CN114889603A

CN114889603A - Vehicle lane changing processing method and device

Info

Publication number: CN114889603A
Application number: CN202210414722.6A
Authority: CN
Inventors: 徐鑫
Original assignee: Beijing Jingdong Qianshi Technology Co Ltd
Current assignee: Beijing Jingdong Qianshi Technology Co Ltd
Priority date: 2022-04-20
Filing date: 2022-04-20
Publication date: 2022-08-12

Abstract

The invention discloses a vehicle lane change processing method and device, and relates to the technical field of automatic driving. One embodiment of the method comprises: receiving current driving environment parameters transmitted by a target vehicle to determine information of other vehicles around the target vehicle; generating a plurality of traffic orders by selecting vehicles and corresponding vehicle information from other vehicles with a target vehicle as a starting point; screening a target passing sequence from the multiple passing sequences, and planning a collision-free track for each vehicle in sequence according to the priority sequence of the vehicles in the target passing sequence so as to prevent the target vehicle and other vehicles from colliding during simultaneous running; and triggering the target vehicle to carry out lane changing operation according to the conflict-free track planned for the target vehicle. According to the method, the Monte Carlo tree searching method and the heuristic rule are adopted, the target passing sequence can be rapidly screened out, corresponding tracks are intensively matched from the preset tracks according to the driving parameters, the generation efficiency of the lane changing decision is improved, and the lane changing safety is ensured.

Description

Vehicle lane changing processing method and device

Technical Field

The invention relates to the technical field of automatic driving, in particular to a vehicle lane change processing method and device.

Background

The lane changing strategy is a key technology for guaranteeing the driving safety of the automatic driving vehicle, and when the automatic driving vehicle runs in a complex dynamic environment, the lane changing strategy for making the vehicle safe and meeting the requirements is still a difficult point. Although many path-changing decision methods are proposed in the prior art, most of the path-changing decision methods belong to optimal planning strategies, so that a solution algorithm is designed or solution is directly carried out by means of commercial software, the solution time is usually long, and even feasible solutions cannot be obtained due to nonlinear constraints.

As an optimization method of the above embodiment, a feasible planning strategy is proposed to determine a passing order according to an artificially designed heuristic rule, and then the trajectories of all vehicles are optimized according to the obtained passing order. However, the traffic sequence given by the heuristic rules mostly follows the first-in first-out principle, which is difficult to generate better traffic performance.

Disclosure of Invention

In view of this, embodiments of the present invention provide a method and an apparatus for processing lane change of a vehicle, which are used for processing conflicts at different levels and simultaneously considering traffic coordination performance and calculation time, so as to at least solve the problem in the prior art that it is difficult to determine a lane change policy of an autonomous vehicle.

To achieve the above object, according to an aspect of an embodiment of the present invention, there is provided a lane change processing method for a vehicle, including:

receiving current driving environment parameters transmitted by a target vehicle to determine information of other vehicles around the target vehicle;

generating a plurality of passing orders from the other vehicles by selecting vehicles and corresponding vehicle information with the target vehicle as a starting point;

screening out a target passing sequence from the multiple passing sequences, and planning a collision-free track for each vehicle in sequence according to the priority sequence of the vehicles in the target passing sequence so as to prevent the target vehicle and the other vehicles from colliding during simultaneous running;

and triggering the target vehicle to carry out lane changing operation according to the conflict-free track planned for the target vehicle.

Optionally, before the determining information of other vehicles located around the target vehicle, the method further includes:

if the road resources in front of the current driving lane of the target vehicle are occupied or the road geometry changes in the current driving environment parameters, triggering lane change logic; or

And if the current driving environment parameters comprise a lane change request transmitted by the target vehicle, triggering lane change logic.

Optionally, the method further includes:

when current driving environment parameters transmitted by a plurality of vehicles are received, preferentially distributing the right of way to the vehicles which are in the key conflict area in front of the driving; the key conflict area is an area where road resources are occupied or the road geometry changes; and

and when the driving front of each vehicle is not a critical conflict area, preferentially distributing the right of way to the vehicle with the driving position at the forefront in the plurality of vehicles.

Optionally, the determining information of other vehicles located around the target vehicle further includes:

and constructing a local conflict area based on the current areas of the target vehicle and the other vehicles, and determining the feasible action of each other vehicle in the local conflict area.

Optionally, the generating a plurality of passing orders from the other vehicles by selecting vehicles and corresponding vehicle information with the target vehicle as a starting point includes:

a plurality of passage sequences are generated by randomly selecting vehicles and each vehicle's possible actions among the other vehicles with the target vehicle as a starting point.

Optionally, the screening out a target passage order from the plurality of passage orders includes:

calling a heuristic rule, and screening one or more passing sequences from the multiple passing sequences according to a first-in first-out principle according to the current positions of the target vehicle and each other vehicle;

calling a Monte Carlo tree search program, determining that nodes without expansion action exist in each screened passing sequence, and creating new nodes as child nodes of the nodes; the child node is used for indicating driving environment parameters of the target vehicle after the non-expansion action is executed;

sampling from vehicles which are not included in each passing sequence by taking the child nodes as starting points, adding the sampled vehicles into each passing sequence, and repeating the random sampling operation until leaf nodes are generated to obtain a simulation score;

in each passing sequence, all nodes on the node and the path from the root node to the node are graded and accumulated according to the simulation grade; the root node is used for indicating the current driving environment parameters of the target vehicle;

and calculating a confidence value according to the simulation times and the accumulated scores of each node, determining a target node with the maximum confidence value, and taking the passing sequence between the target node and the root node as a target passing sequence.

Optionally, before planning a collision-free trajectory for each vehicle in sequence according to the priority order of the vehicles in the target passing order, the method further includes:

calculating a safe vehicle distance between the target vehicle and a vehicle in front of the lane change target lane without collision according to the driving parameters of the target vehicle and the driving parameters of the vehicle in front of the lane change target lane;

in response to the distance between the target vehicle and the vehicle ahead of the lane change target lane being smaller than the safe distance, adjusting the speed of the target vehicle so that the distance between the target vehicle and the vehicle ahead of the lane change target lane is larger than or equal to the safe distance;

and repeating the operation of adjusting the speed to meet the collision constraint condition, so that the distance between the target vehicle and each other vehicle and the corresponding front vehicle meets the collision constraint condition.

Optionally, the planning a collision-free trajectory for each vehicle in turn includes:

searching out a conflict-free track matched with the driving parameters from a preset track set according to the driving parameters of each vehicle; the driving parameters comprise speed, acceleration and vehicle type;

and when the number of the matched conflict-free tracks is multiple, taking the conflict-free track with the highest road right priority in the multiple conflict-free tracks as a target conflict-free track.

To achieve the above object, according to another aspect of an embodiment of the present invention, there is provided a lane change processing apparatus for a vehicle, including:

the determining module is used for receiving the current driving environment parameters transmitted by the target vehicle so as to determine the information of other vehicles around the target vehicle;

a generating module, configured to generate a plurality of passing orders from the other vehicles by selecting a vehicle and corresponding vehicle information from the target vehicle as a starting point;

the processing module is used for screening out a target passing sequence from the multiple passing sequences, and planning a collision-free track for each vehicle in sequence according to the priority sequence of the vehicles in the target passing sequence so that the target vehicle and the other vehicles do not collide when running simultaneously;

and the lane changing module is used for triggering the target vehicle to carry out lane changing operation according to the conflict-free track planned for the target vehicle.

Optionally, the apparatus further includes a lane change determining module, configured to:

And if the current driving environment parameters comprise the lane changing request transmitted by the target vehicle, triggering lane changing logic.

Optionally, the apparatus further includes a lane change priority determining module, configured to:

Optionally, the determining module is further configured to:

Optionally, the generating module is configured to: a plurality of passage sequences are generated by randomly selecting vehicles and each vehicle's possible actions among the other vehicles with the target vehicle as a starting point.

Optionally, the processing module is configured to:

calling a Monte Carlo tree search program, determining that nodes without expansion action exist in each screened passing sequence, and creating new nodes as child nodes of the nodes; the child node is used for indicating driving environment parameters of the target vehicle after the unextended action is executed;

Optionally, the processing module is further configured to:

Optionally, the processing module is configured to:

To achieve the above object, according to still another aspect of embodiments of the present invention, there is provided a lane change decision electronic device for a vehicle.

The electronic device of the embodiment of the invention comprises: one or more processors; and the storage device is used for storing one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors realize any one of the vehicle lane changing processing methods.

To achieve the above object, according to still another aspect of embodiments of the present invention, there is provided a computer-readable medium having a computer program stored thereon, the program, when executed by a processor, implementing any one of the vehicle lane change processing methods described above.

According to the scheme provided by the invention, one embodiment of the invention has the following advantages or beneficial effects: fully considering road complexity and dynamics, carrying out lane change necessity evaluation on a target vehicle, and preparing lane change in advance; the feasible actions of other vehicles around the target vehicle are fully considered, a plurality of passing sequences are generated, and the target passing sequences are accurately determined by relying on a Monte Carlo tree searching method and heuristic rules; under the condition that all vehicles meet the collision constraint condition, a collision-free path is planned again, and the safety of the target vehicle in the lane changing process is guaranteed.

Further effects of the above-mentioned non-conventional alternatives will be described below in connection with the embodiments.

Drawings

The drawings are included to provide a better understanding of the invention and are not to be construed as unduly limiting the invention. Wherein:

FIG. 1 is a schematic flow chart of a vehicle lane-change processing method according to an embodiment of the present invention;

FIG. 2(a) is a schematic view of a typical traffic scenario with a lane change;

FIG. 2(b) is a schematic view of another traffic typical scene with lane change;

FIG. 2(c) is a schematic diagram of finding multiple collision-free tracks that match from a set of tracks;

FIG. 3 is a schematic flow diagram of an alternative lane change process for a vehicle according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of a search iteration using a Monte Carlo tree search method;

FIG. 5 is a schematic flow diagram of an alternative lane-change processing method for a vehicle according to an embodiment of the present invention;

FIG. 6 is a schematic diagram of adjusting vehicle speed to meet crash constraints;

FIG. 7 is a schematic diagram of the major modules of a lane-change processing apparatus for a vehicle according to an embodiment of the present invention;

FIG. 8 is an exemplary system architecture diagram in which embodiments of the present invention may be employed;

FIG. 9 is a schematic block diagram of a computer system suitable for use with a mobile device or server implementing an embodiment of the invention.

Detailed Description

Exemplary embodiments of the present invention are described below with reference to the accompanying drawings, in which various details of embodiments of the invention are included to assist understanding, and which are to be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the invention. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

It should be noted that the embodiments and features of the embodiments may be combined with each other without conflict. According to the technical scheme, the data acquisition, storage, use, processing and the like meet relevant regulations of national laws and regulations.

The words and phrases involved in the present scheme are explained herein as follows:

heuristic rules are as follows: heuristic, i.e. a combination of simplified virtual machines and simplified behavior decision engines. Based on heuristic rules, the method mainly aims at updating the characteristic value identification technology and solving the defect of single characteristic code comparison. The aim is not to detect all unknown viruses, but only to supplement the eigenvalue scanning technique.

Monte Carlo Tree Search (MCTS), a Tree structure based Monte Carlo method, is performed throughout 2 according to the Monte Carlo method ^N (N equals the number of decisions, i.e. tree depth) space to find the optimal tree structure path (feasible solution) based on certain feedback. In summary, MCTS is a heuristic random search algorithm driven by deterministic rules.

And (3) passing sequence: the priority of each vehicle is indicated, and the vehicles with higher rank have the right to occupy the road resource preferentially.

Referring to fig. 1, a main flowchart of a lane change processing method for a vehicle according to an embodiment of the present invention is shown, including the following steps:

s101: receiving current driving environment parameters transmitted by a target vehicle to determine information of other vehicles around the target vehicle;

s102: generating a plurality of passing orders from the other vehicles by selecting vehicles and corresponding vehicle information with the target vehicle as a starting point;

s103: screening a target passing sequence from the multiple passing sequences, and planning a collision-free track for each vehicle in sequence according to the priority sequence of the vehicles in the target passing sequence so as to prevent the target vehicle and the other vehicles from colliding during simultaneous running;

s104: and triggering the target vehicle to carry out lane changing operation according to the conflict-free track planned for the target vehicle.

In the above embodiment, for step S101, the intelligent driving research is currently the leading direction of the development of the automobile field, and as the number of vehicles on the road increases in the future, the traffic congestion problem and the traffic accident will be more prominent, and these problems are caused by the wrong lane change in many cases, so for the automatically driven vehicle, a reasonable lane change strategy is selected, which has an important research value for ensuring safe driving.

In the driving process of each vehicle, real-time driving environment parameters can be obtained by using a sensor, the surrounding environment is sensed, and a traffic scene is identified. The scenes requiring the lane change of the vehicle can be divided into two types of lane change by free lane change and forced lane change: 1) the forced lane change is usually generated by traffic rules and road conditions, and the lane change and obstacle avoidance behaviors which have to be executed by vehicles according to the traffic rules are carried out. For example, when the sensor detects a forward accident, the lane change is forced. 2) The free lane change is that the vehicle can independently select lane keeping or lane changing in the driving process.

According to the scheme, the conflict area is set and divided into a key conflict area and a local conflict area according to the influence of the conflict area on the traffic scene. When a vehicle is about to arrive at a key conflict area, a lane needs to be changed forcibly, the key conflict area is a relatively fixed area, is usually set according to road characteristics and is a bottleneck area of traffic flow, so that the traffic flow is hindered, and common key conflict areas such as city intersections, expressway ramps, working areas and the like need to plan the driving track of the vehicle carefully when the vehicle runs in the areas so as to avoid collision. There are two main types of reasons for critical conflict areas:

1. when road resources are occupied, such as a vehicle accident and a working area shown in fig. 2(a), a part of the area on the road is forbidden, so that the number of lanes is reduced, traffic flow in a certain range is affected, and even the traffic flow of a plurality of roads is blocked. Therefore, all vehicles in the same lane as the accident vehicle or the working area must be changed in advance to avoid collision, and as shown in fig. 2(a), the vehicle B is in an impassable working area in front of the current driving lane, so that the vehicle B needs to be changed to the right in advance.

2. Changes in road geometry, such as road intersections and road merges, result in a dramatic increase in the risk of collisions and a decrease in traffic capacity. Referring to fig. 2(b), the key conflict area is an intersection, the traffic flow and the people flow in the area are large, and a left-turning vehicle G needs to change lanes to a left lane due to traffic regulations.

A local collision zone, as opposed to a critical collision zone, refers to an area where collisions may occur due to vehicle motion or behavior and generally affects the motion of several nearby vehicles only for a short period of time. As shown in fig. 2(a), the vehicle B needs to change lane to the right, and the right adjacent lane is located around the vehicle B and has already driven the vehicle C, so that the current areas of the vehicle B and the vehicle C form a local collision area, or the current areas of the vehicle B, the vehicle C and the vehicle a form a local collision area, and the local collision area disappears after the lane change of the vehicle B is finished; the vehicle E needs to change lane and overtake, the right lane is located around the vehicle E, and the vehicle D already runs, so that the current areas of the vehicle E and the vehicle D form a local conflict area, or the current areas of the vehicle C, the vehicle D and the vehicle E form a local conflict area, and the area disappears after the lane changing and overtaking of the vehicle E are finished. As shown in fig. 2(b), the vehicle G needs to change lanes to the left, and the left lane is located around the vehicle G and the vehicle F has already traveled, so that the vehicle F and the current area of the vehicle G form a local collision zone, and the local collision zone disappears after the lane change of the vehicle G is finished. Namely, the local conflict areas are generated by the action of one vehicle and disappear rapidly along with the ending of the action.

As can be seen from the above description, the critical conflict area has a significant influence on the traffic efficiency of the whole traffic scene, and therefore, in actual operation, it is necessary to determine which vehicles are about to reach the critical conflict area, preferentially allocate right of way to the vehicles about to reach the critical conflict area (i.e., have a right of preferentially occupying road resources), and then feed back the allocation result to the upstream to optimize the movement of the vehicle in the local conflict area.

Furthermore, the central controller on the roadside device is usually configured with a higher-performance computing device than the vehicle-mounted device, so that all vehicles in the scheme are directly communicated with the central controller, real-time driving environment parameters acquired by respective sensors of the vehicles are sent to the central controller, and then the central controller performs a lane change decision for each vehicle through the high-performance computing device configured by the central controller or a cloud computing access mode.

It should be noted that, although the real-time driving environment parameters of each vehicle are processed, a problem of balance between the flexibility of the computing system and the computing capability is involved, in practical application, the real-time driving environment parameters can be realized in a form of a self-organizing network, specifically, the vehicles broadcast state and intention information thereof, all vehicles close to a collision area form a network in a self-organizing form, then, a lead vehicle is dynamically selected to plan collision-free tracks for all vehicles, and the planned tracks are respectively returned to the vehicles. However, considering that the central controller processing mode can achieve higher-speed calculation and faster landing, and ensures that cooperative decision and control of the vehicle group are achieved in real time, the scheme preferably adopts the central controller processing mode according to the requirements of the scene and evaluation important factors.

The central controller receives current driving environment parameters transmitted by a target vehicle, wherein the current driving environment parameters comprise a current driving lane, driving parameters and surrounding scene information; the peripheral scene information is as follows: the lane changing method comprises the following steps of determining the lane changing target lane according to the lane changing target lane, the lane distance between the target vehicle and the vehicle in front of the current lane, the lane distance between the target vehicle and the vehicle in front of the lane changing target lane, the lane distance between the target vehicle and the vehicle behind the lane changing target lane, the vehicle distance between the target vehicle and the vehicle behind the lane changing target lane, the use condition of a steering lamp of the vehicle behind the current lane at the current moment, and driving parameters such as the speed of the target vehicle at the current moment, the speed of the vehicle in front of the current lane, the speed of the vehicle in front of the lane changing target lane, the speed of the vehicle behind the lane changing target lane, the type of the vehicle and the like.

The central controller may first determine whether road resources in front of the current driving lane are occupied and whether the front road geometry changes according to the position of the "current driving lane", and if at least one of the determination results is yes, determine that the front of the target vehicle is a critical collision area, and trigger a forced lane change logic, such as vehicle B in fig. 2(a) and vehicle G in fig. 2 (B). The lane change logic, such as vehicle E in fig. 2(a), may also be triggered based on the "active lane change request" in the current driving environment parameters.

The central controller may receive current driving environment parameters sent by a plurality of vehicles at the same time, but when the plurality of vehicles need to change lanes and the driving front is not a critical conflict area, the vehicles with the driving positions at the forefront among the vehicles need to be considered, and lane change logic is triggered on the vehicles preferentially. As shown in fig. 2(a), it is assumed that none of the vehicles a to E ahead is a critical collision area, and since the vehicle a is the most advanced in the traveling position among the vehicles, the lane change logic is triggered to the vehicle a to assign the right of way preferentially. However, if only the vehicle E and the vehicle D send the current driving environment parameters, the lane change logic is preferentially triggered for the vehicle D because the vehicle D is located forward relative to the vehicle E.

For step S102, the present solution also optimizes the trajectories of all vehicles according to the passing order, and in order to eliminate the infeasible passing order, the following assumptions are made: 1) due to road geometric constraint or traffic rules, the vehicles which need to change lanes only have one possible action (such as lane changing), the vehicles which generate negative influence during lane changing only have one possible action (such as straight running), and the rest vehicles can freely select straight running or lane changing; 2) for other vehicles on the same lane, the front vehicle should have a higher priority for traffic than all the rear vehicles. Besides the basic assumptions, other reasonable assumptions can be customized according to the characteristics of the actual application scene, and the scheme is not limited herein.

There may be two possible actions for a vehicle, and thus a plurality of passing sequences are generated by randomly selecting the vehicles and the possible actions of each vehicle among other vehicles around the target vehicle, with the target vehicle as a starting point. The traffic order represents the priority of each vehicle, with higher ranked vehicles having the right to preferentially occupy road resources. The traffic sequence is usually expressed in the form of a string, for example CAB means that the vehicle C has the highest priority, and when there is a conflict with other vehicles AB, it needs to be slowed down to give way. And when there is no conflict between vehicle a and vehicle C, the performance of the traffic sequences CAB and ACB is the same.

Taking fig. 2(a) as an example, the vehicle B needs to execute lane change logic, and according to the current driving environment parameters, it is determined that the front vehicle in the right lane change target lane is the vehicle a, the rear vehicle is the vehicle C, and the current driving lane of the vehicle B has no front vehicle and no rear vehicle, so that a local collision area can be constructed by the areas where the vehicle a, the vehicle B, and the vehicle C are located. Or the lane change of the vehicle B only has negative influence on the vehicle C, and a local conflict area can be constructed only by the areas where the vehicle B and the vehicle C are located. Considering only the feasible actions of each vehicle in the local conflict area, such as the vehicle A going straight or changing lane, the vehicle B only changing lane, and the vehicle C only going straight, so as to generate a plurality of traffic sequences ABC, ACB, BAC, BCA, CAB, CBA, AB, AC, BA, BC, CA, CB, and since the front of the vehicle B is the key conflict area, the right of way is preferentially distributed to the vehicle B, and four traffic sequences BAC, BCA, BA and BC are screened out.

For step S103, most of the complex traffic scenes are composed of basic scene elements, which include intersections, upper and lower ramps, rotary islands, road segments, and the like. Referring to fig. 2(a) and 2(b), most basic traffic scenarios include a downstream critical conflict zone and multiple local conflict zones. Therefore, an urban road network can be segmented into two typical traffic scenes, namely an intersection scene and a road section scene according to the geometric characteristics of roads, the motion of vehicles can be considered as driving from one traffic scene to the other traffic scene, and the lane change control in each traffic scene can be operated independently.

Therefore, if all typical traffic scenarios can be handled well, control and coordination of more complex traffic scenarios can be achieved through an expansion and combination manner. The scheme provides a group decision based on a double-layer planning framework, the upper layer of the method aims to find the optimal target passing sequence to maximize traffic efficiency, and the lower layer aims to solve the traffic track conflict of each vehicle in a local conflict area according to the target passing sequence.

The problem of finding the optimal target traffic sequence is substantially a tree search problem, in order to accelerate the search process, the upper layer of the group decision based on the double-layer planning frame preferably adopts a mode of combining a Monte Carlo tree search method and heuristic rules, screens out one or more traffic rules which accord with the rule from a plurality of traffic rules through the heuristic rules according to a first-in first-out principle, ensures that the obtained traffic sequence is basically reasonable, and then uses the Monte Carlo tree search method to continuously screen from the one or more traffic rules.

Compared with the existing method of traversing all feasible traffic sequences by adopting an optimal planning strategy, the Monte Carlo tree searching method is more prone to searching the traffic sequences (namely the target traffic sequences) with the potential to become the optimal traffic sequences, and therefore the overall searching efficiency is greatly improved. In the construction process of the search tree, the current optimal target passing sequence is continuously updated, and when the maximum calculation time budget is reached, the search process is ended and the currently screened target passing sequence is returned.

Furthermore, because the structured road limits vehicle movement, the right-of-way assignments can sometimes only be adjusted within a limited space. The traffic sequence is preferably screened through heuristic rules, and then a better solution can be obtained by continuously adjusting the road rights of some vehicles in the sequence, and the properties also ensure that the decision method can find at least one feasible solution for any scene.

Based on the lower layer of the group decision of the double-layer planning framework, an interpretation algorithm from the passing sequence to the track is designed, the interpretation algorithm can rapidly and accurately plan a collision-free track for each vehicle according to the priority sequence of the vehicles in the target passing sequence, and the vehicles with higher priorities can be planned with the track in priority. And the passing sequence given by the upper layer plan is evaluated according to the evaluation result. If the lower layer can not find the corresponding track-changing track for the target passing sequence, a value is fed back to the upper layer to indicate that the passing sequence is not feasible. After the upper layer receives the feedback value, the Monte Carlo tree search method adjusts the search direction to avoid generating similar traffic sequence again.

By iteratively solving the upper and lower layer problems, the current optimal transit order is continuously updated. In actual operation, the interpretation algorithm is complex to operate, so that the optimal solution is difficult to find in a certain time. Considering that the suboptimal solution usually does not reduce the effect obviously, and meanwhile, in order to ensure the real-time performance of the whole solving process, the optimal approximate solution is preferably obtained in a short time instead of obtaining the global optimal solution in a large amount of time. And when the allowable maximum calculation time limit is reached, the searching process is ended, and the searched current optimal target passing sequence and the vehicle collision-free track corresponding to the current optimal target passing sequence are returned.

For step S104, it is difficult for the existing optimization problem to consider multiple feasible lane change tracks, so it is generally assumed that the lane change tracks of all vehicles are the same, but only one feasible lane change track often results in no solution to the problem. In order to solve the problem, a track set is constructed in advance by the scheme, the track set comprises a large number of different lane changing tracks, and each track is generated according to constraint conditions such as vehicle type, initial speed, acceleration, final speed and the like.

Each vehicle can search a plurality of matched collision-free tracks from the track set based on the driving parameters thereof, such as the above-mentioned constraint conditions of vehicle type, initial speed, acceleration, etc., as shown in fig. 2(c), and select one collision-free track with the highest priority of road right from the plurality of collision-free tracks, so that the target vehicle and other vehicles around the target do not collide during the simultaneous driving.

And for the target vehicle, after a conflict-free track with the highest road right priority is screened out, updating the road space-time occupation information according to the conflict-free track changing track of the target vehicle, so that other subsequent vehicles can conveniently make a track changing decision based on the updated road space-time occupation information.

The method provided by the embodiment adopts a mode of combining the Monte Carlo tree searching method and the heuristic rule, thereby quickening the efficiency of screening the target passing sequence and reducing the calculated amount; corresponding tracks can be matched from the preset tracks in a centralized mode according to vehicle driving parameters, the generation efficiency of lane change decisions is improved, and the method has important significance for guaranteeing road traffic safety and improving road traffic capacity.

Referring to fig. 3, a schematic flow chart of an alternative lane-changing processing method for a vehicle according to an embodiment of the present invention is shown, including the following steps:

s301: calling a Monte Carlo tree search program, determining that nodes without expansion action exist in each screened passing sequence, and creating new nodes as child nodes of the nodes; the child node is used for indicating driving environment parameters of the target vehicle after the unextended action is executed;

s302: sampling from vehicles which are not included in each passing sequence by taking the child nodes as starting points, adding the sampled vehicles into each passing sequence, and repeating the random sampling operation until leaf nodes are generated to obtain a simulation score;

s303: in each passing sequence, all nodes on the node and the path from the root node to the node are graded and accumulated according to the simulation grade; the root node is used for indicating the current driving environment parameters of the target vehicle;

s304: and calculating a confidence value according to the simulation times and the accumulated scores of each node, determining a target node with the maximum confidence value, and taking the passing sequence between the target node and the root node as a target passing sequence.

In the above embodiment, in steps S301 to S303, as shown in fig. 2(a), since the road segment scene includes the vehicle a, the vehicle B, the vehicle C, the vehicle D, and the vehicle E, the set { a, B, C, D, E } is formed, and the total number of the sets is 5! 120, when there are a large number of vehicles in the local conflict area that need to consider lane change behavior, the number of passing orders to be calculated will be huge, and at this time, all leaf nodes in one search tree cannot be expanded within a limited calculation time, so this scheme preferably uses the monte carlo tree search method MCTS to search out the optimal target passing order from the generated multiple passing orders.

The MCTS method constructs a search tree in an iterative manner, where each iteration generally includes four steps: selection (Selection), Expansion (Expansion), Simulation (Simulation), Back Propagation (Back Propagation). Each node in the search tree contains three basic information: driving environment parameters, simulation times and accumulated scores.

1. A selection stage: there may be several possibilities for each extended node:

1) if all the feasible actions of the node are already expanded, calculating the UCT values of all the nodes of the node, finding out a child node with the maximum UCT value, continuously checking, and repeatedly iterating downwards.

2) If a node has 20 feasible actions but 19 child nodes are created in the search tree, the node is the node of the current iteration, and the action of the node which is not expanded is determined.

2. And (3) an expansion stage: at the end of the selection phase, a node Y which is most urgently expanded and an action X which is not expanded are found. A new node Yn is created in the search tree as a new child node of Y. The situation of Yn is the situation of the node Y after executing the action X, and for this scheme, the situation is the driving environment parameter of the target vehicle.

3. A simulation stage: to get an initial score for Yn, a random sampling is taken from vehicles that have not been included in the traffic sequence, starting with Yn, until an outcome is obtained that will be the initial score for Yn (i.e., the simulated score). Wins or failures are typically used as scores, with only 1 or 0.

4. And (3) a back propagation stage: after the simulation of Yn is finished, the parent node Y and all nodes on the path from the root node to the node Y add their cumulative scores according to the result of the current simulation. The search tree is expanded every iteration, and the scale of the search tree is continuously increased along with the increase of the number of iterations. And when a certain iteration number or time is up, selecting the best child node under the root node as the decision result.

For step S304, the MCTS method continuously simulates each routine until the end, and the winning rate of the routine is calculated to be W/N according to the total simulation times N of the routine and the successful simulation results W. In order to balance the maximum odds against the new node exploration, a UCT (Upper Confidence Bound algorithm) is introduced.

The confidence interval is the confidence level of the probability calculation result. The UCT can be used to correct the problem of too few samples, such as 3 coins being thrown with the face up, and the probability of the coin being thrown with the face up is calculated to be 100%, due to the error caused by too few samples. The formula used by the UCT is:

wherein, w _i The number of times the simulation result of the node i is successful, i.e. the cumulative score, n _i Is the number of simulations of node i, N _i Is the simulation times of all nodes, c is an exploration constant and has a theoretical value of

Can be adjusted empirically.

Therefore, after the iteration is completed, the node Y in each passing sequence, all nodes on the path from the root node to the node Y, and the cumulative score and the simulation times acquired in the last iteration are input into the formula to obtain the UCT value of each node. And taking the node with the maximum UCT value as a target node, and taking the passing sequence between the target node and the root node as the optimal target passing sequence.

Referring to fig. 4, in the initial stage, the search tree has only one root node R, which represents the current driving environment parameters of the target vehicle, and the root node R is used as the starting node of each passing sequence. Referring to fig. 2(a), assuming that the current local collision area includes a vehicle a, a vehicle B, and a vehicle C, according to a mode of randomly selecting feasible actions of the vehicle and each vehicle, and the front of the vehicle B is a critical collision area, road right allocation is preferentially performed on the vehicle B, four passing sequences of BAC, BCA, BA, and BC are screened out, and because the vehicle a and the vehicle C are located in the same lane, two passing sequences are finally obtained: BA. And BC.

In addition to changing lanes to the front of vehicle a, vehicle B may also change lanes to the front of vehicle C, and node BC may also be expanded if only one traffic sequence BA is currently available. Continuing to randomly generate a direct child node, assumed to be the BCA, at the lower layer thereof using a random sampling method starting from the node BC. If there are more vehicles, node BCA repeats the above process until a leaf node, e.g., node BCADE, is generated. After the leaf node is generated, a simulation result, i.e. simulation success or simulation failure, is obtained.

The classical MCTS method randomly samples vehicles that are not yet included in the traffic sequence and then continuously adds the sampled vehicles to the existing traffic sequence until the complete traffic sequence is formed, without creating branches in the process, but all the way down until the maximum depth of the tree is reached. In the construction process of the search tree, the current optimal passing sequence is continuously updated, and the search process is ended when the upper limit of the iteration times or the maximum calculation time budget is reached (namely the iteration time reaches the preset time). Based on the optimal target traffic sequence, the speed and acceleration curves of all vehicles can be calculated by a traffic sequence-to-trajectory interpretation algorithm in the lower layer plan.

In the method provided by the embodiment, the Monte Carlo tree search method is used to convert the search problem for determining the optimal target traffic sequence into the tree search problem constructed by all feasible traffic sequences, and the nodes with the highest confidence values are selected by simulating a plurality of vehicles and feasible actions thereof, so that the accuracy of determining the target traffic sequence is improved.

Referring to fig. 5, a schematic flow chart of another alternative lane-changing processing method for a vehicle according to an embodiment of the present invention is shown, which includes the following steps:

s501: calculating a safe vehicle distance between the target vehicle and a vehicle in front of the lane change target lane without collision according to the driving parameters of the target vehicle and the driving parameters of the vehicle in front of the lane change target lane;

s502: in response to the fact that the distance between the target vehicle and the vehicle in front of the lane changing target lane is smaller than the safe distance, adjusting the speed of the target vehicle to enable the distance between the target vehicle and the vehicle in front of the lane changing target lane to be larger than or equal to the safe distance;

s503: and repeating the operation of adjusting the speed to meet the collision constraint condition, so that the distance between the target vehicle and each other vehicle and the corresponding front vehicle meets the collision constraint condition.

In the above embodiment, for steps S501 to S503, after the target passing order is screened out based on the monte carlo tree search method, a collision-free trajectory can be planned for each vehicle, so that the decision variables are: the target passing sequence, the expected arrival time, the speed and acceleration curves of each vehicle, the vehicle distance, and the speed, the acceleration and the vehicle distance can be located in the driving parameters of the vehicles.

According to the scheme, an interpretation algorithm from a passing sequence to a track is designed, as shown in fig. 6, a target vehicle adjusts the distance between the target vehicle and a front vehicle of a lane change target lane through acceleration and deceleration actions so as to ensure that a sufficient space exists for realizing lane change, and the method belongs to a distance-based method for judging collision risks. In addition, there are many other types of methods that may be used to evaluate the Risk of Collision during a lane change, such as a time to-Collision (TTC) Based method and a Risk Index (Risk Index-Based) Based method, which may be incorporated into the present scheme. The scheme can also use collision avoidance constraint conditions, which are defined as follows:

d _ij (t)≥F _ij (t)

wherein v is _i (t) is the speed of vehicle i, v _j (t) is the speed of the vehicle j, d _ij (t) is the vehicle distance between vehicle i and vehicle j, F _ij (t) is a safe vehicle distance, which represents the length of a travel-prohibited area between the vehicle i and the vehicle j, wherein the travel-prohibited area refers to a space-time area where the front vehicle has absolute road right, and the rear vehicle cannot enter the travel-prohibited area owned by the front vehicle; ρ is the safe headway, a _i,brake (t) is the average braking acceleration of the vehicle i, a _i,max,brake Is the maximum braking acceleration of the vehicle i, a _i,min,brake Is the minimum braking acceleration of the vehicle i, and the subscript m in the formula represents the maximum or minimum one, which can be arbitrarily selected. In addition, due to the parameter a _i,max,brake 、a _i,min,brake Dependent on vehicle type, thus safe vehicle distance F _ij (t) is also dependent on the vehicle type, i.e. the solution is applicable to different types of vehicle composition scenarios.

It should be noted that the front vehicle includes a front vehicle of a lane where the vehicle is currently located and a front vehicle of a lane change target lane, as shown in fig. 6, the longitudinal speed of the vehicle which does not satisfy the collision avoidance constraint condition is adjusted until the safe lane change condition is satisfied, and the vehicle I needs to change the lane to the right lane, so that the front vehicle is the vehicle J, and the vehicle distance between the vehicle I and the vehicle J needs to satisfy d _ij (t)≥F _ij (t) of (d). And the vehicle I is positioned in front of the vehicle K after changing lanes, so that the front vehicle of the vehicle K is the vehicle I, and the distance between the vehicle K and the vehicle I needs to meet d _ki (t)≥F _ki (t) of (d). In the lane changing lane, the vehicle J is arranged in front of the vehicle K, so the distance between the two vehicles needs to satisfy d _kj (t)≥F _kj (t) but due to d _ij (t)≥F _ij (t)、d _ki (t)≥F _ki (t), the distance between the vehicle K and the vehicle J inevitably satisfies the collision constraint condition, and therefore, it is preferable to reduce the amount of calculation without considering the distance between the vehicle K and the vehicle J, and similarly, the default distance satisfies the constraint collision constraint condition without considering the distance between the vehicle K and the vehicle J and the vehicle I located in the lane where the vehicle I is located and in front of the vehicle I.

All unplanned vehicles track their predecessors according to a following algorithm, such as the Newell following algorithm, a following algorithm based on a particle-spring-damper-clutch model, etc., which typically acts on the first vehicle in the road segment, such as vehicle J in fig. 6, to control the unplanned vehicle to maintain a safe and appropriate distance from the predecessor.

In the actual driving process of the vehicle, relative distances between the target vehicle and a vehicle ahead of the current lane, between the target vehicle and a vehicle ahead of the lane change target lane, and between the target vehicle and a vehicle behind the lane change target lane have larger influence on the lane change behavior decision relative to the influence of relative speeds of the two vehicles. If the relative distance between the target vehicle and the vehicle in front of the current lane and the vehicle in front of the lane change target lane is too large, the target vehicle needs to adopt an acceleration strategy to adjust the speed of the vehicle to avoid the too large vehicle distance, and if the relative distance between the target vehicle and the vehicle behind the lane change target lane is too small, the target vehicle needs to adopt a deceleration strategy to adjust the vehicle distance. And when the vehicle distances between the target vehicle and the vehicle in front of the current lane, between the target vehicle and the vehicle in front of the lane change target lane and between the target vehicle and the vehicle behind the lane change target lane all meet the collision constraint condition, adopting a constant speed strategy form strategy by the target vehicle.

The method provided by the embodiment can effectively identify the condition that the collision constraint condition is not met between the vehicle and the surrounding environment, and the vehicle distance between the vehicles is adjusted through acceleration and deceleration operation to construct a safe lane changing scene, so that the safety of the lane changing process is ensured, and the method has strong application significance in actual road driving.

The scheme aims at the decision-making problem of the road section scene group considering the road change, the proposed solution analyzes the necessity and the safety of the autonomous road change or the forced road change of the automatic driving vehicle, and effectively reduces the solving complexity caused by the track change track:

1. judging whether the road resource in front of the vehicle is occupied or not and whether the road geometry is changed or not, and determining whether a mandatory lane change logic is executed or not according to the judgment result; or executing lane change logic according to a lane change request transmitted by the vehicle;

2. the group decision idea based on double-layer planning is provided, the upper layer of the idea rapidly screens out a target passing sequence from a plurality of passing sequences by adopting a mode of combining a Monte Carlo tree searching method and heuristic rules, and the solving efficiency is improved; the lower layer of the thought emphasizes processing local conflicts caused by the actions of lane changing and the like, an interpretation algorithm from the passing sequence to the track is designed, the running speed of each vehicle is adjusted to enable the vehicle distance of the vehicles to meet the collision constraint condition, the conflict-free tracks of all vehicles can be conveniently and rapidly planned, and the efficient cooperation of vehicle groups is realized.

Referring to fig. 7, a schematic diagram of main modules of a vehicle lane change processing device 700 provided by an embodiment of the present invention is shown, including:

the determining module 701 is used for receiving the current driving environment parameters transmitted by the target vehicle so as to determine information of other vehicles around the target vehicle; optionally, the method may also be used to determine whether a lane change operation needs to be performed, and if the road resource in front of the current driving lane of the target vehicle in the current driving environment parameter is occupied or the road geometry changes, trigger a lane change logic; or if the current driving environment parameters comprise the lane change request transmitted by the target vehicle, triggering lane change logic.

A generating module 702, configured to generate a plurality of passing orders from the other vehicles by selecting a vehicle and corresponding vehicle information from the target vehicle as a starting point; specifically, a plurality of transit sequences are generated by randomly selecting vehicles and the feasible actions of each vehicle.

The processing module 703 is configured to screen out a target passage order from the plurality of passage orders, and plan a collision-free trajectory for each vehicle in sequence according to a vehicle priority order in the target passage order, so that the target vehicle and the other vehicles do not collide when traveling simultaneously; specifically, the method comprises the following steps:

1) the implementation steps of screening out the target passing sequence from the plurality of passing sequences are as follows: calling a heuristic rule, and screening one or more passing sequences from the multiple passing sequences according to a first-in first-out principle according to the current positions of the target vehicle and each other vehicle;

in each passing sequence, all nodes on the node and the path from the root node to the node are graded and accumulated according to the simulation scores; the root node is used for indicating the current driving environment parameters of the target vehicle;

2) For planning a collision-free trajectory, then: calculating a safe vehicle distance between the target vehicle and a vehicle in front of the lane change target lane without collision according to the driving parameters of the target vehicle and the driving parameters of the vehicle in front of the lane change target lane;

And the lane changing module 704 is configured to trigger the target vehicle to perform a lane changing operation according to a conflict-free trajectory planned for the target vehicle.

The device also comprises a channel-changing priority determining module used for:

In the device for implementing the present invention, the determining module is further configured to:

In addition, the detailed implementation of the device in the embodiment of the present invention has been described in detail in the above method, so that the repeated description is not repeated here.

Fig. 8 shows an exemplary system architecture 800 in which embodiments of the invention may be applied, including terminal devices 801, 802, 803, a network 804 and a server 805 (by way of example only).

The terminal devices 801, 802, 803 may be various electronic devices having display screens and supporting web browsing, and are installed with various communication client applications, and users may interact with the server 805 through the network 804 using the terminal devices 801, 802, 803 to receive or transmit messages, and the like.

The network 804 serves to provide a medium for communication links between the terminal devices 801, 802, 803 and the server 805. Network 804 may include various types of connections, such as wire, wireless communication links, or fiber optic cables, to name a few.

The server 805 may be a server providing various services, and it should be noted that the method provided by the embodiment of the present invention is generally executed by the server 805, and accordingly, the apparatus is generally disposed in the server 805.

It should be understood that the number of terminal devices, networks, and servers in fig. 8 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation.

Referring now to FIG. 9, shown is a block diagram of a computer system 900 suitable for use with a terminal device implementing an embodiment of the present invention. The terminal device shown in fig. 9 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present invention.

As shown in fig. 9, the computer system 900 includes a Central Processing Unit (CPU)901 that can perform various appropriate actions and processes in accordance with a program stored in a Read Only Memory (ROM)902 or a program loaded from a storage section 908 into a Random Access Memory (RAM) 903. In the RAM 903, various programs and data necessary for the operation of the system 900 are also stored. The CPU 901, ROM 902, and RAM 903 are connected to each other via a bus 904. An input/output (I/O) interface 905 is also connected to bus 904.

The following components are connected to the I/O interface 905: an input portion 906 including a keyboard, a mouse, and the like; an output section 907 including components such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker; a storage portion 908 including a hard disk and the like; and a communication section 909 including a network interface card such as a LAN card, a modem, or the like. The communication section 909 performs communication processing via a network such as the internet. A drive 910 is also connected to the I/O interface 905 as needed. A removable medium 911 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 910 as necessary, so that a computer program read out therefrom is mounted into the storage section 908 as necessary.

In particular, according to the embodiments of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication section 909, and/or installed from the removable medium 911. The above-described functions defined in the system of the present invention are executed when the computer program is executed by a Central Processing Unit (CPU) 901.

It should be noted that the computer readable medium shown in the present invention can be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present invention, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present invention, however, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing.

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The modules described in the embodiments of the present invention may be implemented by software or hardware. The described modules may also be provided in a processor, which may be described as: a processor comprises a determining module, a generating module, a processing module and a channel changing module. Where the names of these modules do not in some cases constitute a limitation of the module itself, for example, a generating module may also be described as a "traffic order generating module".

As another aspect, the present invention also provides a computer-readable medium that may be contained in the apparatus described in the above embodiments; or may be separate and not incorporated into the device. The computer readable medium carries one or more programs which, when executed by a device, cause the device to perform any of the vehicle lane change processing methods described above.

The computer program product of the present invention includes a computer program that, when executed by a processor, implements the vehicle lane change processing method in the embodiment of the present invention.

The above-described embodiments should not be construed as limiting the scope of the invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions can occur, depending on design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A lane change processing method for a vehicle, comprising:

2. The method of claim 1, further comprising, prior to said determining information of other vehicles located around the target vehicle:

3. The method of claim 2, further comprising:

when current driving environment parameters transmitted by a plurality of vehicles are received, preferentially performing road right distribution on the vehicles in a key conflict area in front of driving; the key conflict area is an area where road resources are occupied or the road geometry changes; and

when the driving front of each vehicle is not a critical conflict area, the road right distribution is preferentially carried out on the vehicle with the most front driving position in the plurality of vehicles.

4. The method of any of claims 1-3, wherein the determining information of other vehicles located around the target vehicle further comprises:

5. The method of claim 1, wherein said screening a target passage order from said plurality of passage orders comprises:

6. The method of claim 1, further comprising, prior to said planning a collision-free trajectory for each vehicle in turn in a vehicle priority order of the target transit order:

7. The method of claim 1 or 6, wherein the planning of collision-free trajectories for each vehicle in turn comprises:

8. A vehicle lane change processing device characterized by comprising:

9. An electronic device, comprising:

one or more processors;

a storage device for storing one or more programs,

when executed by the one or more processors, cause the one or more processors to implement the method of any one of claims 1-7.

10. A computer-readable medium, on which a computer program is stored, which, when being executed by a processor, carries out the method according to any one of claims 1-7.