An autonomous driving vehicle high-speed cruise control method and an electronic device

Abstract

The present invention discloses a high-speed cruise control method and an electronic device for an autonomous vehicle, including: First, obtain real-time vehicle state information and surrounding environment information through sensors and vehicle network communication devices equipped on the autonomous vehicle; secondly, establish a vehicle dynamics model; then establish a hierarchical control strategy for vehicle lane keeping or lane changing, and determine the actions that the vehicle needs to perform according to the surrounding traffic information; propose a reference path for the vehicle to perform lane-changing operations based on a polynomial method to enable the vehicle to perform lane-changing operations while maintaining a safe distance; finally, adopt a deep reinforcement learning method, learn the lane-changing behaviors of skilled drivers, combine with the reference path to determine the optimal cruise path, and complete lane-changing and cruise control based on the vehicle motion control module. This application improves the cruise control stability and safety of autonomous vehicles during high-speed driving and realizes autonomous cruise control of the vehicle.

Description

Technical Field

[0001] The present invention relates to the field of intelligent vehicle driving, and particularly to a high-speed cruise control method and an electronic device for an autonomous vehicle. Background Art

[0002] With the development of automotive electrification and intelligent technologies, autonomous driving, as the most important application of automotive intelligence, is the focus of the development of intelligent vehicles. The cruise control technology of vehicles can effectively reduce the burden on drivers, relieve driver fatigue, improve traffic conditions and reduce environmental pollution. However, complex traffic scenarios greatly challenge the safety and comfort of autonomous vehicles. At present, the cruise control technologies of vehicles include proportional-integral-derivative (PID) control algorithms, adaptive control algorithms, etc., which are mainly for cruise control during the longitudinal driving of vehicles. However, these two control algorithms often rely on accurate models and have the disadvantage of low cruise control accuracy in complex environments.

[0003] The research content of the cruise control technology of autonomous vehicles on highways mainly includes three aspects: (1) vehicle behavior decision-making; (2) vehicle path planning; (3) vehicle motion control. Therefore, the technical research on the above three aspects is the focus of the high-speed cruise control of autonomous vehicles, and also provides a theoretical basis and technical support for the research on the cruise control of future connected vehicle fleets. However, how to ensure the economy of vehicle cruise control, as well as the safety and comfort during vehicle lane-changing, is still extremely challenging. Summary of the Invention

[0004] The present invention proposes a high-speed cruise control method and an electronic device for an autonomous vehicle, realizing the economical and safe driving of a driverless vehicle on a highway. It is proposed to learn the constant speed cruise and vehicle lane-changing strategies from a skilled driver through a deep reinforcement learning algorithm, guide the autonomous vehicle to make an effective decision on the high-speed driving cruise strategy, and keep the vehicle stable in the high-speed cruise scenario.

[0005] To solve the above technical problems, the present invention adopts the following technical solutions:

[0006] Provide a high-speed cruise control method for an autonomous vehicle, including the following steps:

[0007] Step S1: Obtain the real-time state information and road environment information of the autonomous vehicle, and perform real-time precise positioning on the autonomous vehicle in combination with the positioning system;

[0008] Step S2: Upload the obtained real-time state information and road environment information of the autonomous vehicle in Step S1 to the vehicle control unit;

[0009] Step S3: Establish a vehicle dynamics model, which can characterize the longitudinal, lateral, and yaw characteristics of the vehicle;

[0010] Specifically as follows:

[0011] The established vehicle dynamics model is:

[0012] (Formula 1)

[0013] In the formula, the state vector , , , , , , respectively represent the longitudinal speed of the vehicle, the lateral speed of the vehicle, the yaw angle of the vehicle, the yaw angular velocity of the vehicle, the ordinate of the vehicle's earth coordinate system, and the abscissa of the vehicle's earth coordinate system in sequence; the control vector , and represent the longitudinal acceleration of the vehicle and the front wheel steering angle respectively; C is the state parameter matrix, D is the control parameter matrix;

[0014] Step S4: Establish a cruise control decision-making module for the autonomous driving vehicle. By judging whether there is a vehicle to follow in front of the autonomous driving vehicle, the speed of the vehicle to follow, and the relative distance between the autonomous driving vehicle and the vehicle to follow, formulate a hierarchical control strategy for the autonomous driving vehicle to maintain the lane or change lanes;

[0015] Step S5: Establish a motion control method for the autonomous driving vehicle. For the vehicle cruise control decision-making module established in Step S4, formulate a vehicle motion control strategy, design an optimal path planning method for the vehicle, determine the optimal reference path, and perform vehicle motion control to achieve autonomous cruise driving of the vehicle.

[0016] Furthermore, in Step S1, the real-time vehicle state information includes the speed, acceleration, and position of the autonomous driving vehicle; the road environment information includes the road width, road slope, section speed limit, traffic environment information, and the position and speed information of surrounding vehicles.

[0017] Furthermore, in Step S2, upload the real-time information of the autonomous driving vehicle and the road environment information to the control unit. To improve the information storage and processing efficiency, only upload the data within the communication range of 200 meters of the autonomous driving vehicle and the surrounding environment (including road information, other vehicle state information, speed limit information, etc.).

[0018] Furthermore, the establishment of the cruise control decision-making module for the autonomous driving vehicle in Step S4 mainly includes the following steps:

[0019] First, determine whether there is a vehicle that can be followed ahead;

[0020] If there is no following vehicle, then according to the road speed limit regulations, perform lane-keeping constant-speed cruise driving at a preset economic vehicle speed;

[0021] If there is a following vehicle, then based on the relative speed between the vehicle ahead and the autonomous vehicle, and the relative distance between the vehicle ahead and the autonomous vehicle, determine whether to keep following the vehicle in the lane or change lanes for the autonomous vehicle.

[0022] The detailed steps are as follows:

[0023] Step S41: According to the obtained real-time status information of the autonomous vehicle and the road environment information, detect whether there is a vehicle that can be followed ahead of the autonomous vehicle. If there is no vehicle that can be followed ahead, then perform lane-keeping and perform constant-speed cruise driving according to the economic vehicle speed designed for the autonomous vehicle;

[0024] Step S42: According to the obtained real-time status information of the autonomous vehicle and the road environment information, detect whether there is a vehicle that can be followed ahead of the autonomous vehicle. If there is a following vehicle ahead, then first perform lane-keeping and perform constant-speed cruise driving according to the economic vehicle speed designed for the autonomous vehicle, and then determine whether to change lanes based on the relative speed between the vehicle ahead and the autonomous vehicle and the relative distance between the vehicle ahead and the autonomous vehicle;

[0025] Step S43: According to the obtained real-time status information of the autonomous vehicle and the road environment information, if the speed of the following vehicle is lower than the set economic cruise vehicle speed of the autonomous vehicle, then based on the real-time obtained real-time status information of the autonomous vehicle and the surrounding environment information, determine whether there are conditions for changing lanes; if so, perform a lane-changing operation, if not, exit the constant-speed cruise and decelerate to avoid collision;

[0026] Step S44: If the distance between the autonomous vehicle and the following vehicle is lower than the set safety threshold, then based on the real-time obtained status information of the autonomous vehicle and the surrounding environment information, determine whether there are conditions for changing lanes; if so, perform a lane-changing operation, if not, exit the constant-speed cruise and decelerate to avoid collision;

[0027] Furthermore, the cruise mode switching strategy proposed in step S4 is as follows:

[0028] Step S411: Set the economic cruise vehicle speed as , the speed of following the vehicle ahead as , with the unit of km / h. If there is no following vehicle ahead, then perform lane-keeping and constant-speed cruise; the designed economic vehicle speed here is 100 km / h;

[0029] Step S412: If there is a vehicle to follow ahead, when, then maintain the lane and engage in cruise control;

[0030] Step S413: If there is a vehicle to follow ahead, when , then determine whether the distance between the two vehicles exceeds the set safety threshold. If not, maintain the lane; because , the distance between the two vehicles will continue to decrease. If the distance between the two vehicles exceeds the set safety threshold, then determine whether there are lane-changing conditions based on the real-time information of the autonomous vehicle and the road environment information. If not, use the relative distance between the two vehicles as the control target and follow the vehicle at a constant speed for cruise control;

[0031] Step S414: Based on Step S413, if there are lane-changing conditions, then control the autonomous vehicle to perform a lane-changing operation.

[0032] Furthermore, the autonomous vehicle motion control method proposed in Step S5 includes the following steps:

[0033] Step S51: According to the fact that the stability of lane-changing behavior is crucial when the autonomous vehicle is cruising at high speed, propose an inverse reinforcement learning method to learn the steering behavior of experienced drivers and improve the feasibility of the vehicle's planned path and the stability of the vehicle;

[0034] Step S52: According to the characteristics of the vehicle driving on the high-speed lane, construct a candidate path expressed by a polynomial as follows:

[0035] (Formula 2)

[0036] In the formula, is the vehicle's lateral reference position, X is the vehicle's longitudinal coordinate, , , , , and are polynomial coefficients,

[0037] Define the initial position of the vehicle as , and the lateral distance and longitudinal distance involved in the lane-changing process are represented by L and M respectively. If = 0, = 0, then there is the following relationship for the terminal position :

[0038] (Formula 3)

[0039] In the formula, L is the lane width,M is the longitudinal distance for lane change. During the vehicle lane change process, the longitudinal speed of the vehicle remains unchanged, and the lane change time is , and the longitudinal speed of the vehicle is , then:

[0040] (Formula 4)

[0041] Step S53: Use the reverse deep learning algorithm to evaluate the overall performance during the vehicle lane change process, including vehicle safety and driving comfort. Vehicle safety is quantified by the vehicle stable state and potential collision risk.

[0042] Furthermore, the quantization index of the vehicle stability state in step S53 is:

[0043] (Formula 5)

[0044] In the formula, is the stability index, is the lateral speed of the vehicle, is the longitudinal speed of the vehicle.

[0045] Furthermore, the quantization index of the vehicle potential collision risk is:

[0046] (Formula 6)

[0047] In the formula, is the collision risk index, and are the weight factors, and are the lateral coordinate and longitudinal coordinate of the autonomous vehicle, and are the lateral coordinate and longitudinal coordinate of the collision risk vehicle.

[0048] Furthermore, construct the following vector to define the comprehensive performance index of vehicle control as:

[0049] (Formula 7)

[0050] (Formula 8)

[0051] In the formula, is the comprehensive performance index, is the stability index, is the collision risk index, is the total number of sample points collected for each candidate path, n is the total number of collision risk vehicles.

[0052] Normalize Formula 7, and the normalization function is expressed as:

[0053] (Formula 9)

[0054] The reward function in the reverse deep learning algorithm is expressed as:

[0055] (Formula 10)

[0056] In the formula, is the reward function, is the weight coefficient, , which is set as a constant.

[0057] Furthermore, it further includes step S54: taking the driving behavior of experienced drivers during lane change as expert experience to optimize the weight coefficients and , and training in the reverse deep learning algorithm, so that the driving vehicle exhibits lane-changing behaviors similar to those of proficient drivers.

[0058] Furthermore, an electronic device for an autonomous driving vehicle is provided, which is used to implement the above-mentioned autonomous driving vehicle high-speed cruise control method, and includes a memory and a navigator; the memory stores a computer program, and the software program in the memory can run on the vehicle central processing unit; the navigator is used to determine the current position, vehicle speed, and road speed limit of the vehicle.

[0059] Compared with the prior art, the beneficial effects of the present invention are:

[0060] 1. The autonomous driving vehicle high-speed cruise control method and electronic device proposed by the present invention propose an overall framework for autonomous cruise control of autonomous driving vehicles on highways. This framework inherits decision-making, path planning, and motion control modules, and the proposed framework can ensure the performance of high-speed cruise vehicles.

[0061] 2. The present invention proposes a vehicle lane-changing and motion control strategy based on a deep reinforcement learning algorithm. This strategy is to learn the ability of vehicle path planning and vehicle control from proficient drivers. This method takes into account the driving safety and economy of the vehicle, and ensures smooth and safe lane-changing behaviors when the vehicle executes lane changes.

[0062] 3. The technical solution of the present invention solves the technical problems of low control accuracy and poor stability in cruise control in the related art, enhances the adaptability of vehicle cruise control to complex working conditions, and provides a technical solution for researching cooperative cruise control applications between multiple vehicles. Description of the Drawings

[0063] Figure 1It is a framework diagram of the high-speed cruise control method for autonomous vehicles;

[0064] Figure 2 It is a vehicle dynamics model diagram;

[0065] Figure 3 It is a framework diagram of the skilled driver's lane-changing behavior training based on deep reinforcement learning. Specific implementation manners

[0066] The following describes the specific implementation manners of the present invention in conjunction with the accompanying drawings, so that those skilled in the art can better understand the present invention.

[0067] It should be noted that:

[0068] Figure 1 As shown, it is a framework diagram of the high-speed cruise control method for autonomous vehicles proposed by the present invention, including an environmental information acquisition module, a vehicle lane-changing decision-making module, a path planning module, and a motion control module.

[0069] The environmental information acquisition module mainly includes the real-time state information of the autonomous vehicle and the road environment information. The real-time state information of the autonomous vehicle specifically includes: vehicle speed, vehicle acceleration, and vehicle position information; the road environment information specifically includes: road width, road slope, section speed limit, and traffic environment information.

[0070] The vehicle lane-changing decision-making module includes two aspects: lane keeping and vehicle lane-changing.

[0071] The path planning module includes three aspects: reference path, actual path, and path tracker design.

[0072] The vehicle motion control module refers to developing a path tracker to ensure the feasibility and stability of the vehicle driving according to the planned path.

[0073] Figure 2 As shown, it is a vehicle dynamics model, and a general bicycle model is used to describe the dynamic characteristics of the vehicle.

[0074] A cruise control method for autonomous vehicles proposed in this application includes the following steps:

[0075] Step S1: Obtain the real-time state information of the autonomous vehicle and the road environment information, and perform real-time precise positioning of the autonomous vehicle in combination with the positioning system; the vehicle real-time state information includes the speed, acceleration, and position of the autonomous vehicle; the road environment information includes road width, road slope, section speed limit, traffic environment information, and the position and speed information of surrounding vehicles.

[0076] Step S2: Upload the real-time status information and road environment information of the autonomous vehicle obtained in Step S1 to the vehicle control unit; when uploading the real-time information of the autonomous vehicle and the road environment information to the control unit, to improve the information storage and processing efficiency, only the data within the communication range of 200 meters of the autonomous vehicle and the surrounding environment (including road information, other vehicle status information, speed limit information, etc.) is uploaded.

[0077] Step S3: Establish a vehicle dynamics model, which can characterize the longitudinal, lateral, and yaw characteristics of the vehicle;

[0078] Step S4: Establish a cruise control decision-making module for the autonomous vehicle;

[0079] Step S5: Establish a motion control method for the autonomous vehicle; for the vehicle cruise control decision-making module established in Step S4, formulate a vehicle motion control strategy, design an optimal path planning method for the vehicle, determine the optimal reference path, and perform vehicle motion control to achieve autonomous cruise driving of the vehicle.

[0080] Among them, in Step S3: The vehicle dynamics model established is:

[0081] (Formula 1)

[0082] In the formula, the state vector , , , , , , respectively represent the longitudinal speed of the vehicle, the lateral speed of the vehicle, the yaw angle of the vehicle, the yaw angular velocity of the vehicle, the ordinate of the vehicle in the geodetic coordinate system, and the abscissa of the vehicle in the geodetic coordinate system; the control vector , and respectively represent the longitudinal acceleration of the vehicle and the front wheel steering angle; C is the state parameter matrix, D is the control parameter matrix.

[0083] In Step S4: To establish a cruise control decision-making module for the autonomous vehicle, it mainly includes the following steps:

[0084] The detailed steps are as follows:

[0085] Step S41: According to the obtained real-time status information and road environment information of the autonomous vehicle, detect whether there is a following vehicle in front of the autonomous vehicle. If there is no following vehicle in front, perform lane keeping and cruise at a constant speed according to the designed economic speed of the autonomous vehicle;

[0086] Step S42: Based on the acquired real-time status information of the autonomous vehicle and road environment information, detect whether there is a following vehicle in front of the autonomous vehicle. If there is a following vehicle in front, first maintain the lane, cruise at a constant speed according to the designed economic speed of the autonomous vehicle, and then determine whether a lane change is needed based on the relative speed between the vehicle in front and the autonomous vehicle and the relative distance between the vehicle in front and the autonomous vehicle;

[0087] Step S43: Based on the acquired real-time status information of the autonomous vehicle and road environment information, if the speed of the vehicle that can be followed is lower than the designed economic cruise speed of the autonomous vehicle, then determine whether there are lane-changing conditions based on the acquired real-time status information of the autonomous vehicle and the surrounding environment information; if so, perform a lane-changing operation, if not, exit the cruise and decelerate to avoid collision;

[0088] Step S44: If the distance between the autonomous vehicle and the following vehicle is lower than the set safety threshold, then determine whether there are lane-changing conditions based on the real-time acquired status information of the autonomous vehicle and the surrounding environment information; if so, perform a lane-changing operation, if not, exit the cruise and decelerate to avoid collision;

[0089] Among them, the cruise mode switching strategy proposed in Step S4 is as follows:

[0090] Step S411: Set the economic cruise speed as and the speed of following the vehicle in front as , with the unit of km / h. If there is no following vehicle in front, then maintain the lane and cruise at a constant speed. Here, the designed economic cruise speed is 100 km / h;

[0091] Step S412: If there is a vehicle that can be followed in front, when , then maintain the lane and cruise at a constant speed;

[0092] Step S413: If there is a vehicle that can be followed in front, when , then determine whether the distance between the two vehicles exceeds the set safety threshold. If not, then maintain the lane; because , the distance between the two vehicles will continuously decrease. If the distance between the two vehicles exceeds the set safety threshold, then determine whether there are lane-changing conditions based on the acquired real-time autonomous vehicle information and road environment information. If not, then use the relative distance between the two vehicles as the control target and follow the vehicle in front for cruising;

[0093] Step S414: Based on Step S413, if there are lane-changing conditions, then control the autonomous vehicle to perform a lane-changing operation.

[0094] The autonomous vehicle motion control method proposed in Step S5 includes the following steps:

[0095] Step S51: According to the feature that the stability of lane-changing behavior is crucial when an autonomous vehicle is cruising at high speed, an inverse reinforcement learning method is proposed to learn the steering behavior of experienced drivers, so as to improve the feasibility of the vehicle's planned path and the stability of the vehicle.

[0096] Step S52: According to the characteristics of the vehicle driving on the highway lane, construct a candidate path expressed by polynomials as follows:

[0097] (Formula 2)

[0098] In the formula, is the vehicle's lateral reference position, X is the vehicle's longitudinal coordinate, , , , , and are polynomial coefficients,

[0099] Define the initial position of the vehicle as , the lateral distance and longitudinal distance involved in the lane-changing process are represented by L and M respectively. If = 0, = 0, then the terminal position has the following relationship:

[0100] (Formula 3)

[0101] In the formula, L is the lane width, M is the longitudinal distance of lane change. During the lane-changing process of the vehicle, the longitudinal speed of the vehicle remains unchanged, the lane-changing time is , and the longitudinal speed of the vehicle is , then:

[0102] (Formula 4)

[0103] Step S53: Use the reverse deep learning algorithm to evaluate the overall performance during the vehicle lane-changing process, including vehicle safety and driving comfort. The vehicle safety is quantified by the vehicle's stable state and potential collision risk.

[0104] The quantization index of the vehicle stability state is:

[0105] (Formula 5)

[0106] In the formula, is the stability index, is the vehicle's lateral speed, is the longitudinal speed of the vehicle.

[0107] The vehicle potential collision risk quantification index is:

[0108] (Formula 6)

[0109] In the formula, is the collision risk index, and are weight factors, and are the lateral coordinate and longitudinal coordinate of the autonomous vehicle, and are the lateral coordinate and longitudinal coordinate of the collision risk vehicle.

[0110] Construct the following vector to define the comprehensive performance index of vehicle control as:

[0111] (Formula 7)

[0112] (Formula 8)

[0113] In the formula, is the comprehensive performance index, is the stability index, is the collision risk index, is the total number of sample points collected for each candidate path, n is the total number of collision risk vehicles.

[0114] Perform normalization processing on Formula 7, and the normalization function is expressed as:

[0115] (Formula 9)

[0116] The reward function in the reverse depth algorithm is expressed as:

[0117] (Formula 10)

[0118] In the formula, is the reward function, is the weight coefficient, is the transpose of, , set as a constant.

[0119] Step S54: Take the driving behavior of an experienced driver during lane change as expert experience to optimize the weight coefficients and , training is carried out in the reverse deep learning algorithm, so that the driving vehicle exhibits lane-changing behaviors similar to those of a skilled driver. The proposed training framework diagram for the lane-changing behavior of a skilled driver based on deep reinforcement learning is as Figure 3 shown.

[0120] Figure 3 There are three contents shown as follows, including ① Based on the constructed polynomial to represent candidate paths, applying the deep reinforcement learning method to learn the path planning experience of skilled drivers and formulating a path planning method for autonomous driving vehicles; ② Combining the vehicle dynamics model and the path planning scheme determined in ① to establish a cruise control lane-changing strategy and formulate a reward function for the path planning and motion control of autonomous driving vehicles; ③ According to the cruise and lane-changing strategies determined in ② and the determined reward function, carrying out the deep reinforcement learning architecture design and weight coefficient optimization for the cruise control of autonomous driving vehicles, and finally forming the high-speed cruise control ability of autonomous driving vehicles. Among them, the lane-changing strategy and path planning strategy are designed based on the deep reinforcement learning system framework, which improves the motion stability of autonomous driving vehicles and the probability of potential collision avoidance of vehicles.

[0121] This application also provides an electronic device for an autonomous driving vehicle, including a memory and a navigator. The memory stores computer programs, the processor is the central processing unit of the autonomous driving vehicle, and the software programs in the memory can run on the in-vehicle central processor. The navigator is used to determine the current position of the vehicle, the vehicle speed, and the road speed limit.

[0122] It should be noted that the various module units proposed in the present invention can be integrated into one processing unit, or each unit can exist independently, or two or more can be integrated into one unit.

[0123] The advantages of this example are as follows:

[0124] To improve the safety and economy of the cruise driving of autonomous driving vehicles, an overall framework for autonomous cruise control of autonomous driving vehicles on highways is proposed. This framework inherits the decision-making, path planning, and motion control modules, and the proposed framework can ensure the performance of high-speed cruise vehicles.

[0125] The present invention solves the problems of low control accuracy and poor stability existing in the existing related cruise control technologies, enhances the adaptability of the vehicle cruise control method to complex working conditions, and improves the lane-changing safety and driving comfort. The simulation results confirm that the proposed method can effectively reduce the vehicle collision risk while achieving stable control.

[0126] The above are only the feasible implementation examples of the present invention, and do not limit the scope of the rights of the present invention accordingly. All equivalent structural changes made by using the content of this invention book and the attached drawings are included in the scope of the rights of the present invention.

Claims

1. A high-speed cruise control method for an autonomous driving vehicle, characterized in that: The following steps are involved: Step S1: Acquire the real-time status information and road environment information of the autonomous driving vehicle, and accurately locate the autonomous driving vehicle in real time in combination with the positioning system; Step S2: uploading the real-time status information of the autonomous driving vehicle and the road environment information obtained in step S1 to the vehicle control unit; Step S3: Establish a vehicle dynamics model, which can characterize the longitudinal, lateral and yaw characteristics of the vehicle, as follows: The vehicle dynamics model is established as: (Formula 1) In the formula, the state vector , , , , , , The control vectors respectively represent the longitudinal velocity of the vehicle, the lateral velocity of the vehicle, the yaw angle of the vehicle, the yaw rate of the vehicle, the ordinate of the vehicle geodetic coordinate system, and the abscissa of the vehicle geodetic coordinate system. , and They represent the vehicle longitudinal acceleration and the front wheel steering angle respectively; C is the state parameter matrix, D is the control parameter matrix; Step S4: establishing a cruise control decision module for the autonomous driving vehicle, formulating a hierarchical control strategy for lane keeping or lane changing of the autonomous driving vehicle by judging whether there is a followable vehicle in front of the autonomous driving vehicle, the speed of the followable vehicle, and the relative distance between the autonomous driving vehicle and the followable vehicle; Step S5: Establish an autonomous driving vehicle motion control method, formulate a vehicle motion control strategy for the vehicle cruise control decision module established in step S4, design a vehicle optimal path planning method, determine the optimal reference path, and perform vehicle motion control to achieve autonomous cruising of the vehicle.

2. The high-speed cruise control method for an autonomous driving vehicle according to claim 1, characterized in that: The vehicle real-time status information described in step S1 includes the speed, acceleration and position of the autonomous driving vehicle; the road environment information includes road width, road slope, traffic environment information, and the position and speed information of surrounding vehicles.

3. The high-speed cruise control method for an autonomous driving vehicle according to claim 1, characterized in that: The road environment information uploaded to the vehicle-mounted control unit in step S2 is data within a 200-meter communication range around the driving vehicle.

4. The method for high-speed cruise control of an autonomous driving vehicle according to claim 1, characterized in that: The step S4 of establishing the autonomous driving vehicle cruise control decision module mainly includes the following steps: Step S41: Based on the acquired real-time status information and road environment information of the autonomous driving vehicle, detect whether there is a vehicle that can be followed in front of the autonomous driving vehicle. If there is no vehicle that can be followed in front, perform lane keeping and cruise at an economic speed designed for the autonomous driving vehicle. If there is a vehicle that can be followed in front, first perform lane keeping and cruise at an economic speed designed for the autonomous driving vehicle. Then, determine whether lane changing is required based on the relative speed of the vehicle in front and the autonomous driving vehicle, and the relative distance between the vehicle in front and the autonomous driving vehicle, and formulate a cruise mode switching strategy. Step S42: The cruise mode switching strategy includes: according to the acquired real-time status information of the autonomous driving vehicle and the road environment information, if the speed of the followable vehicle is less than the economic cruise speed set by the autonomous driving vehicle, judging whether there is a lane change condition according to the real-time status information of the autonomous driving vehicle and the surrounding environment information acquired in real time; if so, performing a lane change operation; if not, exiting the cruise control and slowing down to avoid collision; Step S43: The cruise mode switching strategy includes: if the distance between the autonomous driving vehicle and the following vehicle is less than the set safety threshold, judging whether there are lane change conditions based on the real-time autonomous driving vehicle status information and surrounding environment information; if so, performing lane change operation; if not, exiting cruise control and slowing down to avoid collision.

5. The high-speed cruise control method for an autonomous driving vehicle according to claim 4, characterized in that: The cruise mode switching strategy described in steps S41-S43 is as follows: Step S411: Set the economic cruising speed to , the speed of the vehicle being followed is , in km / h. If there is no vehicle ahead that can follow, the vehicle will maintain the lane and cruise at a constant speed; Step S412: If there is a vehicle ahead that can be followed, , then lane keeping and cruise control; Step S413: If there is a vehicle ahead that can be followed, , then determine whether the distance between the two vehicles is greater than the set safety threshold. If it is greater than the safety threshold, the lane is maintained; , the distance between the two vehicles will continue to decrease. If the distance between the two vehicles is less than the set safety threshold, the real-time autonomous driving vehicle and road environment information obtained will be used to determine whether there are lane change conditions. If not, the relative distance between the two vehicles will be used as the control target, and the vehicle will cruise; Step S414: Based on step S413, if there are lane changing conditions, the automatic driving vehicle is controlled to perform lane changing operations.

6. The method for high-speed cruise control of an autonomous driving vehicle according to claim 1, characterized in that: The automatic driving vehicle motion control method proposed in step S5 is characterized by comprising the following steps: Step S51: According to the characteristic that the stability of lane changing behavior is crucial when the autonomous driving vehicle is cruising at high speed, a method based on inverse reinforcement learning is proposed to learn the steering behavior of experienced drivers, thereby improving the feasibility of the vehicle's planned path and the stability of the vehicle; Step S52: Based on the driving characteristics of the vehicle in the high-speed lane, a candidate path expressed in a polynomial is constructed as follows: (Formula 2) In the formula, is the vehicle lateral reference position, X is the longitudinal coordinate of the vehicle, , , , , and are the polynomial coefficients, The initial position of the vehicle is defined as , the lateral distance and longitudinal distance involved in the lane change process are respectively L and M Indicates that if =0, =0, then the terminal position The following relationship exists: (Formula 3) In the formula, L is the lane width, M is the longitudinal distance of lane change. During the lane change process, the longitudinal speed of the vehicle remains unchanged. The lane change time is , the longitudinal speed of the vehicle is ,but: (Formula 4) Step S53: Use the reverse deep learning algorithm to evaluate the overall performance of the vehicle during the lane change process, including vehicle safety and driving comfort. Vehicle safety is quantified by vehicle stability status and potential collision risk.

7. The method for high-speed cruise control of an autonomous driving vehicle according to claim 6, characterized in that: In step S53 The vehicle stability state quantitative index is: (Formula 5) In the formula, is a stability indicator, is the vehicle lateral velocity, is the vehicle longitudinal velocity; The quantitative index of the potential collision risk of the vehicle is: (Formula 6) In the formula, is the collision risk indicator, and is the weight factor, and are the lateral and longitudinal coordinates of the autonomous vehicle, and are the lateral and longitudinal coordinates of the collision risk vehicle; The following vector is constructed to define the comprehensive performance index of vehicle control: (Formula 7) (Formula 8) In the formula, As a comprehensive performance indicator, is a stability indicator, is the collision risk indicator, is the total number of sample points collected for each candidate path, n is the total number of vehicles at risk of collision; Formula 7 is normalized, and the normalized function is expressed as: (Formula 9) The reward function in the inverse depth algorithm is expressed as: (Formula 10) In the formula, is the reward function, is the weight coefficient, for The transpose of , set to a constant.

8. The method for high-speed cruise control of an autonomous driving vehicle according to claim 7, characterized in that: The step S54 is also included: using the driving behavior of experienced drivers in lane change as expert experience to optimize the weight coefficients in step S53 and , trained in an inverse deep learning algorithm so that the driving vehicle exhibits lane-changing behavior similar to that of an experienced driver.

9. An electronic device for an autonomous driving vehicle, used to implement the method for high-speed cruise control of an autonomous driving vehicle according to any one of claims 1 to 8, characterized in that: It includes a memory and a navigator. The memory stores a computer program. The computer program in the memory can be run on the central processing unit of the autonomous driving vehicle. The navigator is used to determine the current position, vehicle speed and road speed limit of the vehicle.

Images