if the network layer protocol is unreliable, the transport layer can still provide reliable services for applications. The transport layer provides logical communication between application processes, that is, end-to-end communication. Logical communication means that the communication between the transport layers is to transmit data in the horizontal direction, but in fact there is no horizontal physical connection between the two transport layers.

[0308] When the transport layer uses connection-oriented TCP, this logical communication channel is equivalent to a full-duplex reliable channel. The server manages the data statistics and provides corresponding services for each vehicle. At the same time, the data can be jointly analyzed to become a local Internet of Vehicles service business, providing efficient, accurate and timely data services to the user group.

[0309] Step 3: Work of the application layer: The application layer removes the inherent format of all the transmission data of the network transmission layer, that is, the application of the remaining Data part after the SF, FF, CF flag bits, Len and padding bits, and gives the Data part practical meaning;

[0310] The network layer is responsible for data transmission, while the application layer is concerned with obtaining information from the data. The request in the application layer refers to the information sent from the client to the server; the response in the application layer refers to the information sent from the server to the client. The client usually refers to the diagnostic instrument, host computer and test tool, and the server usually refers to the BCM, GW and PEPS body electronic components. The standard application layer service refers to ISO 15765-3 and ISO 14229-1 documents and is used in the scenarios of testing, monitoring, diagnosis and vehicle online refresh.

Claims

1. A method for setting up an autonomous parking system for a large electric truck based on field-side collaboration, characterized in that: The method comprises the following steps: The first step is to install the Internet of Vehicles platform, which includes an information receiving module, an information sorting module, a high-precision map platform and an information sending module. The information receiving module and the information sending module are connected to the field system and the on-board computer through the network to realize real-time information transmission and the application and acceptance of commands. The information sorting module includes a parking space storage unit and a parking management unit. The information sorting module will process and summarize the parking information of vehicles in the parking lot according to the received information. The parking space storage unit uses the detection of the field support system to obtain the number and location of the vacant parking space. After the vacant parking space is identified and located, the information processing module stores the parking space information in the parking space database and waits for the vehicle to apply; When the autonomous vehicle issues a parking application command, the parking space recommendation unit uses a sorting algorithm that approximates the ideal solution to obtain the optimal sorting of free parking spaces and recommends the optimal parking space to the autonomous vehicle. The Internet of Vehicles platform will update the parking space database and no longer release the recommended parking spaces to other vehicles. The high-precision map platform is responsible for monitoring the map data of the parking lot. The high-precision map platform locates the parking spaces and the positioning coordinates of the applying vehicles uploaded to the Internet of Vehicles platform. The Internet of Vehicles platform will send the relevant information to the corresponding vehicle by accessing the pre-stored map. The second step is to establish a parking lot management module, which includes a parking space detection unit at the parking lot and a roadside data unit; The parking space detection unit at the yard detects the vacancy status and parking space number of the parking space through the parking space detection sensor, and pre-processes the parking space signal; The roadside data unit is responsible for receiving the parking space application command issued by the vehicle networking platform and the parking space information released by the platform; The third step is to build an autonomous parking control system. The autonomous parking control system includes building a vehicle model, performing path planning, and transmitting driving information to the motion execution part. It is the control center of the entire autonomous parking system. The specific process is as follows: The vehicle's autonomous parking control system first determines the vehicle's own properties, such as the vehicle's wheelbase L and the maximum steering angle θ of the front wheels. max To determine the minimum turning radius of the car, and then build a real-time model of the vehicle based on the radius, the speed and heading angle of the vehicle and the relevant position coordinates during movement, so as to control the vehicle. Then the vehicle plans the path according to its own model and the map information and target parking space information released by the Internet of Vehicles platform. At the same time, it needs to consider the collision avoidance situation. The Hybrid A* algorithm is used, which will gradually perform heuristic search in a continuous coordinate system, expand subnodes and perform collision detection on the path based on the existing obstacle map. The nodes corresponding to the collision-free path will be added to the expansion tree to generate a large number of potential paths. Finally, the loss function is designed to construct the Voronoi diagram and the conjugate gradient method is used for non-parametric interpolation to improve the quality of the generated path in order to obtain the target path. The fourth step is to carry out trajectory tracking control. The vehicle's trajectory tracking control system controls the vehicle's speed and steering angle according to calculations, so that the vehicle can travel along the corresponding path; The vehicle's trajectory tracking control system regards the vehicle as a bicycle model with the rear axle as the center under ideal conditions. Then, based on the Pure Pursuit trajectory tracking control method, the sensors collect the vehicle's geometric center position coordinates, the angle between the speed and the horizontal line, the front wheel angle, the angle between the vehicle body and the target point, the forward sight distance and the movement speed at the set time in real time, so that the vehicle has an accurate driving route when moving. Finally, the trajectory tracking control system calculates the target values ​​of the current vehicle actuators of throttle, brake and steering according to the input target trajectory, and transmits the target data to the vehicle actuator ECU through CAN. The actuator controls the action of the actuator according to the target value. Step 5: Build a real-time vehicle-side collision avoidance system. When the vehicle-side collision avoidance system finds an unavoidable obstacle, it will automatically sound an alarm and take braking or evasive measures to avoid a collision. The real-time vehicle-side collision avoidance system is a method for vehicles to avoid collisions or a solution when encountering unavoidable situations. The real-time vehicle-side collision avoidance system includes a signal acquisition module, a data processing module and an execution module. The execution module includes several decision-making execution methods such as starting the alarm, closing the window and emergency braking. The signal acquisition module collects the real-time status information of the vehicle itself, surrounding vehicles and obstacles and the current road information in real time based on its own vehicle sensors, and continuously transmits the real-time data to the data processing module of the real-time vehicle-side collision avoidance system. The data processing module calculates the critical safety distance for collision avoidance based on its own vehicle speed and the real-time distance between the vehicle and the obstacle, completes the collision avoidance decision based on the current longitudinal distance between the vehicle and the obstacle, and plans the decision-making method for the vehicle. If the distance allows braking to avoid collision, the vehicle implements the allowed deceleration to slow down. If the distance is too short to avoid collision, the vehicle will be forced to trigger several decisions such as alarms and emergency braking. Step 6: Establishment of the information processing system, which is responsible for collecting relevant information of the vehicle during driving and sending it to various systems for sharing; The information processing system consists of three parts: information collection layer, transmission layer and application layer. The collection layer is responsible for collecting the real-time operating parameters of the vehicle, road environment parameters and prediction parameters and converting them into electrical signals. The transmission layer is responsible for connecting various systems and providing information transmission services, that is, realizing the aggregation, analysis, processing and transmission of input and output data. The application layer refers to the integration of various information, the close integration of business logic and resources, and the presentation of information processing results to users in the form of rich multimedia.

2. The method for setting up an autonomous parking system for large electric trucks based on field-side collaboration according to claim 1, characterized in that: The autonomous driving vehicle is equipped with an on-board computer, which is also the control center of the autonomous driving vehicle. The on-board computer has built-in codes for the autonomous parking control system and trajectory tracking control system and the vehicle-side collision avoidance system that are input in advance. The on-board computer is connected to the Internet of Vehicles platform through the network, and the parking lot management module and the Internet of Vehicles platform are also connected through the network. The field sensors inside the parking lot collect vehicle driving information and parking space information in real time and send them to the Internet of Vehicles platform for everyone to see and use. At the same time, the high-precision map platform in the Internet of Vehicles platform will also share the detailed map of the parking lot for vehicles to use. The specific process is as follows: After the on-board computer receives the map information and parking space information published by the Internet of Vehicles platform, it shares it with the downlink system. First, the autonomous parking control system plans a reasonable driving path through the pre-established vehicle motion model and the received map information and parking space information. The path is transmitted to the trajectory tracking control system in the computer. The trajectory tracking control system starts the vehicle, and the vehicle-side sensors start running. The data from the combined effect of the on-board camera and the field-side sensors is published on the Internet of Vehicles platform and sent to the on-board computer at the same time to calculate the real-time distance between the vehicle and the obstacle. After the vehicle moves, the vehicle-side sensors monitor the vehicle and collect the The vehicle speed information and steering angle information are converted into a trajectory tracking control system that inputs electrical signals. The trajectory tracking control system is adjusted according to the driving path and position information, and the speed and steering angle at the next moment are calculated, and input to the vehicle through electrical signals. At the same time, the vehicle speed information, steering angle information, the distance between the vehicle and the obstacle, and the location map information are shared with the vehicle-side collision avoidance system. The vehicle-side collision avoidance system runs in real time to calculate whether it is a safe distance. If everything is normal, it continues to drive. If it is less than the safe distance, the vehicle-side collision avoidance system will directly transmit electrical signals to the braking system, motor and alarm switch to force braking, emergency stop of the motor and alarm.

3. The method for setting up an autonomous parking system for large electric trucks based on field-side collaboration according to claim 1, characterized in that: The Internet of Vehicles platform is connected via a network and is used to accept parking instructions from car owners. The information receiving module and information sending module of the Internet of Vehicles platform are simultaneously connected to the unmanned vehicle and the parking lot management module via a network, so as to realize the connection between the field-side sensor and the vehicle-side sensor, and is used to monitor the vehicle's driving path in all directions and count available parking spaces, obtain information about the vehicle's surrounding environment, and cooperate to guide the vehicle to avoid other vehicles and obstacles and drive into the parking space. The Internet of Vehicles platform is used to obtain available parking spaces from the parking lot management module and send them to the vehicle, and is also used to perform real-time control of vehicles parked in the parking lot; The information collation module in the Internet of Vehicles platform is responsible for the management of vehicles and parking spaces and the optimal recommendation of parking spaces. It includes three parts: parking space identification / positioning unit, parking space recommendation unit and parking management unit. The parking space identification and positioning unit uses artificial intelligence technology to identify parking spaces and obtain the number and location of vacant parking spaces. After the vacant parking spaces are identified and located, the Internet of Vehicles platform stores the parking space information in the server and waits for parking space application commands. When the autonomous vehicle issues a parking request command, it is transmitted to the IoV platform through the network layer. The parking space recommendation unit in the IoV platform uses a sorting algorithm that approximates the ideal solution to obtain the optimal sorting of free parking spaces, and sends the optimal parking space to the autonomous vehicle through the network. At the same time, the IoV platform will update the parking space database and no longer publish the recommended parking spaces to other vehicles. The high-precision map platform belongs to the downstream part of the Internet of Vehicles platform. It can output lane-level parking lot map data to support low-latency, high-precision path planning. The Internet of Vehicles platform transmits the vehicle's parking application command to the information receiving module, and then the high-precision map platform sends the parking lot's road condition information to the on-board navigation system in real time through the network. The on-board navigation system determines the feasible parking path based on the parking position and the real-time road condition information of the parking lot. Based on the parking space uploaded to the Internet of Vehicles platform and the positioning coordinates of the applying vehicle, the Internet of Vehicles platform will guide the vehicle to automatically drive to the parking space and complete parking by accessing the pre-stored map.

4. The method for setting up an autonomous parking system for large electric trucks based on field-side collaboration according to claim 1, characterized in that: The parking lot is a parking lot with the function of guiding and managing unmanned vehicles. In order to reasonably utilize parking spaces in the parking lot and reduce waste of resources, the parking lot management system is provided with a field-end parking space detection unit. The field-end parking space detection unit is a parking space geomagnetic detection device. The parking space geomagnetic detection device is composed of a geomagnetic sensor, a power supply, a single-chip microcomputer and a communication unit. The field-end parking space detection unit determines the parking status by detecting the disturbance of the earth's magnetic field near the parking space when the vehicle arrives at the parking space, and can upload the real-time parking space usage information to the parking lot management module. The field-end parking space detection unit is responsible for counting and reporting the information of parked vehicles in the parking lot, including license plate number, parking time, vehicle location and related information of autonomously parked vehicles when driving. After the field-end parking space detection unit collects the vehicle-related information, it is transmitted to the Internet of Vehicles platform in real time through the network. The roadside data unit is responsible for monitoring the vehicle driving information on the mobile phone and transmitting it to the Internet of Vehicles platform for use.

5. The method for setting up an autonomous parking system for large electric trucks based on field-side collaboration according to claim 1, characterized in that: The working process of the autonomous parking control system includes the following steps: Step 1: Determine the minimum turning radius: The minimum turning radius of a car refers to the radius of the circle that the center plane of the outer steering wheel of the car travels on the supporting plane when the front wheel steering angle reaches the maximum and the car makes a turning motion at the lowest stable speed. R A Indicates the minimum turning radius of the car, and its calculation formula is as follows: Where: L represents the vehicle wheelbase, θ max Indicates the maximum steering angle of the front wheels; Step 2: Establish a vehicle motion model: Take the center point of the rear axle of the vehicle as the reference point and establish the vehicle model based on the vehicle's own information. The bicycle model is established based on the following assumptions: 1) The movement of the vehicle in the vertical direction, i.e. the Z-axis direction, is not considered, that is, the movement of the vehicle is assumed to be a movement on a two-dimensional plane; 2) Assume that the left and right tires of the vehicle have the same steering angle and speed at any time; in this way, the movement of the left and right tires of the vehicle is combined into one tire to describe; 3) Assume that the vehicle speed changes slowly and ignore the transfer of front and rear axle loads; 4) Assume that the vehicle body and suspension system are both rigid systems; 5) Assume that the vehicle's movement and steering are driven by the front wheels; In the geodetic coordinate system OXY, the center point of the rear axle of the vehicle (x, y) is taken as the reference point, the speed of the center point of the rear axle is v, and the coordinates of the center point of the front axle of the vehicle are (x f ,y f ), the heading angle of the vehicle is θ, the wheelbase of the vehicle is L, the extension line of the rear axle of the vehicle and the extension line perpendicular to the front wheel turning angle intersect at a point P, which is the instantaneous steering center of the vehicle, and the distance between point Р and the center point of the rear axle is the steering radius R; The vehicle kinematic model is expressed as follows: δ is the equivalent turning angle of the front wheel. According to the geometric relationship, the angle corresponding to the instantaneous turning center of the vehicle is also δ. Therefore, the relationship between the equivalent turning angle of the front wheel and the wheelbase and turning radius is expressed as: The relationship between the coordinates of the center point of the front axle and the center point of the rear axle of the vehicle: x f =x+L·cosθ and f =y+L·sinθ The component speeds of the center point of the vehicle's rear axle on the horizontal and vertical axes in the geodetic coordinate system are: Since the vehicle does not skid, the vertical speed of the front and rear wheels is zero; From the above calculations, we can conclude that: In summary, the kinematic equation of the vehicle is: In this formula, [x, y, θ] T is the state quantity of the vehicle, indicating the position and posture of the vehicle, [v,δ] T To control the amount, the table makes the vehicle's speed and front wheel angle; Where (x, y) represents the coordinates of the vehicle's geometric center position; (x f ,y f ) represents the coordinates of the center position of the front wheel of the vehicle, δ is the equivalent turning angle of the front wheel; v represents the vehicle speed; L represents the wheelbase; θ represents the heading angle of the vehicle; R represents the radius of the circle followed by the rear axle at a given turning angle; Step 3, path planning: The HybridA* algorithm adds the initial posture point (x1, y1, θ1) to the open list, adds the node with the highest priority in the open list to the closed list, and determines whether the current node is less than the threshold. If it is less than the threshold, the Reeds-Shepp curve or Dubins curve is used to connect with the target posture point (x2, y2, θ2), and perform collision detection. If there is no collision with the obstacle, the curve is retained and a path is generated. If there is a collision with the obstacle, the curve is abandoned, and the node is searched again from the open list and re-expanded. Finally, a feasible path from the starting state to the target state that meets the vehicle kinematic constraints is generated. The specific process is as follows: (1) First, the vehicle path planning system in the on-board computer makes a judgment based on the relevant map provided by the terminal, the current positioning information, and the target storage location. The control end converts the map information provided by the terminal into a two-dimensional grid map that can be used by the vehicle, and obtains the initial posture (x1, y1, θ1) and target posture (x2, y2, θ2) of the vehicle; (2) The Hybrid A* algorithm selects the maximum of the two as the final heuristic information, H(n) = max(Reeds-Shepp, A*). The first type of "considering constraints but not obstacles" is to prevent the robot from reaching the target from the wrong direction, and the second type of "considering obstacles but not constraints" is to prevent wasting time in dead ends or U-shaped obstacles. (3) The Reeds-Shepp curve used in (2) above not only explores the vehicle's heuristic information, but also provides an interval path between two nodes that conforms to the vehicle's kinematic model in the subsequent node expansion; In the path of the Reeds-Shepp curve, cusps are allowed to exist. To indicate cusps, superscripts need to be added to the letters in the field to indicate the direction of movement; The differential equation according to the vehicle kinematic model is as follows: Its integral form is as follows: The points are: ψ(t)=t The corresponding paths are as follows: In order to be more precise, The somewhat redundant convention is used in that the sign of t matches the direction; If t>0: the vehicle moves forward, the symbol is +, if t<0: the vehicle moves backward, the symbol is -; The Reeds and Shepp curve has 48 fields in total. The "|" in the phrase indicates that the vehicle's direction of motion changes from forward to reverse or from reverse to forward. Some of these 48 fields have two formulas for the actual path of the field type. For any given situation, there are at most 68 formulas. The unification of position and posture is as follows: Before calculation, the vehicle's posture must be normalized. Assume that the vehicle's initial posture is (x1, y1, θ1), the target posture is (x2, y2, θ2), and the vehicle's radius is r = ρ. The normalization process is the process of vector translation and rotation, so that the transformed starting posture is (0, 0, 0), the target posture is (x, y, φ), and the vehicle's turning radius is r = 1; x=((x2-x1)*cos(θ1)+(y2-y1)*sin(θ1)) / ρ y=(-(x2-x1)*sin(θ1)+(y2-y1)*cos(θ1)) / ρ φ=θ2-θ1 The symmetric relationship is used to reduce the complexity of the solution, as follows: The principle of solving the Reeds-Shepp curve is that after changing the target point, it is equivalent to the original target point A being symmetrical about the y-axis and the x-axis, and a new target point B is obtained, and a set of new paths t, u, v are generated, using timefilp symmetry, that is, sign symmetry and reflect symmetry; This simplifies the calculation; Step 4, collision detection: replace the vehicle body with three overlapping circles. For each state, calculate whether the distance from the obstacle grid point to the center of the circle is greater than the radius to perform collision detection. The union of the three circles contains the vehicle body. Just calculate whether the distance from the obstacle point in the grid to the center of the circle is less than the radius. If it is less than the radius, it means that a collision occurs, otherwise no collision occurs. The calculation amount is small and the accuracy is high.

6. The method for setting up an autonomous parking system for large electric trucks based on field-side collaboration according to claim 1, characterized in that: The specific working process of the trajectory tracking control system is as follows: According to the given path, it is calculated as a reference trajectory with continuous curvature. The front and rear axles of the vehicle are both used as control points. The lateral error is controlled by the angle tracking error with the center of the front axle of the vehicle as the control point, and the longitudinal error is controlled by the speed tracking error with the center of the rear axle of the vehicle as the control point. The controller changes the motor power by controlling the motor voltage to achieve the control of the speed and the front wheel angle, decoupling the control of the vehicle angle and vehicle speed to the dual-axis control with the center of the front and rear axles as the control points, achieving the separation control of the vehicle speed and the angular speed, reducing the mutual influence of the two controls, and making full use of the freedom of the front and rear axles of the vehicle. The process includes: Step 1: According to the real-time posture of the vehicle and the posture of the parking point, a vehicle kinematic model with speed and angular velocity constraints is combined to obtain a reference posture sequence of the vehicle with the rear axle as the geometric center; Step 2: Convert the geometric center point reference pose sequence into the vehicle's rear axle center point reference trajectory through the pose transformation relationship; Step 3: The vehicle kinematic model is expressed as follows: y(t+Δt)=y(t)+v t sin(δ t )Δt x(t+Δt)=x(t)+v t cos(δ t )Δt Among them, (x, y) represents the coordinates of the geometric center of the vehicle; δ represents the angle between the speed and the horizontal line; β represents the front wheel turning angle; α represents the angle between the vehicle body and the target point; t represents a certain moment; Δt represents the time interval; L represents the wheelbase; L f represents the foresight distance; v t represents the speed at time t; R represents the radius of the circle followed by the rear axle at a given turning angle; Step 4: The speed and corner speed constraints are expressed as follows: V(t+Δt)=V(t)+aΔt V≤V max β≤β max V represents the speed value; V max Assume the maximum speed; β max The maximum front wheel turning angle is set; Step 5: The turning radius and foresight distance constraints are expressed as follows: R≤R max L f =kv c Among them, R max is the maximum turning radius of the vehicle; k represents the proportionality coefficient between the forward sight distance and the speed; Step 6: The posture transformation relationship is expressed by the following formula: Among them, x f With y f Indicates the position coordinates of the center point of the front axle; x r With y r Indicates the position coordinates of the center point of the rear axle; According to the reference trajectory of the rear axle center, the preview method of feedback control is adopted to obtain the rear wheel angle control value; The process includes: Calculate the distance d from the center point of the rear axle to the closest point A on the reference trajectory of the center point of the rear axle; Calculate the distance from this point to satisfy the foresight distance L f The next point B; Take point B of the reference trajectory of the vehicle's rear axle center point as the rear wheel preview point P; The front wheel turning angle β of the vehicle is calculated based on the preset rear wheel preview coefficient k, the distance d, and the position coordinates of the rear wheel preview point P; Finally, the vehicle updates its own status information in real time, updates the current speed and steering angle, obtains the speed control value of the controlled object through the PID controller, and finally drives into the parking space.

7. The method for setting up an autonomous parking system for large electric trucks based on field-side collaboration according to claim 1, characterized in that: The working process of the vehicle-end collision avoidance system is as follows: The vehicle-side collision avoidance system consists of three parts: signal acquisition module, data processing module and execution module. The working process of the specific part includes the following steps: Step 1: The signal acquisition module works: The signal acquisition module uses radar and camera technology to automatically measure the vehicle speed and the distance to the obstacle; Step 2: Work of the data processing module: The data processing module processes the instantaneous speed and determines the safe distance. If it is less than the safe distance, it issues a command to alarm or brake. The specific process is as follows: The coordinate system is established with the center of mass of the vehicle as the reference point; the vehicle collision time TTC is introduced to reflect the degree of danger of the driving condition. -1 , combined with the driving information and the critical indicator of the vehicle's danger level, the coordinated collision avoidance braking deceleration threshold a is calculated and output by the following formula x-max In the above formula, v m is the initial speed of the vehicle, v n is the speed of the obstacle. Here, the obstacle is considered as a fixed object with no speed, so v n = 0, S0 is the longitudinal distance between the vehicle and the obstacle, and 0.8 is the TTC of the vehicle in danger level -1 critical indicators; Based on the ideal braking deceleration a' x and braking deceleration threshold a x-max , after comparison, the expected braking deceleration a for cooperative collision avoidance under the current road conditions is output x-fine : when |a' x |≤|a x-max |, let a x-fine =a' x ; when |a' x |>|a x-max |, let a x-fine =a x-max ; Based on the expected braking deceleration a of cooperative collision avoidance x-fine , complete the collaborative collision avoidance trajectory planning, expressed as: In the above formula, x'(t) and y'(t) are the coordinates of the vehicle at time t, v m is the initial speed of the vehicle, t' m is the completion time of collaborative collision avoidance, X(t' m ) is the longitudinal displacement of the center of mass of the ego vehicle when the coordinated collision avoidance is completed; According to the planned trajectory, considering that the ego vehicle smoothly and effectively completes the collision avoidance, the critical collision scenario in the collision avoidance process is analyzed, and the critical safety distance D of collision avoidance is calculated; Based on the critical safety distance Dcooperative for cooperative collision avoidance, the relationship between the longitudinal distance S0 between the ego vehicle and the obstacle collected in real time and the critical safety distance Dcooperative for cooperative collision avoidance is compared; Step 3: Work of the execution module: The execution module receives the instructions of the data processing module and implements them, issuing instructions for alarm, closing windows and automatic braking; If S0<D coordination, collision is unavoidable, control the vehicle to brake as hard as possible to reduce collision damage; execute the module and issue a corresponding alarm; If S0>D coordination, the collision can be avoided and the vehicle continues to move.

8. The method for setting up an autonomous parking system for large electric trucks based on field-side collaboration according to claim 1, characterized in that: The working process of the information processing system is as follows: The information processing system consists of three layers, from low to high, they are the collection layer, the transmission layer and the application layer. The working steps of each layer are as follows: Step 1, the work of the collection layer: the collection layer is responsible for collecting the vehicle speed, steering angle, positioning information, mileage, engine speed and temperature inside the vehicle. The collection layer is completed by vehicle sensors. The sensors on the transmission include speed sensors, temperature sensors, shaft speed sensors and pressure sensors. The sensors on the steering gear include angle sensors, torque sensors and hydraulic sensors. The sensors on the suspension include speed sensors, acceleration sensors, body height sensors, roll angle sensors and angle sensors. The vehicle sensor is the input device of the vehicle computer system. It converts various working condition information of the vehicle, vehicle speed, temperature of various media and engine operating conditions into electrical signals and inputs them into the vehicle computer. The collected data will also be uploaded to the backend server for unified processing and analysis to obtain the business data required by the user, providing reliable data support for the vehicle to complete subsequent work; Step 2: Transport layer work: The transport layer is composed of network servers and WEB services. From the perspective of communication and information processing, the transport layer provides communication services to the application layer above. The transport layer is located above the network layer. It provides logical communication between processes running on different hosts, while the network layer provides logical communication between hosts. Even if the network layer protocol is unreliable, the transport layer can still provide reliable services for applications. The transport layer provides logical communication between application processes, that is, end-to-end communication. Logical communication means that the communication between the transport layers is to transmit data in the horizontal direction, but in fact there is no horizontal physical connection between the two transport layers. When the transport layer uses connection-oriented TCP, this logical communication channel is equivalent to a full-duplex reliable channel. The server manages the data statistics and provides corresponding services for each vehicle. At the same time, it can jointly analyze the data and become a local Internet of Vehicles service business. Step 3: Work of the application layer: The application layer removes the inherent format of all the transmission data of the network transmission layer, that is, the application of the remaining Data part after the SF, FF, CF flag bits, Len and padding bits, and gives the Data part practical meaning; The network layer is responsible for data transmission, while the application layer is concerned with obtaining information from the data. The request in the application layer refers to the information sent from the client to the server; the response in the application layer refers to the information sent from the server to the client. The client refers to the diagnostic instrument, host computer and test tools, and the server refers to the BCM, GW and PEPS body electronic components. The standard application layer service refers to ISO 15765-3 and ISO 14229-1 documents and is used in testing, monitoring, diagnosis and vehicle online refresh scenarios.

Images