Unmanned bus path planning method, device and equipment and storage medium

Abstract

The invention discloses an unmanned bus path planning method, device and equipment and a storage medium, relates to the technical field of path planning, and discloses an unmanned bus path planning method comprising the following steps: constructing a multi-target model; determining a rule adherence weight, a personnel injury weight and a regional movement style weight of the multi-target model according to the road end sensing information; determining global path planning information according to the rule adherence weight, the personnel injury weight and the regional movement style weight; determining local path planning information according to the environment perception information; and completing unmanned bus path planning according to the global path planning information and the local path planning information. According to the scheme, the safety of path planning of the unmanned bus in a complex environment can be kept, meanwhile, the anthropomorphism of path planning is improved, and safe, efficient and stable driving of the unmanned bus in the whole driving process is achieved.

Description

Technical Field

[0001] The present application relates to the technical field of path planning, and particularly to an unmanned bus path planning method, device, equipment and storage medium. Background Art

[0002] With the continuous development of autonomous driving technology, the vehicle traveling control method has become one of the crucial research topics in the intelligent transportation system. The vehicle traveling control method based on trajectory tracking is of great significance in intelligent transportation applications such as autonomous driving and driverless buses. Especially in the driverless scenario, how to ensure the safe driving of the vehicle through trajectory tracking and path planning has become an important topic in the industry development.

[0003] In the prior art, some solutions generate a vehicle traveling control strategy by calculating the trajectory tracking result of the vehicle and the target path of the vehicle itself, and combining the deviation coefficient and the human body distance value around the vehicle, but there are potential safety hazards in complex urban environments. Some solutions propose a dispatching simulation system for driverless buses based on reinforcement learning, and use a deep reinforcement learning algorithm to optimize the dispatching strategy, but there is a lack of sufficient anthropomorphic consideration for crowded areas, schools, etc., which may lead to ethical safety hazards. How to improve the anthropomorphism of route planning while maintaining the safety of the path planning of driverless buses in complex environments has become a problem to be solved.

[0004] The above content is only used to assist in understanding the technical solution of the present application, and does not represent an admission that the above content is the prior art. Summary of the Invention

[0005] The main purpose of the present application is to provide an invention name, aiming to solve the technical problem of how to improve the anthropomorphism of route planning while maintaining the safety of the path planning of driverless buses in complex environments.

[0006] To achieve the above purpose, the present application proposes an unmanned bus path planning method, and the method includes:

[0007] Construct a multi-objective model, and determine the rule compliance weight, personnel injury weight, and regional movement style weight of the multi-objective model according to the roadside perception information;

[0008] Determine the global path planning information according to the rule compliance weight, the personnel injury weight, and the regional movement style weight;

[0009] Determine the local path planning information according to the environmental perception information;

[0010] Complete the unmanned bus path planning according to the global path planning information and the local path planning information.

[0011] In one embodiment, the step of determining the rule compliance weight, the personnel injury weight, and the regional movement style weight of the multi-objective model according to the roadside perception information includes:

[0012] Determine the road segment type information, the stationary obstacle information on the road segment, and the dynamic obstacle information on the road segment according to the roadside perception information;

[0013] Determine the rule compliance weight, the personnel injury weight, and the regional movement style weight of the multi-objective model according to the road segment type information, the stationary obstacle information, and the dynamic obstacle information.

[0014] In one embodiment, the step of determining the rule compliance weight, the personnel injury weight, and the regional movement style weight of the multi-objective model according to the road segment type information, the stationary obstacle information, and the dynamic obstacle information includes:

[0015] Evaluate the importance of rule compliance according to the road segment type information to obtain the rule compliance weight of the multi-objective model;

[0016] Evaluate the potential injury risk according to the stationary obstacle information and the dynamic obstacle information to obtain the personnel injury weight of the multi-objective model;

[0017] Evaluate the traffic order style according to the dynamic obstacle information to obtain the regional movement style weight of the multi-objective model.

[0018] In one embodiment, the step of determining the local path planning information according to the environmental perception information includes:

[0019] Determine the leading vehicle state variables, the ego vehicle state variables, and the regional environmental variables according to the environmental perception information;

[0020] Determine the predicted acceleration of the leading vehicle and the control change amount of the ego vehicle according to the leading vehicle state variables, the ego vehicle state variables, and the regional environmental variables;

[0021] Determine the local path planning information according to the predicted acceleration of the leading vehicle and the control change amount of the ego vehicle.

[0022] In one embodiment, the step of determining the leading vehicle state variables, the ego vehicle state variables, and the regional environmental variables according to the environmental perception information includes:

[0023] Obtain the longitudinal speed of the leading vehicle, the lateral offset of the leading vehicle, the relative distance of the leading vehicle, the longitudinal speed of the ego vehicle, the control lateral angle of the ego vehicle, the longitudinal acceleration of the ego vehicle, and the weather information according to the environmental perception information;

[0024] Determine the leading vehicle state variables according to the longitudinal speed of the leading vehicle, the lateral offset of the leading vehicle, and the relative distance of the leading vehicle;

[0025] Determine the ego vehicle state variables based on the ego vehicle longitudinal speed, the ego vehicle controlled lateral angle, and the ego vehicle longitudinal acceleration;

[0026] Determine the regional environmental variables according to the weather information.

[0027] In one embodiment, the step of determining the predicted acceleration of the leading vehicle and the control change amount of the ego vehicle according to the leading vehicle state variables, the ego vehicle state variables, and the regional environmental variables includes:

[0028] Predict the behavior of the leading vehicle according to the leading vehicle state variables and the regional environmental variables through a leading vehicle behavior prediction network;

[0029] Add model random noise to the behavior of the leading vehicle to obtain the predicted acceleration of the leading vehicle;

[0030] Determine the control change amount of the ego vehicle according to the ego vehicle state variables, the predicted acceleration of the leading vehicle, and the regional environmental variables through an ego vehicle control response network.

[0031] In one embodiment, the step of determining the local path planning information according to the predicted acceleration of the leading vehicle and the control change amount of the ego vehicle includes:

[0032] Calculate the leading vehicle acceleration comparison term based on the predicted acceleration of the leading vehicle through a time decay weight function and a safety interval weighted norm;

[0033] Calculate the ego vehicle steering response comparison term based on the control change amount of the ego vehicle through a causal response delay function and a curvature adaptive weighted norm;

[0034] Obtain a causal loss comparison function according to the comparison term weight coefficient, the leading vehicle acceleration comparison term, and the ego vehicle steering response comparison term;

[0035] Optimize the leading vehicle behavior prediction network and the ego vehicle control response network based on the causal loss comparison function to obtain local path planning information.

[0036] In addition, to achieve the above object, the present application also proposes an unmanned bus path planning device, and the unmanned bus path planning device includes:

[0037] A data processing module, configured to construct a multi-object model and determine the rule compliance weight, the personnel injury weight, and the regional movement style weight of the multi-object model according to the roadside perception information;

[0038] A global planning module, configured to determine global path planning information according to the rule compliance weight, the personnel injury weight, and the regional movement style weight;

[0039] A local planning module for determining local path planning information based on environmental perception information;

[0040] A path planning module for completing the path planning of the driverless bus according to the global path planning information and the local path planning information.

[0041] In addition, to achieve the above object, the present application also proposes a driverless bus path planning device, which includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, and the computer program is configured to implement the steps of the driverless bus path planning method as described above.

[0042] In addition, to achieve the above object, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium, and when the computer program is executed by a processor, the steps of the driverless bus path planning method as described above are implemented.

[0043] In addition, to achieve the above object, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the driverless bus path planning method as described above are implemented.

[0044] One or more technical solutions proposed by the present application have at least the following technical effects:

[0045] By constructing a multi-objective model and determining the rule compliance weight, personnel injury weight, and regional movement style weight of the multi-objective model according to the roadside perception information; determining the global path planning information according to the rule compliance weight, the personnel injury weight, and the regional movement style weight; determining the local path planning information according to the environmental perception information; and completing the path planning of the driverless bus according to the global path planning information and the local path planning information. By performing global path planning on the driverless bus, the general driving direction of the vehicle is obtained, and by performing local path planning on the driverless bus, the real-time driving trajectory of the vehicle is obtained. By combining the global path planning information and the local path planning information, the planned path can not only ensure that the overall driving route of the driverless bus conforms to the absoluteness of respecting rules, minimizing personnel injuries, and adaptability of movement style, but also enable the driverless bus to respond to local environmental changes in real time, and can improve the anthropomorphism of route planning while maintaining the safety of the driverless bus path planning in a complex environment, realizing the safe, efficient, and stable driving of the driverless bus during the entire driving process. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The accompanying drawings herein are incorporated into the specification and form a part of the specification, showing embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application.

[0047] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0048] Figure 1 It is a schematic flowchart provided for the first embodiment of the method for path planning of the driverless bus in the present application;

[0049] Figure 2 It is a schematic flowchart provided for the second embodiment of the method for path planning of the driverless bus in the present application;

[0050] Figure 3 It is a schematic brief flowchart of the method for path planning of the driverless bus provided for the second embodiment of the present application;

[0051] Figure 4 It is a schematic module structure diagram of the device for path planning of the driverless bus in the embodiment of the present application;

[0052] Figure 5 It is a schematic device structure diagram of the hardware operating environment involved in the method for path planning of the driverless bus in the embodiment of the present application.

[0053] The implementation, functional features, and advantages of the objectives of the present application will be further described with reference to the embodiments and the drawings. Specific embodiments

[0054] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0055] To better understand the technical solutions of the present application, the following will be described in detail with reference to the drawings in the specification and specific embodiments.

[0056] The main solution of the embodiment of the present application is: constructing a multi-objective model, and determining the rule compliance weight, personnel injury weight, and regional movement style weight of the multi-objective model according to the roadside perception information; determining the global path planning information according to the rule compliance weight, the personnel injury weight, and the regional movement style weight; determining the local path planning information according to the environment perception information; and completing the path planning of the driverless bus according to the global path planning information and the local path planning information.

[0057] In this embodiment, for the convenience of description, the following will be described with the vehicle end recognition as the execution subject.

[0058] In the prior art, some solutions generate a driving control strategy for a vehicle by calculating the trajectory tracking result of the vehicle and the target path of the host vehicle, and combining the deviation coefficient and the human body distance value around the vehicle. However, there are safety hazards in complex urban environments. Some solutions propose a scheduling simulation system for driverless buses based on reinforcement learning, and use a deep reinforcement learning algorithm to optimize the scheduling strategy. However, there is insufficient anthropomorphic consideration for crowded areas, schools, etc., which may lead to ethical safety hazards. How to improve the anthropomorphism of route planning while maintaining the safety of the path planning of driverless buses in complex environments has become a problem to be solved.

[0059] The present application provides a solution. By performing global path planning on the driverless bus to obtain the general driving direction of the vehicle, and performing local path planning on the driverless bus to obtain the real-time driving trajectory of the vehicle, and combining the global path planning information and the local path planning information, the obtained planned path can not only ensure that the overall driving route of the driverless bus conforms to the absoluteness of respecting rules, minimizing human injury, and adapting to the movement style, but also enable the driverless bus to respond to changes in the local environment in real time, and can improve the anthropomorphism of route planning while maintaining the safety of the path planning of the driverless bus in complex environments, so as to realize the safe, efficient and stable driving of the driverless bus during the entire driving process.

[0060] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication, and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device, a vehicle terminal, etc. that can implement the above functions. Hereinafter, the vehicle terminal is taken as an example to illustrate this embodiment and the following embodiments.

[0061] Based on this, the embodiment of the present application provides a method for path planning of a driverless bus, referring to Figure 1 , Figure 1 which is a schematic flowchart of the first embodiment of the method for path planning of a driverless bus in the present application.

[0062] In this embodiment, the method for path planning of the driverless bus includes steps S10 to S40:

[0063] Step S10, constructing a multi-objective model, and determining the rule compliance weight, the human injury weight, and the regional movement style weight of the multi-objective model according to the road-end perception information;

[0064] It should be noted that the roadside perception information can be collected in real time through roadside sensors and processed in real time by a roadside controller. The roadside sensors include information perception devices such as lidar, cameras, millimeter-wave radars, and global positioning systems fixedly deployed on road infrastructure. The roadside controller is installed behind the roadside sensors and is used to integrate and analyze the raw data collected by the roadside sensors to obtain structured roadside perception information and transmit it to the vehicle end of the driverless bus. Among them, the lidar can measure the distance to an object through a laser beam, generate high-precision three-dimensional point cloud data, and identify obstacle, road sign, and road surface condition information; the camera can capture videos and images to identify road markings, traffic signs, traffic lights, pedestrians, vehicles, etc. information; the millimeter-wave radar can detect the distance, speed, and relative movement of vehicles and pedestrians in low visibility environments (such as haze, rainy days); the global positioning system can accurately provide the location information of the road section.

[0065] It should be understood that the multi-objective model is a path planning model that comprehensively considers ethical, safety, and social factors and balances different objectives through dynamic weights. During the process of path planning for the driverless bus, for a fixed starting point and destination, multiple candidate paths will be generated. According to the multi-objective model, multiple different objective factors, such as rule compliance, personnel injury, and regional driving style, can be comprehensively considered to score the multiple candidate paths and select the path that best meets ethical and social expectations to obtain the panoramic path planning decision.

[0066] It should be noted that the rule compliance weight is an indicator used to measure the importance of rule compliance of candidate paths in the multi-objective model, the personnel injury weight is an indicator used to measure the degree of personnel injury of candidate paths in the multi-objective model, and the regional movement style weight is an indicator used to measure the vehicle driving and personnel movement styles in candidate paths in the multi-objective model. Inputting the roadside perception information of the candidate path into the multi-objective model can obtain the values corresponding to the rule compliance weight, personnel injury weight, and regional movement style weight of the multi-objective model.

[0067] In a feasible implementation manner, step S10 may include steps S11 to S12: [[ID=II]]

[0068] Step S11, determining the road section type information, road section static obstacle information, and road section dynamic obstacle information according to the roadside perception information;

[0069] It should be noted that the road section type information is information about the nature of the road section determined through roadside perception information, including function type information and necessary marking information. The function type information is used to classify the personnel activity functions of the road section, including platform road sections, school road sections, hospital road sections, commercial district road sections, residential area road sections, highway road sections, etc. The necessary marking information is used to mark the degree of passage required for the road section, including mandatory road sections, regularly passed road sections, and non-passable road sections. By identifying the scene of the road section based on the roadside perception information, the function type information can be obtained. By detecting the roadside perception information and matching it with the mandatory platform information of the driverless bus, the mandatory road sections, regularly passed road sections, and non-passable road sections in the candidate path can be identified, and the necessary marking information can be obtained.

[0070] In addition, it should be noted that the road section static obstacle information includes the position information and shape information of all static obstacles on each road section in the candidate path. The static obstacles can be vehicles parked by the roadside, enclosures for road construction, traffic signs, buildings, etc. The road section dynamic obstacle information includes the position information, type, speed information, and moving direction information of dynamic obstacles on the road. The types of dynamic obstacles include vehicles, pedestrians, and animals, etc.

[0071] Step S12: Determine the rule compliance weight, personnel injury weight, and regional movement style weight of the multi-objective model according to the road section type information, the static obstacle information, and the dynamic obstacle information.

[0072] It should be noted that by combining the road section type information, the static obstacle information, and the dynamic obstacle information, the importance of the driverless bus's compliance with rules on each road section of each candidate path can be evaluated to obtain the rule compliance weight of the multi-objective model, the potential injury risk to personnel on each road section of each candidate path can be evaluated to obtain the personnel injury weight, and the moving speed and rule compliance degree of vehicles and personnel on each road section of each candidate path can be evaluated to obtain the regional movement style weight.

[0073] In a feasible implementation manner, step S12 may include steps S121 to S123:

[0074] Step S121: Evaluate the importance of rule compliance according to the road section type information to obtain the rule compliance weight of the multi-objective model;

[0075] It should be understood that the importance of rule compliance refers to the degree of importance of an unmanned bus complying with traffic rules during driving on a certain section of the road. In sections with strict traffic restrictions (such as sections where passage is prohibited), the importance of rule compliance is relatively high. The corresponding rule compliance weight is larger, and the smaller the rule compliance weight, the higher the selection priority of the corresponding section. In particular, for the necessary sections of the path planning of the unmanned bus, the rule compliance weight is 0, and for sections where passage is impossible, the rule compliance weight is 1. The vehicle end maintains a rule compliance weight query table. After obtaining the functional type information of a certain section and the necessary marking information based on the section type information, it will first determine whether the section is a necessary section or a non-passable section through the necessary marking information. If the section is a necessary section or a non-passable section, the rule compliance weight is the corresponding special value 0 or 1. If the section is not a necessary section for the unmanned bus, the rule compliance weight is queried from the rule compliance weight query table based on the functional type information and time information, where the time information is used to judge the degree of crowd density.

[0076] Exemplarily, assume that W d is the rule compliance weight. For sections where passage is prohibited, W d = 1; for intersections and school sections at the school gate during crowded times, W d = 0.9; for sections with few people and normal vehicle passage, W d = 0.05; for the platform sections that the unmanned bus must pass through, W d = 0.

[0077] Step S122, evaluate the potential injury risk according to the static obstacle information and the dynamic obstacle information, and obtain the personnel injury weight of the multi-objective model;

[0078] It should be understood that the potential injury risk refers to the degree of risk of an unmanned bus causing injury to personnel during driving on a certain section of the road. The smaller the degree of risk of causing injury to personnel, the smaller the personnel injury weight, and the higher the selection priority of the corresponding section. According to the static obstacle information and the dynamic obstacle information, the static obstacle position information, the static obstacle shape information, the dynamic obstacle position information, the dynamic obstacle speed information, the dynamic obstacle moving direction information, and the dynamic obstacle type information can be obtained.

[0079] It should be noted that by combining the static obstacle position information, the static obstacle shape information, the dynamic obstacle position information, the dynamic obstacle speed information, the dynamic obstacle moving direction information, and the dynamic obstacle type information, the degree of injury to personnel (including passengers and pedestrians) when the unmanned bus collides during driving on this section of the road can be predicted, and the estimated injury value can be obtained. Based on the injury coefficient, the estimated injury values of each person in the section are summed to obtain the personnel injury weight of the multi-objective model.

[0080] In addition, it should be noted that the injury coefficient can be preset according to the attributes of pedestrians and passengers. For vulnerable groups such as children and the elderly among pedestrians, their injury coefficients are relatively high; for ordinary adults, the injury coefficients are relatively low. For passengers, it is set according to factors such as their positions in the vehicle and the use of safety equipment. For example, the injury coefficient of a passenger sitting in the front row without wearing a seatbelt is higher than that of a passenger sitting in the back row and wearing a seatbelt.

[0081] Exemplarily, assume that U is the personnel injury weight, and U = |-∑(c i ·h i )|. Among them, c i represents the injury coefficient for humans (pedestrians, passengers), and h i represents the estimated injury value.

[0082] Step S123, evaluate the traffic order style according to the dynamic obstacle information to obtain the regional movement style weight of the multi-objective model.

[0083] It should be noted that by combining the traffic signal data at the road end and the dynamic obstacle information, pedestrian movement style data and vehicle driving habit data can be obtained, such as the average walking speed of pedestrians, the compliance frequency with traffic lights when crossing the road, the average vehicle speed, the vehicle lane-changing frequency, and whether non-motor vehicles often drive in the motor vehicle lane. According to the pedestrian movement style data and vehicle driving habit data, the traffic order style of this section can be evaluated to obtain the regional movement style weight of the multi-objective model. The better the traffic order, the lower the regional movement style weight, and the greater the selection priority of the corresponding section.

[0084] Exemplarily, assume that W c is the regional movement style weight. For sections with a slow pace and where pedestrians and vehicles highly comply with traffic rules, W c = 0.15; for sections with a fast pace, where non-motor vehicles often drive in the motor vehicle lane and pedestrians frequently cross the lane, W c = 0.75.

[0085] Step S20, determine the global path planning information according to the rule compliance weight, the personnel injury weight, and the regional movement style weight;

[0086] It should be noted that after inputting the road-end perception information of the candidate path into the multi-objective model and obtaining the values corresponding to the rule compliance weight, personnel injury weight, and regional movement style weight of the multi-objective model, the scene risk level in the global planning process can be evaluated by combining the time information and scene type information, and the dynamic weight coefficients corresponding to the rule compliance weight, personnel injury weight, and regional movement style weight can be adjusted accordingly. According to the global path planning model, the rule compliance weight, personnel injury weight, and regional movement style weight are weighted and summed to obtain a multi-objective score value. Based on the multi-objective score value, the candidate paths are scored, and the path that best meets the ethical and social expectations of the current scene is selected from the candidate paths as the target path from the starting point to the destination, and the route navigation information of the target path is obtained to obtain the global path planning information.

[0087] Exemplarily, the construction formula of the multi-objective score value J is as follows:

[0088] J = α·(W d ·R d ) + β·U + γ·W c ·S

[0089] In the formula, W d represents the rule compliance weight, representing the ethical weight index; U represents the personnel injury weight, representing the utilitarian weight index; W c represents the regional movement style weight, representing the pedestrian movement style and vehicle driving habits index of the road section. R d represents the degree of violation of the ethical weight (for example, if traffic regulations are violated, it is counted as 1, and if traffic regulations are not violated, it is counted as 0). S represents the social and cultural acceptability. α, β, and γ represent dynamic weights, which are dynamically adjusted by the scene risk level of the global planning in different times and different scenes. For example, in high-risk scenes such as traffic peaks and complex intersections, the weights related to safety (such as β related to the risk of personnel injury and α related to the ethical weight) will be dynamically increased, while the weights directly related to traffic efficiency (such as γ representing the pedestrian movement style and vehicle driving habits of the road section) will be appropriately reduced to ensure that the driverless bus gives priority to ensuring driving safety in high-risk scenes.

[0090] It should be understood that in the process of global path planning, dynamically balancing the rule compliance weight, personnel injury weight, and regional movement style weight can minimize the damage loss while complying with the rules as much as possible, and ensure traffic safety and traffic efficiency.

[0091] Step S30, determining the local path planning information according to the environmental perception information;

[0092] It should be understood that the environmental perception information is various information about the environment around the driverless bus sensed by in-vehicle sensors, such as the state of the vehicle in front, the state of the own vehicle, and the regional environment, etc. These information are used to determine the local path planning and obtain the local path planning information. The in-vehicle sensors include a forward perception camera installed inside the front windshield of the vehicle, a fisheye camera installed around the vehicle, lidar installed at the four corners of the vehicle, a wheel speed sensor, a steering wheel angle sensor, and an acceleration sensor carried inside the vehicle. The original data collected by the in-vehicle sensors will be transmitted to the intelligent driving controller of the driverless bus, and the environmental perception information will be obtained after data processing.

[0093] Step S40, complete the path planning of the driverless bus according to the global path planning information and the local path planning information.

[0094] It should be understood that during the driving process of the driverless bus, the pre-planned global path planning information will be used as the basic driving route, and at the same time, the local path planning information will be obtained in real time and integrated with the global path planning information. According to the integrated path planning information, control the driving direction, speed, etc. of the driverless bus to make it drive along the path planned by the global path planning information.

[0095] Exemplarily, when driving to an intersection, if the local path planning information indicates that there are pedestrians crossing the road ahead, while the global path planning information shows that the intersection is a straight-ahead route, then according to the combination of the two, control the driverless bus to decelerate or stop at the intersection, and continue to drive along the global route after the pedestrians have passed.

[0096] In this embodiment, by constructing a multi-objective model, and determining the rule compliance weight, the personnel injury weight, and the regional movement style weight of the multi-objective model according to the roadside perception information; determining the global path planning information according to the rule compliance weight, the personnel injury weight, and the regional movement style weight; determining the local path planning information according to the environmental perception information; completing the path planning of the driverless bus according to the global path planning information and the local path planning information. By performing global path planning on the driverless bus, the large driving direction of the vehicle is obtained, and by performing local path planning on the driverless bus, the real-time driving trajectory of the vehicle is obtained. By combining the global path planning information and the local path planning information, the obtained planned path can not only ensure that the overall driving route of the driverless bus conforms to the absoluteness of respecting rules, minimizing personnel injuries, and adaptability of movement style, but also enable the driverless bus to respond to local environmental changes in real time, and can improve the anthropomorphism of route planning while maintaining the safety of the path planning of the driverless bus in a complex environment, realizing the safe, efficient, and stable driving of the driverless bus during the entire driving process.

[0097] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar content as that in the above-mentioned first embodiment can be referred to the above introduction and will not be repeated hereinafter. On this basis, please refer to Figure 2 , step S30 may include steps S31 to S33:

[0098] Step S31, determining the leading vehicle state variable, the ego-vehicle state variable, and the regional environment variable according to the environmental perception information;

[0099] It should be understood that the leading vehicle state variable is a variable used to describe the driving state of the leading vehicle, including the longitudinal speed of the leading vehicle, the lateral offset of the leading vehicle, the relative distance of the leading vehicle, etc. The leading vehicle state variable is a parameter set describing the real-time dynamics of the target vehicle in front. The speed, position, and distance from the vehicle to the ego-vehicle of the leading vehicle can be obtained through in-vehicle sensors. The leading vehicle state variable is used to evaluate the driving behavior and possible changes of the leading vehicle, which affects the local path planning of the driverless bus. For example, when the leading vehicle suddenly brakes or changes lanes, the driverless bus needs to adjust the driving path in time according to the leading vehicle state variable.

[0100] In addition, it should be understood that the ego-vehicle state variable is a parameter set describing the real-time dynamics of the driverless bus, including the longitudinal speed of the ego-vehicle, the lateral control angle of the ego-vehicle, the longitudinal acceleration of the ego-vehicle, etc. The speed, steering wheel angle, and acceleration of the driverless bus can be detected through the sensors installed in the vehicle to obtain the ego-vehicle state variable. The ego-vehicle state variable reflects the current driving situation of the driverless bus and is used to determine the control change amount of the ego-vehicle in the local path planning to adjust the driving path and speed.

[0101] In addition, it should be understood that the regional environment variable is the dynamic parameter of the external environment where the driverless bus is located, such as weather information, road surface friction coefficient, etc. The weather information includes sunny, rainy, snowy, etc. Different weather conditions and friction coefficients will affect the driving performance and safety of the vehicle. For example, when the road surface is wet in rainy days, the braking distance will increase.

[0102] In a feasible implementation manner, step S31 may include steps S311 to S314:

[0103] Step S311, obtaining the longitudinal speed of the leading vehicle, the lateral offset of the leading vehicle, the relative distance of the leading vehicle, the longitudinal speed of the ego-vehicle, the lateral control angle of the ego-vehicle, the longitudinal acceleration of the ego-vehicle, and the weather information according to the environmental perception information;

[0104] It should be understood that the longitudinal speed of the leading vehicle is the speed of the leading vehicle of the driverless bus in the driving direction, the lateral offset of the leading vehicle is the lateral distance of the leading vehicle of the driverless bus relative to the center line of the lane, and the relative distance of the leading vehicle is the longitudinal distance between the driverless bus and the leading vehicle. The longitudinal speed of the host vehicle is the speed of the driverless bus in the driving direction, the lateral control angle of the host vehicle refers to the steering angle of the driverless bus steering wheel, and the longitudinal acceleration of the host vehicle is the acceleration of the driverless bus in the driving direction. The weather information includes comprehensive information of meteorological elements such as rainfall, snowfall, temperature, humidity, visibility, etc. The road surface friction coefficient can be obtained according to the weather information and the road surface information, and the regional environmental variable can be obtained.

[0105] Step S312, determining a leading vehicle state variable according to the longitudinal speed of the leading vehicle, the lateral offset of the leading vehicle, and the relative distance of the leading vehicle;

[0106] Exemplarily, the leading vehicle state variable can be defined as Y t , then the leading vehicle state variable Y t has the following expression:

[0107] Y t = [v front , E front , d front

[0108] In the formula, v front represents the longitudinal speed of the leading vehicle, that is, the longitudinal speed of the forward target vehicle of the driverless bus. E ftont represents the lateral offset of the leading vehicle, that is, the lateral offset of the forward target vehicle of the driverless bus. d front represents the relative distance of the leading vehicle, that is, the relative distance between the driverless bus host vehicle and the forward target vehicle.

[0109] Step S313, determining a host vehicle state variable according to the longitudinal speed of the host vehicle, the lateral control angle of the host vehicle, and the longitudinal acceleration of the host vehicle;

[0110] Exemplarily, the host vehicle state variable can be defined as X t , then the host vehicle state variable is X t has the following expression:

[0111] X t = [v self , E self , a self

[0112] In the formula, v self represents the longitudinal speed of the host vehicle, E self represents the lateral control angle of the host vehicle, and a self represents the longitudinal acceleration of the host vehicle.

[0113] Step S314: Determine the regional environmental variables according to the weather information.

[0114] Exemplarily, the regional environmental variable can be defined as C t , then the expression of the regional environmental variable C t is as follows:

[0115] C t = [weather.friction]

[0116] In the formula, weather represents the weather conditions, including rainfall, snowfall, and temperature, etc., and friction represents the road surface friction coefficient. C t represents the road surface friction coefficient corresponding to different weather conditions such as rainy days and snowy days.

[0117] Step S32: Determine the predicted acceleration of the leading vehicle and the control change amount of the host vehicle according to the leading vehicle state variable, the host vehicle state variable, and the regional environmental variable;

[0118] It should be understood that through the state variables of the leading vehicle and the host vehicle (i.e., the leading vehicle state variable and the host vehicle state variable), combined with the environmental factors where the host vehicle is located (i.e., the regional environmental variable), the future behavior of the leading vehicle and the control requirements of the host vehicle can be deduced. The predicted acceleration of the leading vehicle is the possible acceleration (such as accelerating, decelerating, or maintaining a constant speed) of the leading vehicle at the next moment predicted based on the current dynamic information of the leading vehicle (i.e., the leading vehicle state variable) and combined with environmental factors such as traffic flow and road conditions.

[0119] In addition, it should be understood that the driverless bus can determine the corresponding control change amount according to the predicted behavior of the leading vehicle (i.e., the predicted acceleration of the leading vehicle) and its own dynamic state (i.e., the host vehicle state variable), and obtain the control change amount of the host vehicle to execute control decisions such as lane keeping, lane changing, obstacle avoidance, and turning of the host vehicle. The control change amount of the host vehicle can be the change amount for adjusting the lateral angle of the host vehicle (i.e., the steering wheel angle), which is used to control the trajectory and driving direction of the host vehicle.

[0120] In a feasible implementation manner, step S32 may include: predicting the behavior of the leading vehicle by the leading vehicle behavior prediction network according to the leading vehicle state variable and the regional environmental variable; adding model random noise to the behavior of the leading vehicle to obtain the predicted acceleration of the leading vehicle; determining the control change amount of the host vehicle by the host vehicle control response network according to the host vehicle state variable, the predicted acceleration of the leading vehicle, and the regional environmental variable.

[0121] It should be understood that the leading vehicle behavior prediction network is a neural network model based on deep learning, which is used to predict the future behavior of the target vehicle ahead (such as accelerating, decelerating). The input is the leading vehicle state variable and the regional environmental variable, and the output is the predicted behavior of the leading vehicle (such as acceleration). Among them, the model random noise is the random perturbation artificially added in the leading vehicle behavior prediction network to simulate the real - world uncertainty. The ego - vehicle control response network is a neural network based on causal reasoning, which is used to generate real - time control instructions for the ego - vehicle. The input is the ego - vehicle state variable, the predicted acceleration of the leading vehicle, and the regional environmental variable, and the output is the change amount of the ego - vehicle control.

[0122] It should be noted that the neural network based on causal reasoning can generate accurate ego - vehicle control instructions by analyzing the causal relationship among the ego - vehicle state variable, the predicted acceleration of the leading vehicle, and the regional environmental variable. Through causal reasoning, the neural network can identify the actual impact of different factors (ego - vehicle state variable, predicted acceleration of the leading vehicle, and regional environmental variable) on the change amount of the ego - vehicle control, so as to realize the reasonable prediction and adjustment of the ego - vehicle behavior. This can not only improve the accuracy of control decisions, but also make more effective responses in complex dynamic environments, optimizing the driving safety and efficiency of the ego - vehicle. By constructing a dynamic causal reasoning model, the causal chain decoupling of the front - target of the driverless bus, the environmental state of the driving area of the driverless bus, and the driving state and behavior intention of the driverless bus itself is realized, and at the same time, real - time online incremental learning is supported.

[0123] Exemplarily, the predicted acceleration of the leading vehicle The prediction formula is as follows:

[0124]

[0125] In the formula, Y t represents the leading vehicle state variable, C t represents the regional environmental variable. f θ represents the leading vehicle behavior prediction network, that is, the behavior prediction network of the forward target vehicle of the driverless bus. θ represents the model network parameters of the leading vehicle behavior prediction network, and θ can be adjusted by continuous self - learning. f θ (Y t , C t ) represents the behavior of the leading vehicle. represents the predicted acceleration of the leading vehicle, that is, the acceleration of the forward target vehicle at the next moment. N represents the model random noise, which is used to add uncertainty to the model to avoid overfitting.

[0126] Exemplarily, the calculation formula of the change amount of the ego - vehicle control is as follows:

[0127]

[0128] In the formula, X t represents the ego - vehicle state variable, Indicates the predicted acceleration of the leading vehicle, C t Indicates the regional environmental variable. g μ Indicates the ego-vehicle control response model network, μ represents the model network parameters of the ego-vehicle control response model network, and μ is adjusted by continuous self-learning. ΔE self Indicates the ego-vehicle control change amount, which is used to control the ego-vehicle lateral angle change amount to obtain a local planning path.

[0129] Step S33: Determine the local path planning information according to the predicted acceleration of the leading vehicle and the ego-vehicle control change amount.

[0130] It should be noted that after obtaining the predicted acceleration of the leading vehicle and the ego-vehicle control change amount, the local path planning information can be calculated based on the predicted acceleration of the leading vehicle and the ego-vehicle control change amount. In the process of local path planning for the driverless bus, the control decision for the driverless bus in the current environment is generated according to the ego-vehicle control change amount calculated by the ego-vehicle control response network, which is used to control the ego-vehicle lateral angle change amount to obtain the local path planning information.

[0131] It should be understood that in the process of local path planning for the driverless bus, a causal contrast loss function is constructed according to the predicted acceleration of the leading vehicle and the ego-vehicle control change amount obtained at the current moment, and according to the causal loss contrast function, the network parameters of the leading vehicle behavior prediction equation and the ego-vehicle causal response equation are continuously adjusted to converge, so as to optimize the prediction and control performance of the network.

[0132] In a feasible implementation manner, step S33 may include steps S331 to S334:

[0133] Step S331: Calculate the leading vehicle acceleration comparison term based on the predicted acceleration of the leading vehicle through a time decay weight function and a safety interval weighted norm;

[0134] It should be noted that the time decay weight function is a function used to calculate the leading vehicle acceleration comparison term, which can reduce the confidence of future predictions as the number of prediction steps increases. Since the prediction uncertainty at a farther future time is higher, the time decay weight function gives a higher weight to the acceleration prediction at a closer time when predicting the acceleration of the leading vehicle, and a lower weight to the prediction at a farther time. The introduction of the time decay weight function can improve the accuracy and reliability of the leading vehicle acceleration prediction, making the local path planning more reasonable.

[0135] Exemplarily, the definition formula of the time decay weight function τ(k) is as follows:

[0136] τ(k) = e -β·k , β > 0

[0137] In the formula, β represents a hyperparameter adjusted according to experience or experiment, which controls the attenuation rate, k represents the number of prediction steps, starting from 0 and increasing. τ(k) represents the confidence attenuation of the prediction at the k-th future step.

[0138] In addition, it should be noted that when the safety distance weighted norm is used to measure the magnitude of the predicted error of the acceleration of the vehicle ahead, combined with the safety distance factor, it ensures that the predicted error is within a safe range, avoiding the collision risk caused by inaccurate prediction. It can be represented by ‖·‖ safe which represents the safety distance weighted norm.

[0139] It should be understood that by using the time decay weight function to weight the predicted value of the acceleration of the vehicle ahead and combining the safety distance weighted norm, the comparison term of the acceleration of the vehicle ahead can be calculated. Specifically, through the safety distance weighted norm, the predicted acceleration of the vehicle ahead at the next moment obtained at the current moment and the predicted acceleration of the vehicle ahead at the current moment obtained at the previous moment are differentiated to evaluate whether the safe following distance is satisfied, and the time decay weight function τ(k) is used to gradually reduce the dependence on the predicted acceleration of the vehicle ahead in the future (the influence of more distant predictions is smaller), so as to evaluate whether the safe following distance is satisfied, ensuring that the causal inference large model obtained based on the prediction network of the vehicle ahead behavior and the self-vehicle control response network pays more attention to the accuracy of recent predictions, and at the same time increasing the error penalty when the safety distance is insufficient, avoiding the risk of rear-end collision caused by the sudden deceleration of the vehicle ahead.

[0140] Step S332, calculate the self-vehicle steering response comparison term based on the self-vehicle control change amount through the causal response delay function and the curvature adaptive weighted norm;

[0141] It should be noted that the causal response delay function is a function that reflects the causal delay characteristics of human driving and is used to calculate the self-vehicle steering response comparison term. In actual driving, there is a certain delay in the driver's response to the behavior of the vehicle ahead. The causal response delay function can simulate this delay characteristic, making the self-vehicle control response of the driverless bus closer to human driving behavior and improving the naturalness and safety of driving.

[0142] Exemplarily, the causal response delay function is defined by the following formula:

[0143]

[0144] In the formula, Υ represents a parameter fitted according to the characteristics of human driving behavior or experimental data, and k represents the number of response delay steps, starting from 0 and increasing.

[0145] Additionally, it should be noted that the curvature adaptive weighted norm is used to adaptively adjust the weight according to the road curvature when calculating the error of the ego-vehicle lateral control variation, strengthen the steering angle matching accuracy in high-curvature areas such as curves, ensure driving stability, and can be represented by ‖·‖ curv to represent the curvature adaptive weighted norm.

[0146] It should be understood that by using the causal response delay function to weight the ego-vehicle control variation and combining the curvature adaptive weighted norm, the leading vehicle acceleration comparison term can be calculated. Specifically, through the curvature adaptive weighted norm, the ego-vehicle control variation at the next moment obtained at the current moment and the ego-vehicle control variation at the current moment obtained at the previous moment can be calculated. The rationality of the ego-vehicle control variation is evaluated, and the causal response delay function is combined with the road curvature adaptive rule (higher weight for the steering error in curves) to evaluate the rationality of the ego-vehicle control variation, ensuring that the causal inference large model based on the leading vehicle behavior prediction network and the ego-vehicle control response network can make the response of the driverless bus closer to human driving habits, and at the same time, giving priority to ensuring the steering accuracy in complex road conditions such as curves, avoiding deviating from the lane or losing control. Specifically, when driving on a curve, the curvature is large, and higher requirements are placed on the steering angle matching accuracy of the ego-vehicle. At this time, the curvature adaptive weighted norm will increase accordingly to strengthen the proportion of the steering angle matching accuracy in the loss function, so as to ensure that the driverless bus can better track the planned path when driving on a curve and improve driving stability.

[0147] Step S333, obtain a causal loss comparison function according to the comparison term weight coefficient, the leading vehicle acceleration comparison term, and the ego-vehicle steering response comparison term;

[0148] It should be noted that the causal loss comparison function is a function comprehensively obtained according to the comparison term weight coefficient, the leading vehicle acceleration comparison term, and the ego-vehicle steering response comparison term, and is used to compare and evaluate the rationality of the leading vehicle behavior prediction and the ego-vehicle control response. By weighted summing the leading vehicle acceleration comparison term and the ego-vehicle steering response comparison term, a comprehensive loss value is obtained to optimize the leading vehicle behavior prediction network and the ego-vehicle control response network.

[0149] Additionally, it should be noted that the comparison term weight coefficient includes the leading vehicle acceleration comparison term weight coefficient λ a and the ego-vehicle steering response comparison term weight coefficient λ E . Among them, the causal term weight coefficient λ a is used to control the relative importance of the leading vehicle prediction error in the total loss of the causal loss comparison function, and the ego-vehicle steering response comparison term weight coefficient λ ETo adjust the influence of the steering angle matching accuracy on the total loss of the causal loss contrast function.

[0150] Exemplarily, the causal loss contrast function L causal The calculation formula is as follows:

[0151]

[0152] In the formula, the front term of the loss function is the comparison term of the acceleration of the vehicle in front, and the rear term of the loss function is the comparison term of the steering response of the host vehicle. The meanings of the symbols in the formula have been described above and will not be elaborated here.

[0153] Step S334, optimize the front vehicle behavior prediction network and the host vehicle control response network based on the causal loss contrast function to obtain local path planning information.

[0154] It should be understood that when the loss value of the causal loss contrast function is small, it indicates that the prediction of the front vehicle behavior and the control response of the host vehicle better conform to the causal relationship and the actual driving requirements; when the loss value of the causal loss contrast function is large, the network parameters need to be adjusted to improve the prediction and response effects, and finally accurate local path planning information is obtained. Using the obtained causal loss contrast function, the parameters of the front vehicle behavior prediction network and the host vehicle control response network are iteratively updated through the backpropagation algorithm, continuously adjusting the network parameters to minimize the value of the causal loss contrast function, which can optimize the prediction and control performance of the causal inference large model obtained based on the front vehicle behavior prediction network and the host vehicle control response network, so that when the distance between the host vehicle and the vehicle in front is insufficient, the penalty for the acceleration prediction error is enhanced, and the steering angle matching accuracy can be strengthened for the curved road (high curvature) area, and the optimal local path planning information at the current moment is obtained.

[0155] In this embodiment, by determining the state variables of the vehicle in front, the state variables of the host vehicle, and the regional environment variables according to the environmental perception information, determining the predicted acceleration of the vehicle in front and the control change amount of the host vehicle according to the state variables of the vehicle in front, the state variables of the host vehicle, and the regional environment variables, and determining the local path planning information according to the predicted acceleration of the vehicle in front and the control change amount of the host vehicle, the environmental traffic conditions can be accurately evaluated, traffic conflicts can be reduced, and it can be ensured that the behavior of the host vehicle remains optimal in a dynamically changing environment.

[0156] Exemplarily, to help understand the implementation process of the unmanned bus path planning method obtained by combining this embodiment with the above Embodiment 1, please refer to Figure 3 , Figure 3 A brief flow schematic diagram of an unmanned bus path planning method is provided. Specifically:

[0157] During the global planning process, information about pedestrians, target vehicles, and driverless buses can be sensed through roadside sensors, and this information is input into an ethical large model constructed based on rule compliance weights, personnel injury weights, and regional movement style weights to generate a global path planning that meets ethical and social expectations. During the local planning process, the state of the vehicle in front and the state of the self-vehicle can be sensed through the on-vehicle sensors of the driverless bus, and this information is input into a causal large model constructed based on the prediction network of the vehicle in front's behavior and the control response network of the self-vehicle, and a local path planning is generated in real time through dynamic causal reasoning. Combining the results of global and local planning can obtain the final driving trajectory of the driverless bus, ensuring that the driverless bus can drive safely and stably on the basis of complying with rules, ensuring safety, and adapting to the regional style.

[0158] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the path planning method of the driverless bus in this application. Based on this technical concept, more forms of simple transformation are within the protection scope of this application.

[0159] This application also provides a driverless bus path planning device. Please refer to Figure 4 , and the driverless bus path planning device includes:

[0160] A data processing module 10, configured to construct a multi-objective model and determine the rule compliance weight, personnel injury weight, and regional movement style weight of the multi-objective model according to the roadside perception information;

[0161] A global planning module 20, configured to determine global path planning information according to the rule compliance weight, the personnel injury weight, and the regional movement style weight;

[0162] A local planning module 30, configured to determine local path planning information according to the environmental perception information;

[0163] A path planning module 40, configured to complete the path planning of the driverless bus according to the global path planning information and the local path planning information.

[0164] In an embodiment, the data processing module 10 is further configured to determine road segment type information, road segment static obstacle information, and road segment dynamic obstacle information according to the roadside perception information; and determine the rule compliance weight, personnel injury weight, and regional movement style weight of the multi-objective model according to the road segment type information, the static obstacle information, and the dynamic obstacle information.

[0165] In one embodiment, the data processing module 10 is further configured to evaluate the importance of rule compliance according to the road segment type information to obtain the rule compliance weight of the multi-objective model; evaluate the potential injury risk according to the static obstacle information and the dynamic obstacle information to obtain the personnel injury weight of the multi-objective model; evaluate the traffic order style according to the dynamic obstacle information to obtain the regional movement style weight of the multi-objective model.

[0166] In one embodiment, the local planning module 30 is further configured to determine the leading vehicle state variables, the host vehicle state variables, and the regional environment variables according to the environmental perception information; determine the predicted acceleration of the leading vehicle and the control change amount of the host vehicle according to the leading vehicle state variables, the host vehicle state variables, and the regional environment variables; determine the local path planning information according to the predicted acceleration of the leading vehicle and the control change amount of the host vehicle.

[0167] In one embodiment, the local planning module 30 is further configured to obtain the longitudinal speed of the leading vehicle, the lateral offset of the leading vehicle, the relative distance of the leading vehicle, the longitudinal speed of the host vehicle, the controlled lateral angle of the host vehicle, the longitudinal acceleration of the host vehicle, and the weather information according to the environmental perception information; determine the leading vehicle state variables according to the longitudinal speed of the leading vehicle, the lateral offset of the leading vehicle, and the relative distance of the leading vehicle; determine the host vehicle state variables according to the longitudinal speed of the host vehicle, the controlled lateral angle of the host vehicle, and the longitudinal acceleration of the host vehicle; determine the regional environment variables according to the weather information.

[0168] In one embodiment, the local planning module 30 is further configured to predict the behavior of the leading vehicle according to the leading vehicle state variables and the regional environment variables through the leading vehicle behavior prediction network; add model random noise to the behavior of the leading vehicle to obtain the predicted acceleration of the leading vehicle; determine the control change amount of the host vehicle according to the host vehicle state variables, the predicted acceleration of the leading vehicle, and the regional environment variables through the host vehicle control response network.

[0169] In one embodiment, the local planning module 30 is further configured to calculate the leading vehicle acceleration comparison term based on the predicted acceleration of the leading vehicle through the time decay weight function and the safety interval weighted norm; calculate the host vehicle steering response comparison term based on the control change amount of the host vehicle through the causal response delay function and the curvature adaptive weighted norm; obtain the causal loss comparison function according to the comparison term weight coefficient, the leading vehicle acceleration comparison term, and the host vehicle steering response comparison term; optimize the leading vehicle behavior prediction network and the host vehicle control response network based on the causal loss comparison function to obtain the local path planning information.

[0170] The driverless bus route planning device provided by the present application adopts the driverless bus route planning method in the above embodiment, and can solve the technical problem of how to improve the human-likeness of route planning while maintaining the safety of driverless bus route planning in a complex environment. Compared with the prior art, the beneficial effects of the driverless bus route planning device provided by the present application are the same as those of the driverless bus route planning method provided by the above embodiment, and other technical features in the driverless bus route planning device are the same as the features disclosed in the method of the above embodiment, which will not be elaborated here.

[0171] The present application provides a driverless bus route planning device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the driverless bus route planning method in the first embodiment above.

[0172] Next, refer to Figure 5 , which shows a schematic structural diagram of a driverless bus route planning device suitable for implementing the embodiments of the present application. The driverless bus route planning device in the embodiments of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions, tablet computers), PMPs (Portable Media Players), in-vehicle terminals (such as in-vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 5 The driverless bus route planning device shown is only an example, and should not impose any limitations on the functions and usage scope of the embodiments of the present application.

[0173] As Figure 5As shown, the unmanned bus path planning device may include a processing device 1001 (such as a central processing unit, a graphics processing unit, etc.), which can execute various appropriate actions and processes according to the program stored in the ROM (Read Only Memory) 1002 or the program loaded from the storage device 1003 into the RAM (Random Access Memory) 1004. In the RAM 1004, various programs and data required for the operation of the unmanned bus path planning device are also stored. The processing device 1001, the ROM 1002, and the RAM 1004 are connected to each other through a bus 1005. The input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems may be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, an LCD (Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the unmanned bus path planning device to communicate with other devices wirelessly or wiredly to exchange data. Although the figure shows an unmanned bus path planning device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or had alternatively.

[0174] Specifically, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program codes for executing the methods shown in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device, or installed from the storage device 1003, or installed from the ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the methods of the embodiments disclosed in the present application are executed.

[0175] The unmanned bus path planning device provided by the present application adopts the unmanned bus path planning method in the above-mentioned embodiment, and can solve the technical problem of how to improve the anthropomorphism of route planning while maintaining the safety of unmanned bus path planning in a complex environment. Compared with the prior art, the beneficial effects of the unmanned bus path planning device provided by the present application are the same as those of the unmanned bus path planning method provided by the above-mentioned embodiment, and other technical features in the unmanned bus path planning device are the same as those disclosed in the method of the previous embodiment, and will not be elaborated here.

[0176] It should be understood that each part disclosed in this application can be implemented by hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in a suitable manner in any one or more embodiments or examples.

[0177] As described above, the above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all of them should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0178] This application provides a computer-readable storage medium with computer-readable program instructions (i.e., computer programs) stored thereon, and the computer-readable program instructions are used to execute the unmanned bus path planning method in the above embodiments.

[0179] The computer-readable storage medium provided by this application can be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, RAM (Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable Read Only Memory), or flash memory, optical fibers, CD-ROM (CD-Read Only Memory), optical storage devices, magnetic storage devices, or any suitable combination of the above. In this embodiment, the computer-readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or combined with an instruction execution system, device, or device. The program code contained on the computer-readable storage medium can be transmitted by any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0180] The above computer-readable storage medium can be included in the unmanned bus path planning device; it can also exist separately without being assembled into the unmanned bus path planning device.

[0181] The above computer-readable storage medium carries one or more programs, which, when executed by the unmanned bus path planning device, cause the unmanned bus path planning device to: build a multi-objective model and determine the rule compliance weight, personnel injury weight, and regional movement style weight of the multi-objective model according to the roadside perception information; determine the global path planning information according to the rule compliance weight, the personnel injury weight, and the regional movement style weight; determine the local path planning information according to the environmental perception information; and complete the unmanned bus path planning according to the global path planning information and the local path planning information.

[0182] Computer program code for performing the operations of this application can be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a LAN (Local Area Network) or a WAN (Wide Area Network), or can be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0183] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment, or a part of the code that contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0184] The modules involved in the embodiments of the present application can be implemented in software or in hardware. In some cases, the name of the module does not constitute a limitation on the unit itself.

[0185] The readable storage medium provided by the present application is a computer-readable storage medium. The computer-readable storage medium stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned unmanned bus path planning method, and can solve the technical problem of how to improve the anthropomorphism of route planning while maintaining the safety of unmanned bus path planning in a complex environment. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as those of the unmanned bus path planning method provided by the above embodiments, and will not be elaborated here.

[0186] The present application also provides a computer program product, including a computer program, and the steps of the above-mentioned unmanned bus path planning method are implemented when the computer program is executed by a processor.

[0187] The computer program product provided by the present application can solve the technical problem of how to improve the anthropomorphism of route planning while maintaining the safety of unmanned bus path planning in a complex environment. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as those of the unmanned bus path planning method provided by the above embodiments, and will not be elaborated here.

[0188] The above are only some embodiments of the present application, and do not limit the patent scope of the present application. Any equivalent structural transformation made by using the content of the specification and drawings of the present application under the technical concept of the present application, or any direct / indirect application in other related technical fields, is included in the patent protection scope of the present application.

Claims

1. An unmanned bus path planning method, characterized in that, The unmanned bus path planning method includes: Construct a multi-objective model, and determine the rule compliance weight, personnel injury weight, and regional movement style weight of the multi-objective model according to the roadside perception information; Determine the global path planning information according to the rule compliance weight, the personnel injury weight, and the regional movement style weight; Determine the local path planning information according to the environmental perception information; Complete the unmanned bus path planning according to the global path planning information and the local path planning information.

2. The method according to claim 1, characterized in that, The step of determining the rule compliance weight, personnel injury weight, and regional movement style weight of the multi-objective model according to the roadside perception information includes: Determine the road section type information, road section static obstacle information, and road section dynamic obstacle information according to the roadside perception information; Determine the rule compliance weight, personnel injury weight, and regional movement style weight of the multi-objective model according to the road section type information, the static obstacle information, and the dynamic obstacle information.

3. The method according to claim 2, wherein The step of determining the rule compliance weight, personnel injury weight, and regional movement style weight of the multi-objective model according to the road section type information, the static obstacle information, and the dynamic obstacle information includes: Evaluate the importance of rule compliance according to the road section type information to obtain the rule compliance weight of the multi-objective model; Evaluate the potential injury risk according to the static obstacle information and the dynamic obstacle information to obtain the personnel injury weight of the multi-objective model; Evaluate the traffic order style according to the dynamic obstacle information to obtain the regional movement style weight of the multi-objective model.

4. The method according to claim 1, wherein The step of determining the local path planning information according to the environmental perception information includes: Determine the leading vehicle state variable, the ego vehicle state variable, and the regional environmental variable according to the environmental perception information; Determine the predicted acceleration of the leading vehicle and the control change amount of the ego vehicle according to the leading vehicle state variable, the ego vehicle state variable, and the regional environmental variable; Determine the local path planning information according to the predicted acceleration of the leading vehicle and the control change amount of the ego vehicle.

5. The method according to claim 4, wherein The step of determining the leading vehicle state variable, the ego vehicle state variable, and the regional environmental variable according to the environmental perception information includes: Obtain the longitudinal speed of the leading vehicle, the lateral offset of the leading vehicle, the relative distance of the leading vehicle, the longitudinal speed of the ego vehicle, the control lateral angle of the ego vehicle, the longitudinal acceleration of the ego vehicle, and the weather information according to the environmental perception information; Determine the leading vehicle state variable according to the longitudinal speed of the leading vehicle, the lateral offset of the leading vehicle, and the relative distance of the leading vehicle; Determine the ego vehicle state variable according to the longitudinal speed of the ego vehicle, the control lateral angle of the ego vehicle, and the longitudinal acceleration of the ego vehicle; Determine the regional environmental variable according to the weather information.

6. The method according to claim 4, wherein The step of determining the predicted acceleration of the leading vehicle and the control change amount of the ego vehicle according to the leading vehicle state variable, the ego vehicle state variable, and the regional environmental variable includes: Predict the behavior of the leading vehicle through the leading vehicle behavior prediction network according to the leading vehicle state variable and the regional environmental variable; Add model random noise to the behavior of the leading vehicle to obtain the predicted acceleration of the leading vehicle; Determine the vehicle control variation of the host vehicle based on the host vehicle state variables, the predicted acceleration of the leading vehicle, and the regional environmental variables through the host vehicle control response network.

7. The method according to claim 4, wherein The step of determining the local path planning information based on the predicted acceleration of the leading vehicle and the vehicle control variation of the host vehicle includes: Calculate the leading vehicle acceleration comparison term based on the predicted acceleration of the leading vehicle through a time decay weight function and a safety interval weighted norm; Calculate the host vehicle steering response comparison term based on the vehicle control variation of the host vehicle through a causal response delay function and a curvature adaptive weighted norm; Obtain a causal loss comparison function according to the comparison term weight coefficient, the leading vehicle acceleration comparison term, and the host vehicle steering response comparison term; Optimize the leading vehicle behavior prediction network and the host vehicle control response network based on the causal loss comparison function to obtain local path planning information.

8. A path planning device for an unmanned bus, characterized in that: The device includes: A data processing module, configured to construct a multi-objective model and determine the rule compliance weight, the personnel injury weight, and the regional movement style weight of the multi-objective model according to the roadside perception information; A global planning module, configured to determine the global path planning information according to the rule compliance weight, the personnel injury weight, and the regional movement style weight; A local planning module, configured to determine the local path planning information according to the environmental perception information; A path planning module, configured to complete the path planning of the driverless bus according to the global path planning information and the local path planning information.

9. An unmanned bus path planning device, characterized in that, The device includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the computer program is configured to implement the steps of the driverless bus path planning method according to any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium, and when the computer program is executed by a processor, it implements the steps of the driverless bus path planning method according to any one of claims 1 to 7.

Images