CN113928341B

CN113928341B - Road decision method, system, equipment and medium

Info

Publication number: CN113928341B
Application number: CN202111453193.2A
Authority: CN
Inventors: 李思思; 韩旭
Original assignee: Guangzhou Weride Technology Co Ltd
Current assignee: Guangzhou Weride Technology Co Ltd
Priority date: 2021-11-30
Filing date: 2021-11-30
Publication date: 2023-05-30
Anticipated expiration: 2041-11-30
Also published as: CN113928341A

Abstract

The application discloses a road decision method, a system, equipment and a medium, wherein a travelable road between the current position of an unmanned vehicle and a destination is divided into a plurality of waypoints; determining a target waypoint in the waypoints between the current position of the unmanned vehicle and the destination, and calculating the global cost from the target waypoint to the destination to obtain a waypoint value of the target waypoint; iteratively calculating the waypoint value of the waypoint between the target waypoint and the current position of the unmanned vehicle according to the waypoint value of the target waypoint; the method has the advantages that the action at the current position is determined according to the waypoint value of the next waypoint corresponding to the current position of the unmanned vehicle in real time, the current road decision result of the destination is obtained, the method of model predictive control is adopted in the prior art to obtain the optimal road decision, the optimal road decision is obtained by solving the complex optimization problem, a large amount of calculation capacity is needed to solve the nonlinear optimization problem, and the technical problem of low road decision efficiency is caused.

Description

Road decision method, system, equipment and medium

Technical Field

The application relates to the technical field of unmanned vehicles, in particular to a road decision method, a system, equipment and a medium.

Background

The unmanned vehicle is a comprehensive intelligent platform integrating multiple functions of environment perception and cognition, dynamic planning and decision, behavior control and execution and the like, and the core problems of unmanned vehicle research include environment perception, behavior decision and motion control.

The road decision is a main component of an unmanned vehicle decision technology, the prior art generally adopts a model predictive control method to obtain an optimal road decision, obtains the optimal road decision by solving a complex optimization problem, and needs a large amount of computing capacity to solve a nonlinear optimization problem, so that the road decision is low in efficiency and difficult to be effectively applied to a decision system of an unmanned vehicle.

Disclosure of Invention

The application provides a road decision method, a system, equipment and a medium, which are used for improving the technical problems that the prior art adopts a model predictive control method to obtain an optimal road decision, obtains the optimal road decision by solving a complex optimization problem, and needs a large amount of computing capacity to solve a nonlinear optimization problem, so that the road decision efficiency is low.

In view of this, a first aspect of the present application provides a road decision method, including:

dividing a travelable road between the current position of the unmanned vehicle and the destination into a plurality of waypoints, wherein the travelable road at least comprises a lane, and each lane comprises a plurality of waypoints which are connected in sequence;

Determining a target waypoint in waypoints between the current position of the unmanned vehicle and the destination, and calculating global cost from the target waypoint to the destination to obtain a waypoint value of the target waypoint;

iteratively calculating a waypoint value of the target waypoint to the waypoint between the current position of the unmanned vehicle according to the waypoint value of the target waypoint;

and determining the action at the current position according to the waypoint value of the next waypoint corresponding to the current position of the unmanned vehicle in real time, and obtaining the current road decision result of the destination.

Optionally, the calculating the global cost from the target waypoint to the destination, to obtain the waypoint value of the target waypoint includes:

obtaining the shortest path from the target waypoint to the destination through a graph searching algorithm;

calculating the travel time of the unmanned vehicle from the target waypoint to the destination based on the shortest path and a preset travel speed;

acquiring a global cost of the target waypoint to the destination based on a travel time of the unmanned vehicle from the target waypoint to the destination;

and taking the global cost from the target waypoint to the destination as the waypoint value of the target waypoint.

Optionally, the acquiring the global cost from the target waypoint to the destination based on the travel time of the unmanned vehicle from the target waypoint to the destination includes:

taking the travel time of the unmanned vehicle from the target waypoint to the destination as the global cost of the target waypoint to the destination;

or calculating the global cost from the target waypoint to the destination according to the target information from the target waypoint to the destination and the running time of the unmanned vehicle from the target waypoint to the destination.

Optionally, when iteratively calculating the waypoint value of the waypoint between the target waypoint and the current position of the unmanned vehicle according to the waypoint value of the target waypoint, the calculation process of the waypoint value of the last waypoint corresponding to the target waypoint is as follows:

calculating the short-term cost and the state transition probability of the last waypoint corresponding to the target waypoint;

and superposing the short-term cost of transferring the last waypoint corresponding to the target waypoint on the basis of the waypoint value of the target waypoint, and calculating the waypoint value of the last waypoint corresponding to the target waypoint by combining the state transition probability.

Optionally, a plurality of target waypoints are arranged between the current position of the unmanned vehicle and the destination.

Optionally, the method further comprises:

acquiring a special waypoint influenced by traffic information when the unmanned vehicle runs according to the current road decision result;

when the special waypoint is a waypoint which can be reached in the future when the special waypoint runs according to the current road decision result, determining the next waypoint of the special waypoint according to the current road decision result, and updating the short-term cost from the special waypoint to the next waypoint;

updating the waypoint value of the special waypoint based on the short-term cost from the special waypoint to the next waypoint after updating;

and reversely iterating and updating the waypoint value of each waypoint between the special waypoint and the current position of the unmanned vehicle according to the waypoint value updated by the special waypoint, and returning to the step of determining the action at the current position in real time according to the waypoint value of the next waypoint corresponding to the current position of the unmanned vehicle to obtain the current road decision result driven to the destination.

Optionally, when the traffic information includes static traffic participants;

when the unmanned vehicle runs according to the current road decision result, acquiring the special waypoint influenced by the traffic information, wherein the method comprises the following steps:

And when the unmanned vehicle runs according to the current road decision result, determining a special waypoint influenced by the static traffic participant according to the position of the static traffic participant.

Optionally, the method further comprises:

when the special waypoint influenced by the static traffic participant is a waypoint which can be reached in the future according to the current road decision result and the adjacent lane of the lane where the special waypoint is located cannot pass, dividing a lane separation line between the lane where the special waypoint influenced by the static traffic participant is located and the adjacent lane into a plurality of waypoints which are connected in sequence;

calculating the waypoint value of each waypoint on the lane dividing line according to the waypoint value of the waypoint on the adjacent lane of the lane dividing line, the short-term cost of the transition between the waypoints and the state transition probability, and returning to the step of determining the action at the current position in real time according to the waypoint value of the next waypoint corresponding to the current position of the unmanned vehicle to obtain the current road decision result of driving to the destination.

Optionally, when the traffic information includes dynamic traffic participants;

Determining a target dynamic traffic participant from the dynamic traffic participants when the unmanned vehicle runs according to the current road decision result;

and determining a special waypoint influenced by the target dynamic traffic participant according to the running speed of the target dynamic traffic participant and the running speed of the unmanned vehicle.

Optionally, the determining a target dynamic traffic participant from the dynamic traffic participants includes:

taking a dynamic traffic participant positioned in a preset range in front of the unmanned vehicle as a potential target dynamic traffic participant;

judging whether the running speed of the potential target dynamic traffic participant is smaller than the speed limit value of the lane where the potential target dynamic traffic participant is located, and obtaining a judging result;

calculating the confidence value of the potential target dynamic traffic participant according to the value of the potential target dynamic traffic participant and the judging result;

a target traffic participant is determined from the potential target dynamic traffic participants based on the confidence value.

Optionally, when the special waypoint is a waypoint that will reach in the future according to the current road decision result, determining a next waypoint of the special waypoint according to the current road decision result, and updating the short-term cost from the special waypoint to the next waypoint, including:

When the special waypoint is a waypoint which can be reached in the future according to the current road decision result, determining the next waypoint of the special waypoint according to the current road decision result, and determining the driving distance from the special waypoint to the next waypoint;

and calculating the short-term cost of the unmanned vehicle from the special waypoint to the next waypoint according to the driving distance and the driving speed of the target traffic participant, and obtaining the updated short-term cost from the special waypoint to the next waypoint.

Optionally, the state transition probability includes a lane change success rate, and the method further includes:

and updating the lane change success rate when transferring between waypoints in the preset range of the unmanned vehicle according to the traffic information.

Optionally, the updating the lane change success rate when transferring between waypoints in the preset range of the unmanned vehicle according to the traffic information includes:

acquiring the current distance between the rear side vehicle of the unmanned vehicle and the current yield probability of the rear side vehicle of the unmanned vehicle according to the traffic information, and updating the lane change success rate of the unmanned vehicle at the current waypoint;

updating the lane changing success rate of the remaining waypoints in the preset range of the unmanned vehicle according to the traffic density of the target lane changing lane, wherein the target lane changing lane is a lane after lane changing, and the remaining waypoints in the preset range of the unmanned vehicle are other waypoints except the current waypoints in the preset range of the unmanned vehicle.

Optionally, the calculation process of the current yield probability of the vehicle at the rear side of the unmanned vehicle is as follows:

and calculating the current yield probability of the rear side vehicle according to the current acceleration of the rear side vehicle of the unmanned vehicle and the yield probability of the rear side vehicle at the previous moment, wherein the initial yield probability of the rear side vehicle is obtained through initialization.

A second aspect of the present application provides a road decision system comprising:

the division module is used for dividing a travelable road between the current position of the unmanned vehicle and the destination into a plurality of waypoints, wherein the travelable road at least comprises a lane, and each lane comprises a plurality of waypoints which are connected in sequence;

the first calculation module is used for determining a target waypoint in the waypoints between the current position of the unmanned vehicle and the destination, calculating the global cost from the target waypoint to the destination and obtaining the waypoint value of the target waypoint;

a second calculation module for iteratively calculating a waypoint value of the target waypoint to a waypoint between the current location of the unmanned vehicle according to the waypoint value of the target waypoint;

and the decision module is used for determining the action at the current position according to the waypoint value of the next waypoint corresponding to the current position of the unmanned vehicle in real time to obtain the current road decision result of the destination.

A third aspect of the present application provides a road decision device, the device comprising a processor and a memory;

the memory is used for storing program codes and transmitting the program codes to the processor;

the processor is configured to execute the road decision method according to any one of the first aspects according to instructions in the program code.

A fourth aspect of the present application provides a computer readable storage medium for storing program code which when executed by a processor implements the road decision method of any one of the first aspects.

From the above technical scheme, the application has the following advantages:

the application provides a road decision method, which comprises the following steps: dividing a travelable road between the current position of the unmanned vehicle and the destination into a plurality of waypoints, wherein the travelable road at least comprises a lane, and each lane comprises a plurality of waypoints which are connected in sequence; determining a target waypoint in the waypoints between the current position of the unmanned vehicle and the destination, and calculating the global cost from the target waypoint to the destination to obtain a waypoint value of the target waypoint; iteratively calculating the waypoint value of the waypoint between the target waypoint and the current position of the unmanned vehicle according to the waypoint value of the target waypoint; determining the action at the current position according to the waypoint value of the next waypoint corresponding to the current position of the unmanned vehicle in real time, and obtaining the current road decision result of the destination, wherein the action is left lane change, lane keeping or right lane change.

In the method, a travelable road between an unmanned vehicle and a destination is divided into waypoints, the waypoint value of each waypoint is calculated, so that the unmanned vehicle can carry out road decision at each waypoint according to the waypoint value of the next waypoint, the complex road decision optimization problem is simplified, the waypoint value is calculated in two stages, the global cost from the target waypoint to the destination is calculated in the first stage, the waypoint value of the target waypoint is obtained, the waypoint value from the target waypoint is calculated in the second stage according to the waypoint value of the target waypoint in a reverse iteration manner, the calculation speed of the waypoint value is improved, the road decision efficiency is improved, the optimal road decision is obtained by adopting a model prediction control method in the prior art, the optimal road decision is obtained by solving the complex optimization problem, a large amount of calculation capacity is needed to solve the nonlinear optimization problem, and the technical problem of low road decision efficiency is caused.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive faculty for a person skilled in the art.

Fig. 1 is a schematic flow chart of a road decision method according to an embodiment of the present application;

FIG. 2 is a diagram of a distribution diagram of pivot points according to an embodiment of the present disclosure;

FIG. 3 is a plot of a waypoint distribution provided in an embodiment of the present application;

FIG. 4 is a graph of distribution of waypoint values of the waypoints of FIG. 3 calculated from static traffic information according to an embodiment of the present application;

FIG. 5 is a traffic scenario with static traffic participants provided in an embodiment of the present application;

FIG. 6 is a graph of updated waypoint values of the waypoint values of FIG. 5 according to an embodiment of the present application;

FIG. 7 is a traffic scenario with dynamic traffic participants provided in an embodiment of the present application;

FIG. 8 is a chart of a pre-update waypoint value profile provided by an embodiment of the present application in the presence of a dynamic traffic participant;

FIG. 9 is a plot of updated waypoint values of the waypoint values of FIG. 8 according to an embodiment of the present application;

fig. 10 is a plot of waypoint values in a special traffic scenario according to an embodiment of the present disclosure;

fig. 11 is a schematic structural diagram of a road decision system according to an embodiment of the present application.

Detailed Description

In order to make the present application solution better understood by those skilled in the art, the following description will clearly and completely describe the technical solution in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.

For ease of understanding, referring to fig. 1, an embodiment of the present application provides a road decision method, which includes:

step 101, dividing a travelable road between the current position of the unmanned vehicle and the destination into a plurality of waypoints, wherein the travelable road at least comprises a lane, and each lane comprises a plurality of waypoints which are connected in sequence.

After the destination is determined, dividing a travelable road between the current position of the unmanned vehicle and the destination into a plurality of waypoints, wherein the travelable road at least comprises one lane, each lane comprises a plurality of waypoints which are connected in sequence, and the waypoints of each lane are uniformly distributed. The unmanned vehicle makes road decisions at each waypoint, determines whether to make lane changes, and how to make lane changes.

And 102, determining a target waypoint in the waypoints between the current position of the unmanned vehicle and the destination, and calculating the global cost from the target waypoint to the destination to obtain the waypoint value of the target waypoint.

At least one target waypoint can be determined at a waypoint between the current position of the unmanned vehicle and the destination, when an intersection (including an intersection, a T-shaped intersection and the like) exists between the current position of the unmanned vehicle and the destination, the intersection of a drivable road between the current position of the unmanned vehicle and the destination can be divided into hub centers, and connection points between each road and each hub center are divided into hub points, wherein the connection points comprise an entrance point and an exit point of the hub center, and specifically, referring to fig. 2, a connection line between the hub points and the hub points can be a moving mode of the unmanned vehicle between the hub points, for example, the unmanned vehicle can reach another hub point from the current hub point through actions such as lane change, but the unmanned vehicle cannot perform lane change at the hub center, and the connection relationship between the hub points needs to consider a global map and traffic rules. After the hub center is obtained by dividing, a waypoint corresponding to an entry point of a certain hub center (may be the hub center closest to the unmanned vehicle, such as the hub center 1 in fig. 2) may be selected as the target waypoint, or waypoints corresponding to entry points of a plurality of hub centers may be selected as the target waypoints. It will be appreciated that other waypoints may be selected as the target waypoint, and are not specifically limited herein.

And after the target waypoint is determined, calculating the global cost from the target waypoint to the destination to obtain the waypoint value of the target waypoint. The specific calculation process of the waypoint value of the target waypoint can be as follows:

obtaining the shortest path from the target waypoint to the destination through a graph searching algorithm; calculating the travel time of the unmanned vehicle from the target waypoint to the destination based on the shortest path and the preset travel speed; acquiring global cost from a target waypoint to a destination based on travel time of the unmanned vehicle from the target waypoint to the destination; and taking the global cost from the target waypoint to the destination as the waypoint value of the target waypoint.

Specifically, a global map of the unmanned vehicle to the destination may be obtained, and then the global map may be analyzed by a graph search algorithm (e.g., an a-star algorithm) to obtain a shortest path from the target waypoint to the destination; then, calculating the running time of the unmanned vehicle from the target waypoint to the destination based on the shortest path from the target waypoint to the destination and a preset running speed, wherein the preset running speed can be the speed limit value of the lane; and finally, acquiring the global cost from the target waypoint to the destination based on the travel time of the unmanned vehicle from the target waypoint to the destination, and taking the global cost from the target waypoint to the destination as the waypoint value of the target waypoint.

It should be noted that, when the target waypoint is the entry point of the hub center, the global map may be converted into a search map composed of the hub points; when the target waypoint is not an entry point of the hub center, the global map can be converted into a search map composed of the waypoints, and then the search map is analyzed through a map search algorithm to obtain the shortest path from the target waypoint to the destination.

In one embodiment, the travel time of the unmanned vehicle from the target waypoint to the destination may be taken as the global cost of the target waypoint to the destination. When the target waypoints are the entry points of a certain hub center, the number of the target waypoints is multiple, and after the global cost from each target waypoint to the destination is calculated, the unmanned vehicle can determine which target waypoint enters the hub center through the global cost, so that the unmanned vehicle can reach the destination fastest.

In another embodiment, the global cost of the target waypoint to the destination may be calculated from the target information of the target waypoint to the destination and the travel time of the unmanned vehicle from the target waypoint to the destination. The global cost of the target waypoint to the destination may be determined by a number of factors, such as the distance between the target waypoint to the destination, the number of traffic lights between the target waypoint to the destination, whether there is a toll booth, etc. Therefore, the global cost can be calculated on the basis of the running time in consideration of traffic light quantity information or toll gate information, etc. Specifically, the driving time and the target information are comprehensively considered, and the target information in the embodiment of the application includes traffic light quantity information or toll station information, and the target information can also include other driving requirement related information. The running time and the target information allocation weight may be linearly combined to obtain the global cost, and the specific weight allocation situation may be set according to the actual situation, which is not limited herein.

It will be appreciated that the global cost of the target waypoint depends on the location of the user input destination, and that the global cost of the target waypoint is fixed when the user is not updating the destination.

And step 103, iteratively calculating the waypoint value of the waypoint between the target waypoint and the current position of the unmanned vehicle according to the waypoint value of the target waypoint.

In the embodiment of the application, a selection process of the waypoints is modeled as a Markov decision model, and the waypoints are states in which the unmanned vehicle can be located. The markov decision model may be expressed as < S, a, T, C >, S is the state space of the unmanned vehicle, a= { left transition lane, lane keeping, right transition lane } is the set of actions of the unmanned vehicle, C is a single step cost function for calculating the short term cost required for the unmanned vehicle to transition from one state to another, e.g., C (S, a, S ') is used for calculating the short term cost required for the unmanned vehicle to perform action a from waypoint S to waypoint S'; t is a transition model representing uncertainty caused by the action, e.g., T (s, right transition lane, s') represents a lane change success rate for an unmanned vehicle performing a right transition lane transition to a waypoint at the waypoint.

In the embodiment of the application, when iteratively calculating the waypoint value from the target waypoint to the waypoint between the current position of the unmanned vehicle according to the waypoint value of the target waypoint, only the static traffic information is considered, and the unmanned vehicle is assumed to run in a static and time-invariant traffic environment, wherein only the unmanned vehicle is arranged between the current position of the unmanned vehicle and the destination, and no other traffic participants exist. The calculation process of the waypoint value of the last waypoint corresponding to the target waypoint may be:

S1031, calculating short-term cost and state transition probability of transition from the last waypoint corresponding to the target waypoint.

After the waypoint value of the target waypoint is obtained through calculation, the waypoint value of the last waypoint corresponding to the target waypoint is calculated in a reverse iteration mode, a certain cost is needed to be paid when the unmanned vehicle is transferred between the waypoints, the short-term cost when the waypoints are transferred can be calculated through a single-step cost function C, and the state transfer probability when the waypoints are transferred is determined through a transfer model T. The driving time for executing the left transition road, keeping the lane or transferring the right transition road to the target waypoint at the last waypoint corresponding to the target waypoint can be calculated according to the distance between the last waypoint corresponding to the target waypoint and the speed limit value or the historical driving speed average value of the lane where the target waypoint is located. It should be noted that, in the case where there are a plurality of waypoints in the last waypoint corresponding to the target waypoint. Assuming that the target waypoint is positioned in the middle lane of the three lanes, at the moment, the last waypoint corresponding to the target waypoint comprises the last waypoint of the left lane, the last waypoint of the own lane and the last waypoint of the right lane; if the target waypoint is located on the left side lane of the three lanes, the last waypoint corresponding to the target waypoint comprises the last waypoint of the own lane and the last waypoint of the middle lane.

In one embodiment, the travel time of the last waypoint corresponding to the target waypoint may be directly used as the short-term cost of the last waypoint corresponding to the target waypoint.

In another embodiment, other losses may be considered based on the calculated travel time for the last waypoint corresponding to the target waypoint to transition to the target waypoint, for example, user preference settings that do not want to let the unmanned vehicle travel in the rightmost lane or enter the bus lane, etc., which may result in a certain loss, so the loss resulting from the user preference settings may be increased based on the travel time for the last waypoint corresponding to the target waypoint to transition to the target waypoint to obtain a short-term cost for the last waypoint corresponding to the target waypoint to transition to the target waypoint.

S1032, superposing the short-term cost of transferring the last waypoint corresponding to the target waypoint on the basis of the waypoint value of the target waypoint, and calculating the waypoint value of the last waypoint corresponding to the target waypoint by combining the state transfer probability.

And when the point value of the target point to the point between the current positions of the unmanned vehicles is calculated in a reverse iteration mode, the short-term cost of transferring the last point corresponding to the target point is superposed on the point value of the target point, and the point value of the last point corresponding to the target point is calculated by combining the state transfer probability. In the case that there are a plurality of previous waypoints of the target waypoint, the actions executed by the different waypoints to shift to the target waypoint are different, as shown in fig. 3, the previous waypoint of the target waypoint S0 on the lane (i.e., the left side lane) is S3, the previous waypoint of the adjacent lane (i.e., the middle lane) is S4, the previous waypoint of the target waypoint S1 on the lane (i.e., the middle lane) is S4, the previous waypoints of the adjacent lanes (i.e., the left side lane and the right side lane) are S3 and S5, the previous waypoint of the target waypoint S2 on the lane (i.e., the right side lane) is S5, and the previous waypoint of the adjacent lane (i.e., the middle lane) is S4.

Assuming that the destination waypoints S0, S1, S2 have waypoint values of 100, 50, 80, respectively, the single step cost function is set to C (S, a, S')=1, that is, the unmanned vehicle only needs to pay 1 unit of cost to transfer between the waypoints in the static traffic environment, the lane change success rate is set to 20%, that is, when the unmanned vehicle changes lanes at the waypoints, 20% of chance lane change success occurs.

When calculating the waypoint value of the last waypoint S4 corresponding to the target waypoint, the executable actions of the unmanned vehicle at the waypoint S4 include left transition, lane keeping, and right transition. When a lane-keeping transition to the target waypoint S1 is selected at the waypoint S4, the corresponding waypoint S4 has a waypoint value V (S4) _{Keeping lane} ＝(50+1)*100％＝51；

When the left transition lane is selected to be transferred to the target waypoint S0 at the waypoint S4, the waypoint value of the corresponding waypoint S4 is V (S4) because the lane change success rate is 20 percent _{Left transition lane} ＝(100+1)*20％+(50+1)*80％＝71；

When the right transition lane is selected at the waypoint S4 to be transferred to the waypoint target waypoint S2, the waypoint value of the corresponding waypoint S4 is V (S4) because the lane change success rate is 20 percent _{Right transition lane} ＝(80+1)*20％+(50+1)*80％＝57；

Final V (S4) =min (V (S4) _{Keeping lane} ，V(S4) _{Left transition lane} ，V(S4) _{Right transition lane} ) =51, and therefore, the terminal waypoint S4 has a waypoint value of 51.

The executable actions of the unmanned vehicle at the waypoint S3 include keeping the lane and right transition lane, the waypoint value of the waypoint S3 being V (S3) =min (V (S3) _{Keeping lane} ，V(S3) _{Right transition lane} ) The executable actions at waypoint S5 include left transition and lane keeping, with waypoint S5 having a waypoint value of V (S5) =min (V (S5) _{Keeping lane} ，V(S5) _{Left transition lane} ) The final calculated waypoint values for waypoint S3 and waypoint S5 are shown in fig. 4.

The calculation of the waypoint value for each waypoint between the target waypoint and the current location of the unmanned vehicle can be generalized as:

V(s)＝min _a∈A E _T [C(s，a，s′)+V(s′)]；

wherein V(s) is the waypoint value of waypoint s, C (s, a, s ') is the short term cost of unmanned vehicle performing action a from waypoint s to waypoint s ', V (s ') is the waypoint value of waypoint s ', A is the set of unmanned vehicle's executable actions at waypoint s, E _T (. Cndot.) is a function of the expected value based on the transfer model T.

After the waypoint value of the last waypoint corresponding to the target waypoint is obtained through calculation, the last waypoint corresponding to the target waypoint is taken as the target waypoint, the step S1031 is returned, the waypoint value of the last waypoint corresponding to the new target waypoint is calculated until the last waypoint corresponding to the target waypoint is the current position of the unmanned vehicle, and the waypoint values of all the waypoints between the target waypoint and the current position of the unmanned vehicle are obtained. Referring to fig. 4, after calculating the waypoint value of the last waypoint S3, S4, S5 corresponding to the target waypoint, taking the waypoint S3, S4, S5 as a new target waypoint, calculating the waypoint value of the last waypoint S6, S7, S8 corresponding to the waypoint S3, S4, S5, and calculating the short-term cost and the state transition probability of the waypoint S6, S7, S8 to the corresponding next waypoint, overlapping the corresponding short-term cost on the basis of the waypoint value of the waypoint S3, S4, S5, and calculating the waypoint value of the waypoint S6, S7, S8 in combination with the state transition probability, taking the waypoint S6, S7, S8 as the new target waypoint, calculating the waypoint value of the last point corresponding to the waypoint S6, S7, S8, and so on, and iteratively calculating the waypoint value of each point between the target waypoint and the current position of the unmanned vehicle.

If the waypoint value of each waypoint between the destination and the unmanned vehicle is directly calculated in an iterative manner from the destination direction, an optimal road decision is obtained, and when the unmanned vehicle is far away from the destination, the calculation amount of the process is large, so that the road decision efficiency of the unmanned vehicle is influenced; in the embodiment of the application, the waypoint value is calculated in two parts, one part is to obtain the global cost of the target waypoint through global search of the target waypoint, and the other part is to superimpose the global cost of the target waypoint on the waypoint between the target waypoint and the current position of the unmanned vehicle according to the short-term cost and the state transition probability of the unmanned vehicle for transferring between different waypoints, so that the calculated amount is reduced, the unmanned vehicle can obtain the optimal road decision result through a small amount of calculation, and the road decision efficiency is improved.

And 104, determining the action at the current position according to the waypoint value of the next waypoint corresponding to the current position of the unmanned vehicle in real time, and obtaining the current road decision result of the driving destination.

And carrying out road decision in real time according to the current waypoint value of each waypoint, and determining whether the action at the current position is to change the lane left, keep the lane or change the lane right. Specifically, a minimum waypoint value in the waypoint values of the next waypoints corresponding to the current position of the unmanned vehicle is determined in real time according to the waypoint values of the current lane where the current position of the unmanned vehicle is located and the next waypoints of the adjacent lanes of the current lane, and whether and how to change the lane is determined according to the position of the waypoint corresponding to the minimum waypoint value. Referring to fig. 4, it is assumed that the unmanned vehicle is currently located at the waypoint S4, and according to the waypoint values of the next waypoints S0, S1, S2 corresponding to the waypoint S4, the waypoint value of the waypoint S1 may be determined to be the smallest, and the waypoint S1 is located directly in front of the waypoint S4, i.e., the waypoint S1 and the waypoint S4 are located in the same lane, so that the unmanned vehicle selects to keep the lane straight to the waypoint S1 at the waypoint S4, i.e., the road decision result at the waypoint S4 is to keep the lane. When the unmanned vehicle reaches the waypoint S1, deciding whether and how to change the lane according to the minimum value of the waypoint S1 at the next waypoint of the current lane and the adjacent lane of the current lane, thereby obtaining a road decision result at the waypoint S4, and repeating the steps to make a road decision so as to drive to a destination.

In one embodiment, when the number of target waypoints between the current position of the unmanned vehicle and the destination is 1, after the waypoint value of each waypoint between the target waypoint and the current position of the unmanned vehicle is calculated according to the waypoint value of the target waypoint, carrying out road decision in real time according to the waypoint value of the next waypoint corresponding to the current position of the unmanned vehicle; when the unmanned vehicle runs according to the target waypoint of the road decision result, the waypoint value of each waypoint between the destination and the current position (namely the target waypoint) of the unmanned vehicle can be calculated according to the waypoint value of the destination, and then the road decision is carried out in real time according to the waypoint value of the unmanned vehicle at the next waypoint corresponding to the target waypoint, so that the unmanned vehicle can drive to the destination. The destination waypoint value may be set to 0 or other relatively small value, and the calculation of the destination to the destination waypoint value is similar to the calculation of the destination waypoint to the waypoint between the current location of the unmanned vehicle.

In another embodiment, to further improve the calculation efficiency, a plurality of target waypoints may be set between the current position of the unmanned vehicle and the destination at a time, where each target waypoint is spaced a certain distance along the traveling direction of the unmanned vehicle. After the waypoint values for each target waypoint are calculated in step 102, the target waypoint that the unmanned vehicle first arrives at may be taken as the first target waypoint (i.e., the closest target waypoint to the unmanned vehicle), the second target waypoint that arrives at may be taken as the second target waypoint (i.e., the second closest target waypoint to the unmanned vehicle), and so on. Calculating the waypoint value of each waypoint between the first target waypoint and the current position of the unmanned vehicle according to the waypoint value of the first target waypoint, and then carrying out road decision in real time according to the waypoint value of the next waypoint corresponding to the current position of the unmanned vehicle; when the unmanned vehicle runs to the first target waypoint according to the road decision result, the waypoint value of each waypoint between the second target waypoint and the current position of the unmanned vehicle (namely the first target waypoint) is calculated according to the waypoint value of the second target waypoint, and the like, when the unmanned vehicle reaches the last target waypoint, the waypoint value of each waypoint between the destination and the current position of the unmanned vehicle (the last target waypoint) can be calculated according to the waypoint value of the destination, and then the road decision is carried out in real time according to the waypoint value of the next waypoint corresponding to the last target waypoint, so that the unmanned vehicle runs to the destination.

Taking fig. 2 as an example, assuming that the destination is a certain position in front of the hub center 1, the unmanned vehicle is currently located behind the hub center 2, the hub center 1 and the hub center 2 exist between the destination and the current position of the unmanned vehicle, assuming that the entry points of the hub center 1 and the hub center 2 are selected as target waypoints, determining according to the distances between the entry points of the hub center 1 and the hub center 2 and the current position of the unmanned vehicle, determining that the waypoint corresponding to the entry point of the hub center 2 is a first target waypoint, determining that the waypoint corresponding to the entry point of the hub center 1 is a second target waypoint, and when calculating the waypoint values of other waypoints according to the waypoint values of the target waypoints, firstly, calculating the waypoint values of each waypoint between the entry point of the hub center 2 and the current position of the unmanned vehicle according to the waypoint of the next waypoint corresponding to the current position of the unmanned vehicle, and then making a road decision in real time; then, when the unmanned vehicle runs to a certain entrance point of the hub center 2, the navigation point value of the navigation point between the entrance point of the hub center 1 and the entrance point of the hub center 2 can be calculated according to the navigation point value of the entrance point of the hub center 1, and then a road decision is made; when the unmanned vehicle runs to a certain entrance point of the hub center 1, the waypoint value of the waypoint between the destination and the entrance point of the hub center 1 can be calculated according to the waypoint value of the destination, and then road decision is carried out according to the waypoint value, so that the unmanned vehicle runs to the destination. When the unmanned vehicle is far away from the destination, a plurality of target waypoints can be selected to calculate the waypoint value of each waypoint in stages, and the total calculated amount is shared in the calculation process of each stage, so that the calculation speed is improved, and the decision efficiency is further improved. By setting a plurality of target waypoints, the unmanned vehicle takes each target waypoint as a destination of each stage, and therefore gradually travels through each target waypoint and finally reaches the destination.

In the embodiment of the application, the travelable road between the unmanned vehicle and the destination is divided into the waypoints, the waypoint value of each waypoint is calculated, so that the unmanned vehicle can carry out road decision in each waypoint according to the waypoint value of the next waypoint, the complex road decision optimization problem is simplified, the waypoint value is calculated in two stages, the global cost from the target waypoint to the destination is calculated in the first stage, the waypoint value of the target waypoint is obtained, the waypoint value from the target waypoint is calculated in the second stage according to the waypoint value of the target waypoint in a reverse iteration manner, the calculation speed of the waypoint value is improved, the road decision efficiency is improved, the optimal road decision is obtained by adopting a model prediction control method in the prior art, the optimal road decision is obtained by solving the complex optimization problem, a large amount of calculation capacity is needed to solve the nonlinear optimization problem, and the technical problem of low road decision efficiency is caused.

The above is one embodiment of a road decision method provided in the present application, and the following is another embodiment of a road decision method provided in the present application.

The road decision method provided by the embodiment of the application comprises the following steps:

Step 201, dividing a travelable road between the current position of the unmanned vehicle and the destination into a plurality of waypoints, wherein the travelable road at least comprises a lane, and each lane comprises a plurality of waypoints which are connected in sequence.

Step 202, determining a target waypoint in the waypoints between the current position of the unmanned vehicle and the destination, and calculating the global cost from the target waypoint to the destination to obtain the waypoint value of the target waypoint.

And 203, iteratively calculating the waypoint value of the waypoint between the target waypoint and the current position of the unmanned vehicle according to the waypoint value of the target waypoint.

And 204, determining the action at the current position according to the waypoint value of the next waypoint corresponding to the current position of the unmanned vehicle in real time, and obtaining the current road decision result of the driving destination.

The specific contents of steps 201 to 204 are the same as those of steps 101 to 104, and will not be described here again.

The above steps are to obtain waypoint values based on static traffic information and make road decisions, while the traffic environment in which the unmanned vehicle is in during actual travel is dynamic, time-varying, and there are multiple other traffic participants that can dynamically influence the single step cost function and the unmanned vehicle's transfer model, ultimately affecting the waypoint values for each waypoint. Therefore, in the process of driving the unmanned vehicle, the waypoint value needs to be updated according to the traffic information, and then the road decision result is updated.

Further, the road decision method in the embodiment of the application further includes:

and 205, updating the road decision result according to the traffic information.

The specific updating process is as follows:

s2051, acquiring a special waypoint influenced by traffic information when the unmanned vehicle runs according to the current road decision result.

When the unmanned vehicle runs according to the current road decision result, traffic information can be acquired in real time through a sensor or the Internet of vehicles on the unmanned vehicle.

When the traffic information includes a stationary traffic participant (a roadside-parked vehicle, a traffic cone, or the like), a special waypoint affected by the stationary traffic participant is determined according to the location of the stationary traffic participant while the unmanned vehicle is traveling according to the current road decision result. When the static traffic participant is located at a certain waypoint, the waypoint is a special waypoint, please refer to fig. 5, in a traffic scene, the unmanned vehicle is driven on the right lane, a traffic cone is found to be located 30 meters in front to block the right lane, the unmanned vehicle can predict that the unmanned vehicle cannot drive to the waypoint S2 through the lane keeping at the waypoint S3, cannot drive to the waypoint S2 through the lane keeping at the waypoint S4, cannot drive to the waypoint S1 through the lane keeping at the waypoint S2, and cannot drive to the waypoint S0 through the lane keeping at the waypoint S2. That is, according to the position of the traffic cone, a specific waypoint that will be affected by the traffic cone in the future can be determined as waypoint S2. If a stationary traffic participant is located between two waypoints, for example, a traffic cone is located between waypoint S2 and waypoint S1 in fig. 5, the unmanned vehicle may predict that it is not possible to drive to waypoint S1 in the future at waypoint S2 by maintaining the lane, and thus may determine the affected particular waypoint as waypoint S2.

When the traffic information includes dynamic traffic participants, determining target dynamic traffic participants from the dynamic traffic participants when the unmanned vehicle is driving according to the current road decision result; and determining the special waypoint influenced by the target dynamic traffic participant according to the running speed of the target dynamic traffic participant and the running speed of the unmanned vehicle.

In the unmanned vehicle, there are a plurality of dynamic traffic participants (pedestrians, traveling vehicles, etc.) during traveling, and the calculation amount is very large if the dynamic influence of all the dynamic traffic participants is considered. In order to reduce the calculated amount, improve the updating speed of the waypoint value and further improve the road decision efficiency, in the embodiment of the application, the dynamic traffic participants with the running speed lower than the lane speed limit in the preset range in front of the unmanned vehicle are preferably considered.

Further, the specific process of determining a target dynamic traffic participant from among the dynamic traffic participants may be:

Calculating the confidence value of the potential target dynamic traffic participant according to the prior value of the potential target dynamic traffic participant and the judgment result;

The behavior of the dynamic traffic participant has uncertainty and in determining the particular waypoints that will be affected in the future by the target dynamic traffic participant, it is necessary to determine which target traffic participants are considered for the dynamic impact on the waypoint values. For example, if the front vehicle of the unmanned vehicle starts accelerating only slowly for 1 second, the influence of the front vehicle on the waypoint value is small, and the influence of the front vehicle may be disregarded, and if the front vehicle of the unmanned vehicle slowly travels for a period of time, the dynamic influence of the front vehicle on the waypoint value needs to be considered.

Specifically, after determining the potential target dynamic traffic participant, a priori value may be configured for the potential target dynamic traffic participant, after obtaining a determination result of whether the running speed of the potential target dynamic traffic participant is less than the speed limit value of the lane in which the potential target dynamic traffic participant is located, the determination result may be mapped into a numerical value through a mapping function, for example, the determination result of whether the running speed of the potential target dynamic traffic participant is less than the speed limit value of the lane in which the potential target dynamic traffic participant is located may be mapped into a numerical value 1, and the determination result of whether the running speed of the potential target dynamic traffic participant is greater than or equal to the speed limit value of the lane in which the potential target dynamic traffic participant is located is mapped into a numerical value 0; and then, carrying out weighted summation on the prior value of the potential target dynamic traffic participant and the mapping value corresponding to the judgment result through a preset weight coefficient to obtain the confidence value of the potential target dynamic traffic participant. When the confidence values of the potential target dynamic traffic participants in a period of time are all larger than the preset confidence threshold value, the potential target dynamic traffic participants are used as the target dynamic traffic participants, and the potential target dynamic traffic participants which are accelerated or decelerated suddenly can be prevented from being used as the target dynamic traffic participants.

After the target dynamic traffic participant is determined, a particular waypoint that will be affected by the target dynamic traffic participant in the future is determined based on the travel speed of the target dynamic traffic participant and the travel speed of the unmanned vehicle. Referring to fig. 7, in a traffic scenario, an unmanned vehicle (car 1) is traveling at a speed v ₁ At a constant speed, the front vehicle (car 2) of the unmanned vehicle is driven at a speed v ₂ At constant speed, wherein v ₁ >v ₂ It is assumed that the current decision result of obtaining the unmanned vehicle from the waypoint values calculated from the static traffic information is a kept lane, i.e., the right lane is the currently best lane. The waypoint values of these waypoints do not take into account the dynamic influence of the slowly moving preceding vehicle, due to v ₁ >v ₂ In a future area (the area is estimated by the speed difference between the unmanned vehicle and the front vehicle), namely, a shaded area in fig. 7, the unmanned vehicle approaches the front vehicle, so that the unmanned vehicle is affected by the front vehicle running slowly in the shaded area, and if the unmanned vehicle keeps straight, the speed of the unmanned vehicle needs to be reduced to follow the front vehicle, namely, the unmanned vehicleThe short term cost of the vehicle moving from waypoint S2 to waypoint S1 increases, thereby affecting the waypoint value of waypoint S2, i.e., waypoint S2 is a special waypoint that will be affected by the vehicle ahead in the future.

S2052, when the special waypoint is a waypoint which can be reached in the future according to the current road decision result, determining the next waypoint of the special waypoint according to the current road decision result, and updating the short-term cost from the special waypoint to the next waypoint.

When the special waypoint is the waypoint affected by the static traffic participant, as shown in fig. 5, the special waypoint is the waypoint S2, the next waypoint of the waypoint S2 can be determined to be the waypoint S1 according to the current road decision result (lane keeping), and because the traffic cone exists at the position of the waypoint S2, the unmanned vehicle cannot reach the waypoint S1 from the waypoint S2, the short-term cost C (S2, lane keeping, S1) from the special waypoint S2 to the waypoint S1 can be updated to a larger value (such as 50, 100, etc.), and the specific value can be set according to the actual situation.

Further, when the special waypoint is a waypoint affected by the target dynamic traffic participant, the update process of the short-term cost of the special waypoint to the corresponding next waypoint is:

Taking fig. 7 as an example, the next waypoint of the special waypoint S2 can be determined as the waypoint S1 according to the current road decision result (lane keeping), and the driving speed v of the front vehicle of the unmanned vehicle according to the driving distance d between the special waypoint S2 and the waypoint S1 ₂ The updated short-term cost S/v of the particular waypoint S2 to waypoint S1 may be calculated ₂ Embodiments of the present application further contemplate that the dynamic impact of the targeted dynamic traffic participant may last for a certain period of time,thus, the updated short-term cost C (s, a, s ') that eventually transitions from a particular waypoint s to the next waypoint s' at the execution of action a may be expressed as:

C(s，a，s′)＝β*(d _s′-s /v)，

where β is a cutoff parameter for determining the duration of the dynamic impact of the target dynamic traffic participant, d _s′-s And v is the driving speed of the target dynamic traffic participant for the driving distance from the special waypoint s to the corresponding next waypoint s'.

S2053, updating the waypoint value of the special waypoint based on the short-term cost from the special waypoint to the next waypoint after updating.

According to the above steps, the waypoint value of one waypoint is calculated from the waypoint value of the next waypoint corresponding to the waypoint, the short-term cost of transition between the waypoints and the state transition probability, and after the short-term cost is updated, the corresponding waypoint value is updated. It will be appreciated that if the state transition probabilities are updated, the corresponding waypoint values are updated.

Taking fig. 5 as an example, assuming that in the static traffic environment, the short-term cost of transferring between waypoints is 1, the lane change success rate is 20%, and the short-term cost C after the special waypoint S2 is transferred to the waypoint S1 (S2, keep lane, S1) =100, since the special waypoint S2 cannot be transferred to the waypoint S0, the short-term cost of transferring the special waypoint S2 to the waypoint S0 will also increase, and assuming that the short-term cost C after the special waypoint S2 is transferred to the waypoint S0 (S2, left transition lane, S0) =100.

If a left transition channel is selected at a specific waypoint S2, the updated waypoint value is V (S2) _{Left transition lane} ＝(91+100)*20％+(51+100)*80％＝159；

If lane keeping is selected at the special waypoint S2, the updated waypoint value is V (S2) _{Keeping lane} ＝(51+100)*100％＝151；

Finally, the updated waypoint value for the special waypoint S2 is min (V (S2) _{Left transition lane} ，V(S2) _{Keeping lane} )＝151。

It will be appreciated that if the traffic cone is between waypoint S2 and waypoint S1, i.e., between The special waypoint S2 may pass through the left transition to waypoint S0, at which time the short term cost of the transition of the special waypoint S2 to waypoint S0 remains unchanged, i.e., C (S2, left transition, S0) =1. At this time, if a left transition channel is selected at the special waypoint S2, the updated waypoint value is V (S2) _{Left transition lane} = (91+1) 20++ (51+100) 80% = 139; finally, the updated waypoint value for the special waypoint S2 is min (V (S2) _{Left transition lane} ，V(S2) _{Keeping lane} )＝139。

Taking fig. 8 as an example, assume that in a static traffic environment, the short-term cost of transferring between waypoints is 1, the lane-changing success rate is 20%, the waypoint value calculated according to the static traffic information is shown in fig. 8, and the short-term cost after the calculated special waypoint S2 is transferred to the waypoint S1 is 30.

If a left transition channel is selected at a specific waypoint S2, the updated waypoint value is V (S2) _{Left transition lane} ＝(84+1)*20％+(52+30)*80％＝83；

If lane keeping is selected at the special waypoint S2, the updated waypoint value is V (S2) _{Keeping lane} ＝(52+30)*100％＝82；

Finally, the updated waypoint value for the special waypoint S2 is min (V (S2) _{Left transition lane} ，V(S2) _{Keeping lane} )＝82。

S2054, reversely and iteratively updating the waypoint value of each waypoint between the special waypoint and the current position of the unmanned vehicle according to the updated waypoint value of the special waypoint, and returning to the step 204.

As can be seen from fig. 5, the unmanned vehicle cannot reach the waypoint S2 by keeping the lane at the waypoint S3, so the short-term cost from the waypoint S3 to the special waypoint S2 also needs to be updated, assuming that the updated short-term cost C (S3, keep the lane, S2) =100.

If a left transition channel is selected at the waypoint S3, the updated waypoint value is V (S3) _{Left transition lane} ＝(84+1)*20％+(139+100)*80％＝208；

If lane keeping is selected at the waypoint S3, the updated waypoint value is V (S3) _{Keeping lane} ＝(139+100)*100％＝239；

Finally, the updated waypoint value for waypoint S3 is min (V (S3) _{Left turnLane changing} ，V(S3) _{Keeping lane} )＝208。

The short-term cost from the waypoint S4 to the special waypoint S2 is also required to be updated correspondingly, and the updating process of the waypoint value of the waypoint S4 is similar to that of the waypoint value of the waypoint S3, and will not be described in detail herein. After the waypoint values of the waypoints S3 and S4 are updated, the waypoint values of the waypoints S3 and S4 to the waypoint between the current positions of the unmanned vehicle are iteratively updated in a reverse direction. It should be noted that, the short-term costs corresponding to the waypoints from the waypoint S3, the waypoint S4 to the current location of the unmanned vehicle remain unchanged.

After the waypoint value in fig. 5 is updated, the updated waypoint value is shown in fig. 6, and according to the updated waypoint value, it can be known that the unmanned vehicle transitions from the left lane to the left lane at the current waypoint, and further exceeds the traffic cone. After updating the waypoint values in fig. 8, the updated waypoint values are shown in fig. 9, and it is known from the updated waypoint values in fig. 9 that the unmanned vehicle transitions the left lane to overrun the front slow vehicle.

The calculation of the waypoint values for the waypoints in the static traffic environment does not take into account time, i.e. the influence of the dynamic traffic environment. When there is a dynamic traffic participant in front of the unmanned vehicle, the unmanned vehicle needs to take a huge time to advance from the current waypoint to the next waypoint in front, that is, the short-term cost of transferring the unmanned vehicle between the waypoints is closely related to the traffic environment, the short-term cost is updated according to the traffic information of each frame and is dynamically changed, and accordingly, the waypoint value is dynamically changed. In this embodiment, the update formula of the waypoint value of each waypoint between the specific waypoint and the current location of the unmanned vehicle may be expressed as:

wherein V(s) is the updated waypoint value of the waypoint s, C _t (s, a, s ') is the short term cost of the unmanned vehicle performing action a from waypoint s to waypoint s ' at the current time t, V (s ') is the waypoint value of waypoint sA is the executable action set of the unmanned vehicle at the waypoint s,

for a traffic participant set based on a time-varying transfer model T and at the current time T>

Is a function of the expected value of (a).

The transition model becomes time dependent due to the presence of other traffic participants. At each moment the current waypoint (i.e. the current state) of the unmanned vehicle is known, the state reached by the unmanned vehicle to select a certain executable action (left transition, right transition or lane keeping) is uncertain, e.g. the traffic density of the target lane change approaches its capacity or the vehicle behind the target lane change is approaching rapidly, the unmanned vehicle is not necessarily able to successfully change lane to the target lane change even if a lane change action is made. Therefore, it is necessary to dynamically update the lane change success rate of the transition between waypoints by observing traffic information around the unmanned vehicle. The executable action is determined by the lane in which the unmanned vehicle is located, for example, the unmanned vehicle is in the rightmost lane, the unmanned vehicle has no travelable road on the right side, at this time, the right transition road is an inexecutable action, and the straight and left transition roads are executable actions.

In the embodiment of the application, for the waypoints outside the preset range of the unmanned vehicle, the lane change success rate P (succc) between the waypoints outside the preset range of the unmanned vehicle _·t =1) inheriting the lane change success rate P calculated in the static traffic environment ₀ I.e. P (succc) _·t ＝1)＝P ₀ The method comprises the steps of carrying out a first treatment on the surface of the And updating the lane change success rate of the transfer between the waypoints in the preset range of the unmanned vehicle according to the traffic information for the waypoints in the preset range of the unmanned vehicle.

Specifically, the current distance between the rear side vehicle of the unmanned vehicle and the current yield probability of the rear side vehicle of the unmanned vehicle are obtained according to traffic information, and the lane change success rate of the unmanned vehicle at the current waypoint is updated.

For the lane change success rate of the unmanned vehicle at the current waypoint, the current distance d between the rear side vehicle of the unmanned vehicle and the unmanned vehicle needs to be considered _t And the current yield probability P (success) of the vehicle on the rear side of the unmanned vehicle _·t ＝1|y _t ) (yield willingness y of rear side vehicle) _t The influence of (a), i.e., the lane change success rate P (succc) of the unmanned vehicle at the current waypoint _·t ＝1|d _t ，y _t ) Can be expressed as:

P(succ _·t ＝1|d _t ，y _t )∝P(succ _·t ＝1|d _t )·P(succ _·t ＝1|y _t )；

wherein P (succc) _·t ＝1|d _t ) For controlling lane change success rate, P (vacc) _·t ＝1|y _t ) For controlling the lane change success rate according to the coordination of the rear vehicles, and the symbol is proportional to the number of the lane change success rate.

Further, P (succc) _·t ＝1|d _t ) The calculation formula of (2) can be:

wherein P is ₀ The lane change success rate of the current waypoint is calculated in a static traffic environment, namely the lane change success rate before updating the current waypoint; d, d _safe For safety lane change distance, when d _t ＝d _safe When P (succc) _·t ＝1|d _t )＝P ₀ 。

Further, the current yield probability of the rear side vehicle of the unmanned vehicle is calculated by the following steps:

and calculating the current yield probability of the rear side vehicle according to the current acceleration of the rear side vehicle of the unmanned vehicle and the yield probability of the rear side vehicle at the previous moment, wherein the initial yield probability of the rear side vehicle is obtained through initialization. P (succ) _·t ＝1|y _t ) The calculation formula of (2) can be expressed as:

P(succ _·t ＝1|y _t )＝αP(succ _·t-1 ＝1|y _t-1 )+(1-α)II(a _t ＜0)；

wherein P (succ) _·t ＝1|y _t ) P (succc) is the current yield probability of the vehicle behind the unmanned vehicle _·t-1 ＝1|y _t-1 ) For the yielding probability of the rear side vehicle at the previous moment, alpha is the update rate, a _t For the current acceleration of the rear-side vehicle, II (x) is a mapping function, when the event is true, II (x) =1, when the event is false, II (x) =0, i.e. when a _t When < 0, II (a) _t < 0) =1, when a _t When not less than 0, II (a) _t ＜0)＝0。

The initial yield probabilities of the rear vehicles are obtained through initialization, the initial yield probabilities of different rear vehicles can be the same initial value, and the yield probabilities of the rear vehicles can be updated according to the reaction of the rear vehicles in the driving process.

For the rest waypoints in the preset range of the unmanned vehicle, namely other waypoints except the current waypoint of the unmanned vehicle in the preset range of the unmanned vehicle, according to the traffic density rho of the target lane change lane _t Updating the lane change success rate of the remaining waypoints in the preset range of the unmanned vehicle, wherein the target lane change is a lane after lane change and can be expressed as:

wherein P (succ) _·t ＝1|ρ _t ) Lane change success rate of remaining waypoints in preset range for unmanned vehicles under traffic density at time t, beta is attenuation factor and ρ _t For the traffic density of the target lane change at the time t, delta is the traffic capacity of the target lane change, and P _max Is a lane change success rate threshold.

Further, when the special waypoint affected by the static traffic participant is a waypoint that will be reached in the future according to the current road decision result, and the adjacent lane of the lane where the special waypoint is located cannot pass, the method in the embodiment of the application further includes:

dividing a lane separation line between a lane where a special waypoint influenced by a static traffic participant is located and an adjacent lane into a plurality of waypoints which are connected in sequence; and calculating the waypoint value of each waypoint on the lane separation line according to the waypoint values of the waypoints on the adjacent lanes of the lane separation line, the short-term cost of the transfer between the waypoints and the state transfer probability, and returning to the step 204. The short-term cost of transferring between waypoints on the lane dividing line is higher than the short-term cost of transferring between waypoints on a normal lane, and the specific value can be set according to actual conditions.

For example, as shown in fig. 10, in a traffic scenario, there are two lanes in front of the unmanned vehicle, the unmanned vehicle runs on the right lane, there is a traffic cone in front of the right lane, and the left lane leads to the dead-end lane, so that the calculated waypoint value of each waypoint on the left lane is much higher than the waypoint value of the waypoint on the right lane, that is, the short-term cost of the unmanned vehicle changing from the right lane to the left lane is high, and there is a traffic cone in front of the right lane, and the unmanned vehicle cannot keep straight at all times, in this case, the lane separation line between the left lane and the right lane (i.e., the solid line in fig. 10) can be divided into a plurality of waypoints connected in sequence, then the waypoint value of the waypoint on the lane separation line is calculated by the waypoint value update formula in step S2053, and when the waypoint value of the waypoint on the division line is smaller than the waypoint value of the updated by the right lane and the waypoint value of the left lane, the unmanned vehicle can overrun the cost of the lane on the lane change to the lane. Under a static traffic environment, assuming that the short-term cost of transfer between waypoints on a lane is 1, the short-term cost of transfer between waypoints on a lane dividing line is 30, and the updated short-term cost from a special waypoint on a right lane to a next waypoint is 100 due to the influence of a traffic cone, the waypoint values of the waypoints on the lane dividing calculated based on the updated short-term cost are as shown in fig. 10, and in the scene, after an unmanned vehicle keeps going straight for a period of time, the unmanned vehicle changes the lane to travel on a lane dividing line so as to exceed the traffic cone.

In the embodiment of the application, if the optimal road decision is obtained by adopting a model predictive control method, the optimal road decision is obtained by solving a complex optimization problem, a large amount of computing capacity is required to solve a nonlinear optimization problem, the construction of an environment model is seriously depended, and the method is difficult to be effectively applied to a decision system of an unmanned vehicle. The embodiment of the application solves the optimization problem in two parts, wherein one part obtains the global cost of the target waypoint through global search of the target waypoint, and the other part dynamically corrects the short-term cost and the lane change success rate paid by transferring between different states through observing real-time traffic information, so that the optimization problem of the high-dimensional multi-agent is simplified into the optimization problem of the low-dimensional single agent, and the solving speed is higher. The method has the advantages that the global cost and the short-term cost of a feasible road of the unmanned vehicle are rapidly and quantitatively analyzed in real time, the short-term cost and the global cost of the road are balanced, and the unmanned vehicle can acquire an optimal road decision result through a small amount of calculation, so that the unmanned vehicle can actively change the road, ultra-slowly change the road, actively change the road and leave a potential risk area (such as a construction area, a traffic accident area and the like), actively change the road and avoid a priority vehicle (such as a police car, an ambulance and the like) and the like in an optimal time.

The above is another embodiment of a road decision method provided in the present application, and the following is an embodiment of a road decision system provided in the present application.

Referring to fig. 11, a road decision system provided in an embodiment of the present application includes:

the second calculation module is used for iteratively calculating the waypoint value of the waypoint between the target waypoint and the current position of the unmanned vehicle according to the waypoint value of the target waypoint;

and the decision module is used for determining the action at the current position according to the waypoint value of the next waypoint corresponding to the current position of the unmanned vehicle in real time to obtain the current road decision result of the driving destination, wherein the action is left transition, lane keeping or right transition.

As a further refinement, the first calculation module is specifically configured to:

calculating the travel time of the unmanned vehicle from the target waypoint to the destination based on the shortest path and the preset travel speed;

acquiring global cost from a target waypoint to a destination based on travel time of the unmanned vehicle from the target waypoint to the destination;

As a further improvement, the road decision system in the embodiment of the present application further includes: the waypoint value updating module is used for:

acquiring special waypoints influenced by traffic information when the unmanned vehicle runs according to the current road decision result;

when the special waypoint is a waypoint which can be reached in the future when the vehicle runs according to the current road decision result, determining the next waypoint of the special waypoint according to the current road decision result, and updating the short-term cost from the special waypoint to the next waypoint;

and reversely and iteratively updating the waypoint value of each waypoint between the special waypoint and the current position of the unmanned vehicle according to the updated waypoint value of the special waypoint, and triggering a decision module.

As a further improvement, the road decision system in the embodiment of the present application further includes: a third calculation module for:

when the special waypoints influenced by the static traffic participants are waypoints which can be reached in the future according to the current road decision result and the adjacent lanes of the lanes where the special waypoints are located cannot pass, dividing the lane separation line between the lanes where the special waypoints influenced by the static traffic participants are located and the adjacent lanes into a plurality of waypoints which are connected in sequence;

and calculating the waypoint value of each waypoint on the lane separation line according to the waypoint value of the waypoints on the adjacent lanes of the lane separation line, the short-term cost of the transfer between the waypoints and the state transfer probability, and triggering a decision module.

As a further improvement, the state transition probability includes a lane change success rate, and the road decision system in the embodiment of the present application further includes:

and the lane change success rate updating module is used for updating the lane change success rate when transferring between waypoints in the preset range of the unmanned vehicle according to the traffic information.

As a further improvement, the lane change success rate update module is specifically configured to:

acquiring the current distance between the rear side vehicle of the unmanned vehicle and the current yield probability of the rear side vehicle of the unmanned vehicle according to traffic information, and updating the lane change success rate of the unmanned vehicle at the current waypoint;

Updating the lane changing success rate of the rest waypoints in the preset range of the unmanned vehicle according to the traffic density of the target lane changing lane, wherein the target lane changing lane is the lane after lane changing, and the rest waypoints in the preset range of the unmanned vehicle are other waypoints except the current waypoints in the preset range of the unmanned vehicle.

The embodiment of the application also provides road decision equipment, which comprises a processor and a memory;

the memory is used for storing the program codes and transmitting the program codes to the processor;

the processor is configured to execute the road decision method in the foregoing method embodiment according to instructions in the program code.

The embodiment of the application also provides a computer readable storage medium, and the computer readable storage medium is used for storing program codes, and the program codes are executed by a processor to realize the road decision method in the embodiment of the method.

It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, which are not repeated herein.

The terms "first," "second," "third," "fourth," and the like in the description of the present application and in the above-described figures, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that embodiments of the present application described herein may be capable of operation in sequences other than those illustrated or described herein, for example. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

It should be understood that in this application, "at least one" means one or more, and "a plurality" means two or more. "and/or" for describing the association relationship of the association object, the representation may have three relationships, for example, "a and/or B" may represent: only a, only B and both a and B are present, wherein a, B may be singular or plural. The character "/" generally indicates that the context-dependent object is an "or" relationship. "at least one of" or the like means any combination of these items, including any combination of single item(s) or plural items(s). For example, at least one (one) of a, b or c may represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", wherein a, b, c may be single or plural.

In the several embodiments provided in this application, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of the units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a storage medium, including several instructions to execute all or part of the steps of the methods described in the embodiments of the present application by a computer device (which may be a personal computer, a server, or a network device, etc.). And the aforementioned storage medium includes: u disk, mobile hard disk, read-Only Memory (ROM), random access Memory (Random Access Memory, RAM), magnetic disk or optical disk, etc.

The above embodiments are merely for illustrating the technical solution of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the corresponding technical solutions.

Claims

1. A method of road decision making, comprising:

2. The road decision method of claim 1, wherein the calculating the global cost of the target waypoint to the destination to obtain the waypoint value for the target waypoint comprises:

3. The road decision method of claim 2, wherein the obtaining the global cost of the target waypoint to the destination based on the travel time of the unmanned vehicle from the target waypoint to the destination comprises:

4. The road decision method according to claim 1, wherein, when iteratively calculating the waypoint value of the waypoint between the target waypoint and the current location of the unmanned vehicle according to the waypoint value of the target waypoint, the calculation process of the waypoint value of the last waypoint corresponding to the target waypoint is:

5. The road decision method of claim 1, wherein a plurality of the target waypoints are provided between the current location of the unmanned vehicle and the destination.

6. The road decision method of claim 4, further comprising:

7. The road decision method of claim 6, wherein when the traffic information includes static traffic participants;

8. The road decision method of claim 7, further comprising:

9. The road decision method of claim 6, wherein when the traffic information includes dynamic traffic participants;

10. The road decision method of claim 9, wherein said determining a target dynamic traffic participant from said dynamic traffic participants comprises:

11. The road decision method of claim 9, wherein when the special waypoint is a waypoint that will be reached in the future when traveling according to the current road decision result, determining a next waypoint of the special waypoint according to the current road decision result, and updating the short-term cost of the special waypoint to the next waypoint comprises:

12. The road decision method of claim 6, wherein the state transition probabilities include lane change success rates, the method further comprising:

13. The road decision method as claimed in claim 12, wherein updating the lane change success rate at the time of transition between waypoints within the preset range of the unmanned vehicle according to the traffic information comprises:

14. The road decision method of claim 13, wherein the calculation process of the current yield probability of the rear side vehicle of the unmanned vehicle is:

15. A road decision making system, comprising:

16. A road decision device, characterized in that the device comprises a processor and a memory;

The processor is configured to perform the road decision method of any of claims 1-14 according to instructions in the program code.

17. A computer readable storage medium, characterized in that the computer readable storage medium is for storing a program code which, when executed by a processor, implements the road decision method of any of claims 1-14.