WO2022166239A1

WO2022166239A1 - Vehicle travel scheme planning method and apparatus, and storage medium

Info

Publication number: WO2022166239A1
Application number: PCT/CN2021/122425
Authority: WO
Inventors: 丁涛
Original assignee: 华为技术有限公司
Priority date: 2021-02-03
Filing date: 2021-09-30
Publication date: 2022-08-11
Also published as: CN114861514A

Abstract

A vehicle travel scheme planning method and apparatus, and a storage medium. The method comprises: acquiring a navigation route from a starting point to a destination at a designated moment, wherein the navigation route comprises one or more road segment units, and each of the one or more road segment units is a road segment between two waypoints; acquiring historical traffic data information of the navigation route; on the basis of the historical traffic data information, evaluating the trajectory cost of a road traffic model of each of the one or more road segment units by taking time as a baseline, so as to obtain an evaluation result that meets a planning requirement; according to the evaluation result that meets a planning requirement, determining a departure time and an arrival time respectively corresponding to the two waypoints of each road segment unit; and according to the departure time and the arrival time respectively corresponding to the two waypoints of each road segment unit, determining a travel scheme, which meets the planning requirement, on the navigation route.

Description

A planning method, device and storage medium for a vehicle driving scheme

This application claims the priority of the Chinese patent application filed with the State Intellectual Property Office of China on February 3, 2021, the application number is 202110150312.0, and the application name is "a planning method, device and storage medium for a vehicle driving scheme", all of which are The contents are incorporated herein by reference.

technical field

The present application relates to the technical field of automatic driving, and in particular, to a planning method, device and storage medium for a vehicle driving scheme.

Background technique

Self-driving car is a kind of intelligent car, also known as self-driving car, unmanned car, mainly relying on the control equipment based on computer system in the car to achieve the purpose of driverless. Autonomous vehicles can often sense their surroundings and navigate without human intervention. The control device used to control unmanned driving is used as a vehicle terminal device, sometimes also called electronic control unit (ECU), domain control unit (DCU), or mobile data center (mobile data center, MDC) etc. Before the unmanned vehicle automatically drives on the road, the control device can plan a route that can avoid obstacles and conform to the vehicle dynamics, and control the car to drive meticulously according to the planned trajectory. It’s a bit similar to how the brain issues an order for the hand to take something. As for how to take it, the hand itself completes it.

When planning a route, you first need to plan a navigation route, and the navigation route may be a route with a large span. For example, if a passenger needs to go home from the airport, open the map stored in the on-board terminal, search for "airport" to "home", and there are 35 kilometers, the control device will plan out several routes from the airport to home and display them on the map of the on-board terminal. A route is a navigation route. After the unmanned vehicle determines a navigation route from these routes and automatically drives on the road, it needs to make an evasion according to whether there is a car in front of the current route, and it needs to consider the traffic lights, when to accelerate and when to decelerate. Therefore, it is also necessary to plan a short-term route according to the current real-time driving situation, and the short-term route is often within the range of 100-200 meters. Since the unmanned vehicle is driving, other vehicles are also driving, so the previously planned short-term route may not always be suitable, and the short-term route will be dynamically adjusted every 0.1 seconds.

However, when using the existing scheme to plan the navigation route, how to reasonably plan the driving scheme and improve the driving experience is a problem that needs to be solved.

SUMMARY OF THE INVENTION

In order to solve the above problems, the embodiments of the present application provide a planning method, device and storage medium for a vehicle driving scheme.

In a first aspect, an embodiment of the present application provides a method for planning a vehicle driving scheme, the method includes: the method includes: acquiring a navigation route from a starting point to a destination at a specified time, the navigation route includes one or more There are road section units, each of the one or more road section units is a road section between two road points respectively; obtain historical traffic data information of the navigation route; based on the historical traffic data information Or the road traffic model of each road segment unit in the multiple road segment units uses time as the baseline to evaluate the trajectory cost to obtain an evaluation result that meets the planning requirements; according to the evaluation results that meet the planning requirements, determine the two roads for each road segment unit. The travel time and the arrival time corresponding to the points respectively; according to the travel time and the arrival time respectively corresponding to the two waypoints of each road segment unit, the driving plan on the navigation route that meets the planning requirements is determined.

In this way, when the unmanned vehicle carries out the driving planning of automatic driving, it can plan based on the historical traffic experience of the current road, and obtain the optimal driving trajectory in the whole process, so that the unmanned vehicle can reach the specific destination within the planned time and obtain the most optimal driving trajectory. Excellent driving experience.

In a possible implementation manner, the determining, according to the travel time and the arrival time respectively corresponding to the two waypoints in each road segment unit, the driving plan on the navigation route that meets the planning requirement includes: according to the The travel time and arrival time corresponding to the two waypoints in each road segment unit respectively determine the driving speed of the vehicle on each road segment unit; according to the travel time and arrival time corresponding to the two waypoints in each road segment unit respectively The time and the travel speed of the vehicle on each road segment unit are used to determine a travel plan on the navigation route that meets the planning requirements.

In a possible implementation manner, the historical traffic data information of the navigation route includes one or more of the following:

On the one or more road section units, the vehicle control information, travel information, and traffic participant information and traffic information around the vehicle at multiple different times in a historical time period.

In this way, the historical traffic data information is classified with the time as the baseline and the fixed-length road segment unit as the statistical unit, so as to facilitate the classification statistics and feature extraction of the data with the time as the dimension.

In a possible implementation manner, the one or more road segment units are obtained by dividing the navigation route according to the length of the road segment, or the one or more road segment units are obtained by dividing the navigation route by traffic flow or traffic elements route obtained.

In this way, with the fixed-length road section unit as the statistical unit, the historical traffic data information of each road section unit is obtained, which is convenient for the classification statistics and feature extraction of the subsequent data. Taking the road section units divided according to traffic flow and traffic elements as the statistical unit, the historical traffic data information of each road section unit is obtained, which is convenient for the classification statistics and feature extraction of the subsequent data.

In a possible implementation manner, the method further includes establishing a road traffic model of each of the one or more road segment units according to the historical traffic data information.

In a possible implementation manner, the establishing a road traffic model of each of the one or more road segment units according to the historical traffic data information includes: taking time as a baseline from the historical traffic data information Extract one or more traffic features on each road segment unit from the pass model.

In this way, the modeling elements can be determined based on the traffic characteristics with the time as the baseline, and the road traffic models at different times with the time as the baseline can be determined based on the modeling elements, thereby ensuring the accuracy and comprehensiveness of the road traffic model.

In a possible implementation manner, establishing, according to one or more traffic characteristics on each road segment unit, multiple time-based road traffic models on each road segment unit, including: One or more traffic features on each road segment unit are accumulated to establish a plurality of road traffic models on each road segment unit based on time.

In this way, the modeled element values can be determined by selecting several traffic characteristics based on the time as the baseline, and the road traffic models at different times can be determined based on several modeled element values, so as to ensure the accuracy and comprehensiveness of the road traffic model. sex.

In a possible implementation manner, establishing, according to one or more traffic characteristics on each road segment unit, multiple time-based road traffic models on each road segment unit, including: One or more traffic features on each road segment unit establish a time-based road traffic model through an RNN cyclic neural network.

In a possible implementation manner, the method further includes: establishing an evaluation system of the road traffic model according to the historical traffic data information, where the evaluation system of the road traffic model includes the one or more The trajectory cost score corresponding to the traffic feature; the trajectory cost score corresponding to the one or more traffic features at different times is set, and an evaluation system for the road traffic model is established.

In this way, specific evaluation criteria of the road traffic model are established according to the traffic characteristics, so as to determine the scoring criteria of each traffic feature at different times, so as to evaluate the pros and cons of the road traffic model planning the travel plans at different times.

In a possible implementation, the traffic characteristics include one or more of the following: vehicle control characteristics, travel characteristics, traffic participant characteristics, traffic characteristics, arrival time characteristics, traffic light waiting time characteristics, and green light travel time characteristics.

In this way, comprehensive modeling data is covered by a plurality of different traffic characteristics, thereby ensuring the accuracy and comprehensiveness of the road traffic model.

In a possible implementation manner, the trajectory cost evaluation is performed on the road traffic model of each road segment unit based on the historical traffic data information with a time as a baseline, and an evaluation result that meets planning requirements is obtained, including: according to the road The traffic model evaluation system sums the trajectory cost scores corresponding to a plurality of the traffic features of each road traffic model with the time as the baseline or weighted sum, and obtains the traffic model of each road segment unit at different times. Comprehensive trajectory cost value; compare the comprehensive trajectory cost value with time as a baseline to obtain an evaluation result that meets the planning requirements.

In this way, the trajectory cost of the road traffic model at different times with the time as the baseline can be evaluated to obtain a road traffic model that meets the planning requirements.

In a possible implementation manner, the two waypoints of each road segment unit are a start waypoint and an end waypoint, respectively, and the two waypoints of each road segment unit are determined according to the evaluation result that meets the planning requirements. The travel time and arrival time corresponding to the road points respectively include: obtaining the time information corresponding to the road traffic model of each road segment unit according to the evaluation result that meets the planning requirements, and determining the starting waypoint of each road segment unit. Travel time; according to the travel time of the starting waypoint, combined with the road speed limit, specify the arrival time of the ending waypoint of each road segment unit; obtain the travel from the starting waypoint to the ending waypoint time and arrival time.

In this way, the trajectory cost evaluation can be performed on the road traffic models at different times with the time as the baseline based on the traffic characteristics, so as to obtain a road traffic model that meets the planning requirements.

In a possible implementation manner, the two waypoints of each road segment unit are a start waypoint and an end waypoint, respectively, and the two waypoints of each road segment unit are determined according to the evaluation result that meets the planning requirements. The travel time and arrival time corresponding to each waypoint, including:

According to the evaluation result that meets the planning requirements, the time information corresponding to the road traffic model of each road segment unit is obtained, and the travel time of each road segment unit at the starting waypoint is determined; The travel time of the starting waypoint of the unit, combined with the road speed limit situation, calculates the interval of the arrival time of the ending waypoint of each road segment unit; according to the interval of the arrival time of the ending waypoint of each road segment unit, Planning the travel time of the next road segment unit of each road segment unit; taking the travel time of the next road segment unit of each road segment unit as the arrival time of each road segment unit, obtain two waypoints for each road segment unit Corresponding travel time and arrival time, respectively.

In a possible implementation manner, the determining the travel speed of the vehicle on each road segment unit according to the travel time and the arrival time corresponding to the two waypoints in each road segment unit respectively includes: according to the The travel time and arrival time corresponding to the two waypoints of each road section unit respectively, and the speed of the vehicle on the each road section unit is determined by using the speed optimization formula, and the speed optimization formula is:

Among them, f is the optimization result of the optimization function, and the optimization goal is to minimize f when planning the speed of the unmanned vehicle; _si represents the travel of the ith actual waypoint;

Represents the _i -th planned waypoint trip; ws represents the position deviation weight;

represents the acceleration bias weight;

Indicates the acceleration of the vehicle, and the subscript i is the planned waypoint of the ith;

Indicates the weight of the deviation of the speed value of the vehicle in the acceleration change;

represents the speed of the acceleration change, the subscript i→i+1 is the planned waypoint from the ith to the ith+1; w _t represents the arrival time deviation weight; t _i represents the travel from the ith waypoint to the destination the estimated time;

Represents the planned arrival time when traveling from the i-th waypoint.

In a possible implementation manner, using the speed optimization formula to determine the driving speed of the vehicle on each road segment unit, further comprising:

Set waypoint trip _si , speed

acceleration

the upper and lower limits of ;

Set waypoint trip _si , speed

acceleration

The relationship between the constraints is:

set acceleration

and the speed of acceleration change

The relational constraints are:

In this way, the data information provided by the road traffic model that meets the planning requirements can be used, and the traffic target can be formulated in combination with the current environmental information.

In the second aspect, the embodiments of the present application provide an apparatus for planning a driving scheme of a vehicle, which may include a module for executing the corresponding module in any of the foregoing embodiments, and the module may be software, hardware, or software and hardware. For beneficial effects, reference can be made to the description in the first aspect. For example, the apparatus may include: a route determination module for acquiring a navigation route from a starting point to a destination at a specified time, the navigation route including one or more road segment units, each road segment in the one or more road segment units The units are respectively a road section between two waypoints; a data acquisition module is used for acquiring historical traffic data information of the navigation route; a model evaluation module is used for the one or more road segments based on the historical traffic data information The road traffic model of each road segment unit in the unit conducts trajectory cost evaluation with time as the baseline, and obtains an evaluation result that meets the planning requirements; the time planning module is used to determine the two units of each road segment unit according to the evaluation result that meets the planning requirements. travel time and arrival time corresponding to each waypoint respectively; and a scheme planning module, configured to determine the travel time and arrival time corresponding to the two waypoints in each road segment unit respectively, which satisfies the planning requirement on the navigation route driving plan.

In a third aspect, the embodiments of the present application provide an electronic device, and reference may be made to the description in the first aspect for beneficial effects. The electronic device includes a memory and a processor; the processor is configured to execute computer-executed instructions stored in the memory, and the processor executes the computer-executed instructions to execute the unmanned driving plan described in any one of the foregoing embodiments Methods.

In a fourth aspect, an embodiment of the present application provides a vehicle, the vehicle including the device described in the second aspect or the third aspect.

In a fifth aspect, an embodiment of the present application provides a storage medium, and reference may be made to the description in the first aspect for beneficial effects. The storage medium includes a readable storage medium and a computer program stored in the readable storage medium, where the computer program is used to implement the method for unmanned driving planning described in any one of the foregoing embodiments.

In a sixth aspect, there is provided a computer program product comprising instructions which, when run on a computer, cause the computer to perform the methods of the above aspects.

Description of drawings

In order to more clearly illustrate the technical solutions of the various embodiments disclosed in the present application, the following briefly introduces the accompanying drawings used in the description of the embodiments. Obviously, the accompanying drawings in the following description are only disclosed in the present application. For various embodiments, for those of ordinary skill in the art, other drawings can also be obtained according to these drawings without any creative effort.

Figure 1 is a schematic diagram of the navigation route scene from A->B obtained by an unmanned vehicle from a high-precision map in the first scheme;

2 is a flowchart of a method for planning a vehicle driving scheme provided by an embodiment of the present application;

3 is a flowchart of a trajectory cost assessment using time as a baseline for a method for planning a vehicle driving scheme provided by an embodiment of the present application;

4 is a flowchart of constructing a road traffic model of a method for planning a vehicle driving scheme provided by an embodiment of the present application;

5 is a Time-Cost curve diagram of traveling from a starting point in a method for planning a vehicle driving scheme provided by an embodiment of the present application;

FIG. 6 is a scene diagram of traveling from a starting point in a planning method of a vehicle driving scheme provided by an embodiment of the present application, and dividing a fixed route from the starting point to the ending point according to the traffic conditions;

7 is a functional structural diagram of an apparatus for planning a vehicle driving scheme provided by an embodiment of the present application;

FIG. 8 is a schematic diagram of an electronic device for planning a vehicle driving scheme provided by an embodiment of the present application.

Detailed ways

The high-precision map is an electronic map with high accuracy specially provided for automatic driving, with an accuracy of centimeter level, and usually the accuracy needs to be at the lane line level. Common high-precision map data information includes lane lines, center lines, traffic lights, road signs, stop lines, and lane line types.

In a possible solution, the planning module of the control device obtains the navigation route from the starting point to the ending point based on the high-precision map. Then, according to the navigation route from the starting point to the ending point, the trajectory waypoint is generated on the high-precision map, and one or more traffic routes are obtained. The point set of waypoints generated when the unmanned vehicle performs navigation planning and path planning, which consists of multiple waypoints. These waypoints can be traffic signs or moving obstacles in front of the vehicle. Each waypoint represents A real road coordinate, which links the waypoints in the point set one by one to form one or more travel routes. Wherein, the start point, end point, destination and waypoint on the navigation route may also be included in the point set of waypoints.

The planning module can perform cost evaluation on the moving trajectory based on various constraints, so that the moving trajectory of the autonomous driving of the unmanned vehicle can be planned according to the cost evaluation result. The constraints may include collision-related constraints, comfort-related constraints, speed-limit-related constraints, and legal compliance-related constraints. The trajectory cost of all constraints on a driving trajectory is accumulated, and the obtained value is the comprehensive cost of the driving trajectory. Sort different driving trajectories according to the value of the comprehensive cost, and select the driving trajectory corresponding to the minimum comprehensive cost as the final short-term optimal driving trajectory according to the sorting.

The trajectory cost corresponding to each constraint condition can be set separately. For example, the trajectory cost of collision-related constraints is Cost1. The larger the value of Cost1, the greater the probability of collision with obstacles on the driving trajectory. For example, Cost1=6 indicates the probability of collision with obstacles on the driving trajectory. If it is larger, Cost1=3 indicates that the probability of collision on the driving trajectory is less than that of the driving trajectory corresponding to Cost1=6; for another example, the trajectory cost of the comfort-related constraint is Cost2, and the larger the Cost2 value, indicates the comfort of the passengers on the driving trajectory. The lower the degree is, for example, Cost2=6 indicates that the transition between the waypoints and the speed change in the driving trajectory make passengers feel extremely uncomfortable, and Cost2=0 indicates that the transition between the waypoints and the speed change in the driving trajectory make the passengers feel uncomfortable. Comfort; for another example, the trajectory cost of speed limit-related constraints is Cost3. The larger the value of Cost3, the more speed-limited driving is included in the action trajectory. For example, Cost3=5 indicates that the driving trajectory includes speed-limited maneuvering. , Cost3=0 indicates that the driving trajectory does not include speed-limited maneuvering; for another example, the trajectory cost of legal compliance related constraints is Cost4. =5 indicates that the driving trajectory includes maneuvering that violates traffic laws, and Cost4=0 indicates that the driving trajectory does not include maneuvering that violates traffic laws. It should be noted that the above descriptions are all examples, and are not intended to be limiting.

In the above scheme, the automatic driving planning scheme of the unmanned vehicle has established the full trajectory and the temporary trajectory for automatic driving through the full planning and the local planning, respectively. First, the unmanned vehicle uses the high-precision map and the current traffic flow information to formulate the entire trajectory from the starting point to the destination point; then, through the sensor data, high-precision map, high-precision positioning and driving trajectory prediction, the current environment of the unmanned vehicle is located. Carry out local planning and formulate the ideal navigation route for the moment.

However, the unmanned vehicle that uses the navigation route planned by this solution to drive automatically, even if it is on the same road, the same starting point and ending point, at different times and under different circumstances, the passenger's ride experience and driving effect may be completely different. different.

Taking a certain road as an example, there are multiple traffic lights on the road, and the unmanned vehicle travels at different speeds at different times. Waypoints vary in time. The partially planned driving trajectory of some unmanned vehicles may be to drive fast all the time, but it happens that every intersection will encounter a red light, and it needs to stop and wait for a few seconds, then start slowly and continue driving. Due to frequent acceleration, deceleration, parking and waiting, it brings an extremely poor ride experience to passengers. The partially planned driving trajectory of some unmanned vehicles may be to drive smoothly all the time, and every intersection just encounters a green light, so that there is no need to stop all the way, and the driving experience is excellent.

The following introduces a method for planning a vehicle driving scheme provided by an embodiment of the present application.

For a navigation route from the starting point to the destination at the current moment, when planning the driving trajectory, the planning requirement is to make the unmanned vehicle obtain the global optimal or better comprehensive travel effect in one or more of the following aspects as much as possible: the fastest Arrival, minimum start and stop, minimum emergency braking, minimum traffic jam, etc., so that passengers in the car can get a sense of driving experience that meets the planning requirements. Among them, the driving experience can be measured according to the average braking acceleration and deceleration, the number of braking accelerations, the turning range, the waiting time, the average speed, the number of avoidance and other indicators.

The embodiment of the present application contemplates a planning method for a vehicle driving scheme, which can evaluate the trajectory cost of the road segment at different times on the basis of historical communication data information on each road segment on the navigation route for unmanned vehicles, and according to the evaluation result Obtain the ideal travel time and ideal arrival time of a waypoint on the road segment; obtain a driving plan that meets the planning requirements according to the ideal travel time and ideal arrival time.

In most scenarios, the ideal arrival time and ideal exit time of a waypoint are at the same time. For example, passengers want to encounter a green light at every intersection, so that they can drive smoothly without stopping all the way. When planning a waypoint, there is no need to stop. At the intersection of traffic lights, the time period when the green light is on is set as the time constraint of the ideal arrival time and ideal exit time of the waypoint. When planning the driving scheme, the speed is solved with time as one of the constraints, and a series of driving schemes that can reach the specific destination on the road section within the specified time are obtained.

The historical traffic data information is a collection of data and behavior of vehicles in the history of a road segment. For example, the vehicle control information, itinerary information, and traffic information in the surrounding environment, traffic participant information, etc. of the unmanned vehicle at the historical traffic time on the road section. Based on this information, it is possible to know when to travel, where there will be no traffic jams, and when to travel. You won't encounter a lot of pedestrians, vehicles, etc.

The unmanned vehicle can continuously collect the historical traffic data information of the unmanned vehicle itself on the navigation route through the sensors and the domain controller of the vehicle, including vehicle control information, such as accelerator, brake, steering wheel, turn signal, etc., and Environmental information related to itinerary, traffic, traffic participants, etc. A set of road traffic models based on time is established based on the above historical traffic data information.

The modeling elements required to build a road traffic model involve the vehicle control information, itinerary information, traffic information in the surrounding environment, traffic participant information, etc. Specifically, the vehicle control information of the unmanned vehicle itself includes the number of emergency brakes, the number of emergency avoidance, the number of large accelerators, the number of large brakes, etc.; the travel information includes the rate of change of trajectory speed and trajectory curvature, average vehicle speed and arrival time, etc.; the surrounding environment The traffic information includes the number of occurrences of blocking obstacles and dangerous obstacles, the number of red lights, the number of green lights and the number of yellow lights, etc.; the traffic participant information includes the number of occurrences and waiting times of other passing objects such as vehicles and pedestrians.

Among them, the emergency braking is a sudden braking phenomenon caused by the sudden braking of the vehicle in order to avoid traffic participants, obstacles, etc. during the driving process. Emergency avoidance is the phenomenon of sharp turning caused by the sudden turning of the steering wheel in order to avoid traffic participants, obstacles, etc. during the driving process. The blocking obstacle is the obstacle that causes the unmanned vehicle to brake and stop, and is called the blocking obstacle of the unmanned vehicle. Dangerous obstacles are obstacles that are too close to the unmanned vehicle, or the trajectory intersects with the unmanned vehicle, and may cause the unmanned vehicle to take emergency takeover, braking, and evasion.

In the embodiment of the present application, the road traffic model describes the traffic information of a road segment at different times in a time period, and the traffic information may include one or more traffic characteristics of the current road segment, and the traffic characteristics may be expressed as time changes Therefore, the evaluation score of the trajectory cost corresponding to the traffic feature varies with time. After the planning scheme of a road segment is completed, the next waypoint can be used as the destination, and the road traffic model of the next road segment can be used to evaluate the trajectory cost at different times, and calculate the ideal travel time and ideal arrival time of the next waypoint. Plan the driving plan of the unmanned vehicle. This cycle continues until the end point, and a driving plan that meets the planning requirements throughout the navigation route is obtained.

An unmanned vehicle that uses the travel plan planned by the evaluation result of the road traffic model can reach a specific destination within the planned time and obtain a better driving experience.

Before executing the planning method of the vehicle driving scheme provided by the embodiment of the present application, the control device of the unmanned vehicle also establishes a road traffic model and an evaluation system of the road traffic model according to historical traffic data information.

The method is described in detail below.

FIG. 2 is a flowchart of a method for planning a vehicle driving scheme provided by an embodiment of the present application. As shown in FIG. 2 , the control device of the unmanned vehicle performs the following steps to plan the driving scheme:

S21: Acquire a navigation route from a starting point to a destination at a specified time, where the navigation route includes one or more road segment units, and each road segment unit is a road segment between two waypoints on the navigation route.

Specifically, the control device of the unmanned vehicle can obtain the navigation route from the starting point to the destination at the current moment from the high-precision map, and divide the navigation route into multiple road segments. The road segment between every two waypoints is recorded as a road segment unit.

In a possible implementation, the navigation route from the starting point to the destination can be divided into multiple road segments according to the length of the road segment. Correspondingly, the two waypoints are the starting point and the end point of the two ends of the road segment. The 200-meter section length divides the navigation route.

Optionally, the navigation route from the starting point to the destination may be divided into a plurality of road segments according to traffic flow or traffic elements. For example, the navigation route can be divided into multiple road segments according to the waypoints before and after the traffic light intersection. Correspondingly, the two waypoints can be the locations of the traffic light intersections at both ends of the road segment, or according to the waypoints before and after the road segment where traffic congestion often occurs. The waypoint divides the navigation route into multiple sections. Accordingly, the two waypoints can be the starting point and the end point at both ends of the congested section; or the navigation route is divided into multiple sections according to the waypoints before and after the intersection. The two waypoints can be where the intersections at both ends of the road segment are located.

S22: Acquire historical traffic data information on the current navigation route.

Specifically, the control device of the unmanned vehicle collects historical traffic data information based on time on each road segment of the current navigation route.

Exemplarily, the historical traffic data information at a certain moment on each road segment may include the vehicle control information and travel information of the unmanned vehicle passing on the road segment; it may include the traffic information of other vehicles participating in the traffic, for example, Driving speed, quantity, etc., the traffic information of other vehicles is recorded as the traffic participant information, and it can also include the traffic information of the road section at the moment, for example, the duration of traffic lights, the duration of congestion, etc.

In a possible implementation, the vehicle control information and itinerary information at a certain time can be collected through the sensors of the unmanned vehicle and the vehicle domain controller; or all traffic at a certain time can be obtained through the intelligent network connection or through the data stored in the cloud Participants' information and traffic information at a certain time.

It should be understood that the historical traffic data information based on time includes the historical traffic data information of multiple moments within a certain set period of time.

S23 , based on the historical traffic data information, perform a trajectory cost evaluation on the road traffic model of each road section with time as a baseline, and obtain an evaluation result that meets the planning requirements. As shown in FIG. 3 , S23 is specifically implemented by executing the following steps S231-S233.

S231, the road traffic model of the unmanned vehicle is determined as:

Formula (1) Key _i→j is the road traffic model of the i-th road segment to the j-th road segment, Key _n is the road traffic model of the n-th road segment, and key _nm is the m-th traffic feature on the n-th road segment. i, j and m are natural numbers.

S232, according to the road traffic model evaluation system, sum up the trajectory cost scores corresponding to multiple traffic features of the current road traffic model with time as the baseline (the weight is 1), and obtain a comprehensive combination of each road traffic model of each road segment Trajectory cost value.

The evaluation formula corresponding to the road traffic model of the unmanned vehicle in formula (1) is:

In formula (2), Cost _i→j is the comprehensive trajectory cost of the navigation route from the i-th road segment to the j-th road segment, and Cost _i is the trajectory cost of the road segment starting from the i-th waypoint. W ₁ represents the weight value of the traffic feature key _n1 , cost _nt1 represents the trajectory cost of the traffic feature key _n1 of the n-th road segment at time t, W ₂ represents the weight value of the traffic feature key _n2 , and cost _it2 represents the traffic of the n-th road segment at time t The trajectory cost of the feature key _n2 , W _m represents the weight value of the passing feature key _nm , and cost _ntm represents the trajectory cost of the passing feature key _nm of the nth road section at time t. Formula (2) shows that the comprehensive trajectory cost of the navigation route from the ith waypoint to the jth waypoint is the accumulation of the trajectory cost of the road traffic model corresponding to the ith road segment to the jth road segment; The value of the comprehensive trajectory cost between the road segment and the jth road segment is also the value of the weighted summation of the trajectory cost scores corresponding to all the traffic features on the route.

In a possible implementation, according to the road traffic model evaluation system, with time as the baseline, the vehicle control characteristics, travel characteristics, traffic participant characteristics and the trajectory cost scores corresponding to the traffic characteristics at different times on the current road segment are calculated as The time as the baseline is weighted and summed to obtain the comprehensive trajectory cost of the road traffic model of each road segment with the time as the baseline.

In a possible implementation manner, the statistical threshold corresponding to the arrival time feature, the statistical threshold corresponding to the traffic light waiting time feature, and the statistical threshold corresponding to the green light passing time feature can be obtained according to historical traffic data statistics; according to the road traffic model evaluation system, Taking time t as the baseline, the trajectory cost scores corresponding to the statistical thresholds of the arrival time characteristics, traffic light waiting time characteristics and green light passing time characteristics of each road segment at time t are weighted and summed to obtain the comprehensive trajectory cost of the road traffic model.

S233: Evaluate the comprehensive trajectory cost value at different times on the current road section with the time as the baseline, and obtain an evaluation result that meets the planning requirements.

The planning requirement is to make the unmanned vehicle obtain the overall optimal or better comprehensive travel effect in one or more of the following aspects as far as possible: the fastest arrival, the least start and stop, the least sudden braking, the least traffic jam, etc. A driving experience that meets the planning requirements can be obtained.

Whether it meets the planning requirements can be assessed by the score of the comprehensive trajectory cost of the road traffic model.

In a possible implementation manner, the comprehensive trajectory cost values of the road traffic models on the current road section at different times can be compared, and the time t corresponding to the larger comprehensive trajectory cost value is used as the evaluation result that meets the planning requirements.

For example, in the evaluation system of the road traffic model, the trajectory cost score of each traffic feature is set based on the principle that the larger the trajectory cost score is, the more it meets the planning requirements, then the time t corresponding to the larger comprehensive trajectory cost value is the one that meets the planning requirements. Evaluation results of planning requirements.

In a possible implementation manner, the comprehensive trajectory cost values of the road traffic models on the current road section at different times can be compared, and the time t corresponding to the smaller comprehensive trajectory cost value is used as the evaluation result that meets the planning requirements.

For example: in the evaluation system of the road traffic model, the trajectory cost score of each traffic feature is set based on the principle that the smaller the trajectory cost score, the more in line with the planning requirements, then the time t corresponding to the smaller comprehensive trajectory cost value is consistent with the planning requirements. Evaluation results of planning requirements.

S24, obtaining the travel time and arrival time of each road segment unit from the starting waypoint to the ending waypoint according to the evaluation result that meets the planning requirements;

Specifically, the control device of the unmanned vehicle combines the current location and the current travel time, and selects the time corresponding to the evaluation result that meets the planning requirements as the ideal travel time for the current road point. At the same time, combined with the current planning of the unmanned vehicle and the speed limit of the road, the ideal arrival time to the next waypoint is obtained by planning.

In a possible implementation manner, the time information corresponding to the road traffic model can be obtained according to the evaluation results that meet the planning requirements, and the travel time of the starting waypoint of each road segment can be determined; according to the starting waypoint of each road segment According to the speed limit of the road, specify the arrival time of the end waypoint; obtain the travel time and arrival time from the start waypoint to the end waypoint.

In a possible implementation manner, the time information corresponding to the road traffic model is obtained according to the evaluation result that meets the planning requirements, and the travel time of the starting waypoint of each road section is determined; The travel time of the road point, combined with the road speed limit, calculate the interval of the arrival time of the terminal road point of each road segment; according to the interval of the arrival time of the terminal road point, take the terminal road point of the next road segment of each road segment For the destination, plan the travel time of the end waypoint; the travel time of the end waypoint is the arrival time of each road segment, and obtain the travel time and arrival time of each road segment from the starting waypoint to the end waypoint.

S25 , according to the travel time and the arrival time respectively corresponding to the two waypoints in each road segment unit, determine a travel plan on the navigation route that meets the planning requirements. Specifically, it is realized by executing the following steps S251-S252.

S251 , calculating according to the travel time and the arrival time of the current road point by using the constraints among the position, speed and time to obtain the travel speed of the vehicle on the current road section.

The corresponding speed of the trajectory waypoint is planned according to the travel time of the current waypoint and the ideal arrival time of the next waypoint.

Specifically, the ideal arrival time is used as an input parameter and input into the speed optimization formula. When generating the trajectory, a reasonable corresponding speed of the trajectory waypoint is planned to achieve the goal of reaching the target location at the ideal arrival time.

The speed optimization formula is:

represents the acceleration bias weight;

Represents the planned arrival time when traveling from the i-th waypoint.

The constraints for setting formula (3) include:

Set the constraint of the position s of the starting point, which is derived from the planning starting point;

Set waypoint trip _si , speed

acceleration

the upper and lower limits of ;

Set waypoint trip _si , speed

acceleration

The relationship between the constraints is:

Among them, Δt is the difference between the travel time and the arrival time planned from the ith waypoint;

set speed

acceleration

relationship constraints;

set acceleration

and the speed of acceleration change

The relational constraints are:

The constraints of formulas (4)-(6) can make the value of s at each moment as close as possible to the position of the waypoint corresponding to the previously planned driving trajectory. And it is ensured that the value of the acceleration and the speed of the acceleration change are as small as possible, and the driving comfort is ensured under the condition of the minimum arrival time error. In these constraints, trade-offs are made, and the optimization result of the driving speed is finally obtained.

S252: Plan on the navigation route according to the travel time, arrival time and travel speed of each road segment unit, and obtain a travel plan that meets the planning requirements.

Specifically, the control device of the unmanned vehicle determines whether the end point is reached, and if the determination result is "No", the trajectory planning of the next road segment is performed, and S24 is executed. If the judgment result is "Yes", plan the navigation route according to the travel time, arrival time and driving speed of each road section, obtain a driving plan that meets the planning requirements, and complete the driving trajectory planning of this unmanned vehicle.

The method for planning the driving trajectory of an unmanned vehicle provided by the embodiment of the present application is based on the road traffic model of the unmanned vehicle, optimizes the travel time of the navigation route, and uses the time as a constraint condition to participate in the planning of the driving trajectory and the driving speed, so as to achieve the position-specific , speed and time triple constraints and planning. This method makes use of current and historical traffic experience to make the trajectory planning of the unmanned vehicle globally optimal, and the driving experience and traffic efficiency are greatly improved.

Before executing the planning method of the vehicle driving scheme provided by the embodiment of the present application, a road traffic model is also established according to the historical traffic data information.

Specifically, after collecting historical traffic data information and selecting traffic characteristic keys and thresholds, the control device of the unmanned vehicle selects a modeling formula to establish a series of time baseline road traffic models for the current road segment. Exemplarily, you can choose to model the cumulative sum of multiple traffic features to obtain a road traffic model as:

In formula (7), Key _all is the road traffic model of the planned route, key _n is the road traffic model of the nth road segment, and length is the number of road segments obtained by segmenting the entire road.

During modeling, a number of different traffic features can be extracted for modeling through historical traffic data information.

Exemplarily, the corresponding vehicle control features and itinerary features can be extracted by collecting the historical traffic data information of the vehicle itself on a certain planned route, or the traffic participant features and traffic features can be extracted from the data stored in the intelligent network connection and the cloud. .

In a possible implementation manner, as shown in FIG. 4 , establishing a road traffic model according to the historical traffic data information includes the following steps:

S41 , extracting one or more traffic features on each road segment from the historical traffic data information using time as a baseline.

Optionally, the traffic features on each road section unit are extracted from the data information of historical traffic experience based on time; traffic features include vehicle control features, travel features, traffic participant features, traffic features, arrival time features, traffic lights Waiting time characteristics and green light transit time characteristics.

S42: Accumulate one or more traffic features on each road segment unit to establish multiple road traffic models on each road segment with time as a baseline.

Optionally, with time as the baseline, select one or more traffic features from the vehicle control features, travel features, traffic participant features, traffic features, arrival time features, traffic light waiting time features, and green light transit time features on each road segment unit. The features are used as modeling elements to accumulate traffic features to generate a time-based road traffic model.

Optionally, the road traffic model can be obtained by modeling using recurrent neural networks (RNN).

According to an embodiment of the present application, a method for planning a vehicle driving scheme is provided, wherein the road traffic model adopts a plurality of different traffic feature modeling, and the data corresponding to the plurality of different traffic features covers comprehensive historical traffic data information, which can Ensure the accuracy of the road traffic model; evaluate and plan the driving plan according to the accurate road traffic model, so that the unmanned vehicle can achieve a good traffic effect when driving, so that the passengers traveling by the unmanned vehicle can obtain a satisfactory driving experience.

Before executing the planning method of the vehicle driving scheme provided by the embodiment of the present application, an evaluation system of the road traffic model is also established according to the historical traffic data information.

A road traffic model evaluation system can be established by formulating trajectory cost scoring criteria for traffic characteristics and thresholds. According to a number of different traffic characteristics and thresholds, the trajectory cost scoring criteria for traffic characteristics and thresholds are formulated.

Before making the road traffic model, according to the data corresponding to a plurality of different traffic characteristics combined with the actual law, the traffic characteristics required by the road traffic model of the road section and the statistical threshold corresponding to the traffic characteristics are selected.

After collecting historical traffic data information and selecting traffic characteristics and thresholds, a modeling formula is selected to establish a series of time baseline road traffic models with the current road point as the starting point. Exemplarily, it is possible to choose to evaluate the sum of the trajectory costs corresponding to the traffic features, and the evaluation formula corresponding to the road traffic model in formula (7) is:

In formula (8), Cost _all is the comprehensive trajectory cost of the planned route, Cost _n is the trajectory cost of the nth road segment, and length is the number of road segments obtained by segmenting the entire road.

In a possible implementation, one or more traffic features can be extracted from historical traffic data information; the trajectory cost score corresponding to one or more traffic features is set with time as the baseline, and an evaluation system of the road traffic model is established .

When establishing the evaluation system of the road traffic model, the trajectory cost score of each traffic feature is set based on the principle that the smaller the trajectory cost score, the more in line with the planning requirements. Or the larger the trajectory cost score is, the more in line with the planning requirements, the trajectory cost score of each traffic feature is set based on the principle.

In a possible implementation, traffic features can be extracted from historical traffic data information, and the traffic features include vehicle control features, travel features, traffic participant features and traffic features, arrival time features, and traffic light waiting time features of unmanned vehicles. and green light transit time characteristics. Based on the time as the baseline, set the vehicle control characteristics, travel characteristics, traffic participant characteristics and traffic characteristics, arrival time characteristics, traffic light waiting time characteristics and green light passing time characteristics corresponding to the trajectory cost scores at different times, and establish the evaluation of the road traffic model. system.

For example, combine the historical traffic data information to analyze the threshold of arrival time, and obtain the threshold of the traffic feature of the arrival time feature threshold1=5min. In the planning of the trajectory cost corresponding to the traffic feature, two points will be added if the trajectory cost exceeds threshold1, and no points will be added if it is lower than threshold1. . .

Combine the historical traffic data information to analyze the threshold value of the traffic light waiting time feature, and obtain the traffic light waiting time threshold value threshold2=2min. In the planning of the trajectory cost corresponding to the traffic feature, 3 points are added for exceeding the threshold2, and no points are added below the threshold2. .

Combine the historical traffic data information to analyze the threshold value of the green light traffic time feature, and obtain the threshold value of the traffic feature of the green light traffic time. Threshold 3 = 7min. In the planning of the trajectory cost corresponding to the traffic feature, 2 points are added for exceeding the threshold3, and less than the threshold3 is not added. point.

The embodiments of the present application provide a planning method for a vehicle driving scheme, make a road traffic model of the unmanned vehicle, and plan the driving trajectory of the unmanned vehicle. Specific steps are as follows:

S501, select a relevant modeling element as a traffic feature (key) of a road traffic model of the unmanned vehicle, set a corresponding evaluation score (Value) of a trajectory cost (cost) for the traffic feature, and set a value for some specific building blocks The pass characteristic of the modulo element sets the relevant threshold (threshold). As shown in Table 1, the modeling elements involved in the modeling of the road traffic model of unmanned vehicles include vehicle control characteristics and environmental characteristics.

Table 1

For example, the traffic features of the vehicle control features in Table 1 include "emergency avoidance", and the corresponding evaluation score of the trajectory cost is "+5"; during local planning, if the unmanned vehicle is currently known according to the historical traffic data information If an emergency avoidance has occurred at the roadpoint once, the trajectory cost of the road traffic model of the current road section will be increased by 5 points. If the unmanned vehicle makes multiple emergency avoidance at the roadpoint, the trajectory cost of the road traffic model of the road section will be increased by 5 points. *m points, m is the number of times of emergency avoidance.

The traffic features of the vehicle control features in Table 1 include the "arrival time" feature, and the corresponding evaluation score is ">threshold+2"; during local planning, if the unmanned vehicle is calculated according to the historical traffic data information and the planned speed information When the time to reach the current target waypoint exceeds the threshold (threshold), 2 points are added to the trajectory cost cost of the road traffic model of the current road segment. For example, set the threshold of arrival time to 2 minutes. If the time for the unmanned vehicle to reach the current target waypoint at the current speed exceeds the specified arrival time by more than 2 minutes, the trajectory cost of the current road traffic model will add 2 points. , no additional points are added for trajectory costs below the threshold.

The traffic features of the environmental information in Table 1 include the "red light" feature, and its corresponding evaluation score is "+3"; in local planning, if the unmanned vehicle encounters a red light before reaching the target waypoint, then The trajectory cost of the road traffic model adds 3 points. If the unmanned vehicle encounters multiple red lights before reaching the target waypoint, the cost of the trajectory cost of the road traffic model of this road segment is added by 3*m points, where m is the number of red lights.

Table 1 also shows the traffic characteristics of other vehicle control information and environmental information and their evaluation scores. The evaluation scores of the specific trajectory cost are similar to the above examples, and will not be listed one by one.

In Table 1, traffic characteristics of other related elements and their evaluation scores may be added according to the planning scheme. The cost calculation algorithm of the specific road traffic model is similar to the above example, or those skilled in the art can obtain the above examples based on the common knowledge. The trajectory cost algorithm corresponding to the road traffic model is also within the protection scope of the present application.

S502 , according to the traffic characteristics of the above-mentioned relevant modeling elements and the evaluation score of the corresponding trajectory cost, and in combination with the modeling equation of formula (6), a simple road traffic model of the unmanned vehicle is established. Exemplarily, the road traffic model is modeled by using the traffic time feature and the waiting time feature as the modeling element values, and the current road traffic model of the unmanned vehicle is evaluated as:

In formula (9), Cost _i→j is the comprehensive trajectory cost of the planned route from the i-th road segment to the j-th road segment, and Cost _n is the trajectory cost corresponding to the n-th road segment. W1 represents the weight value of the transit time feature, Cost _nt1 represents the trajectory cost of the transit time of the nth road segment at time t, _W2 represents the weight value of the waiting time feature, and Cost _nt2 represents the trajectory corresponding to the waiting time feature of the nth road _segment at time t Cost, wait time features include red light wait time and/or congestion wait time. Formula (9) shows the road traffic model evaluation formula of the route between the i-th road point to the j-th road point, which is the accumulation of the trajectory cost of the road traffic model of each road segment corresponding to the i-th road segment to the j-th road segment. ; The value of the comprehensive trajectory cost of the route between the i-th road segment to the j-th road segment is the accumulation of the trajectory cost of each road segment corresponding to the i-th road segment to the j-th road segment; Similarly, the i-th road segment to the j-th road segment The value of the comprehensive trajectory cost of the route between them is also the value of the weighted summation of the trajectory cost scores corresponding to the travel time feature and the waiting time feature on the planned route.

S503, modeling is performed according to 7:50 in the morning to 8:50 in the morning, and the trajectory cost Cost _i→j of the road traffic model from the ith road segment to the j th road segment at different travel times is calculated according to formula (9). According to the historical traffic data information, it is known that 8:00 in the morning – 8:30 in the morning is the morning peak with serious traffic congestion, the weight value W ₂ of the waiting time feature is larger, and the waiting time trajectory cost Cost due to serious traffic congestion _nt2 is larger.

FIG. 5 is a Time-Cost curve diagram of traveling from the starting point of a certain planned route. As shown in FIG. 5 , the horizontal axis is the time axis (Time), and the vertical axis is the trajectory cost axis (Cost). When the travel time is 7:50, the trajectory cost Cost _i→j of the road traffic model is calculated as 15; when the travel time is 8:00, the trajectory cost Cost _i→j of the road traffic model is calculated as 20; when the travel time is At 8:10, calculate the trajectory cost of the road traffic model Cost _i→j is 32; when the travel time is 8:20, calculate the trajectory cost of the road traffic model Cost _i→j is 46; when the travel time is 8:30 , Calculate the trajectory cost of the road traffic model Cost _i→j is 35; when the travel time is 8:40, calculate the trajectory cost of the road traffic model Cost _i→j is 20; when the travel time is 8:50, calculate the road traffic The trajectory cost Cost _i→j of the model is 18.

Analysis of the Time-Cost curve shown in Figure 5 shows that when the travel time is 7:50, the trajectory cost Cost _i→j of the road traffic model is the smallest, and the riding experience and driving effect of passengers are the best; when the travel time is At 8:20, the trajectory cost Cost _i→j of the road traffic model is the largest, and the riding experience and driving effect of passengers are the worst.

S504, it is determined that the unmanned vehicle user starts to travel from the starting point at 7:50, and the fixed route from the starting point to the ending point is divided into three sections according to the traffic conditions. As shown in FIG. 6 , the three road segments are respectively the starting point->D1 of waypoint 1, D2 of waypoint 1->waypoint 2, and D3 of waypoint 2->end point.

Specifically, waypoint 1 is before the traffic light intersection, D1 represents the first road segment from the starting point to the traffic light intersection; waypoint 2 is after the traffic light intersection, D2 represents the second road segment from the traffic light intersection to the back of the traffic light intersection; D3 represents from the traffic light intersection The third section after the intersection to the end.

S505, take 7:50 as the starting point for travel time, calculate according to the speed limit of the road section and the trajectory planning of the unmanned vehicle, and obtain the time range of arriving at road point 1 in the range of 8:10-8:30, and in the time range of 8:10-8:30. Within the range, the trajectory cost Cost ₁ of D1 is calculated according to the road traffic model of the unmanned vehicle, and the ideal travel time point of road point 1 is 8:20. At this time, the green light can be encountered, so that the vehicle can drive smoothly without stopping; The time 8:20 is used as the constraint condition of speed trajectory optimization, and is brought into formulas (3)-(5) to solve the speed, and a path starting at 7:50 and arriving at waypoint 1 at 8:20 meets the planning requirements driving plan.

S506, then travel along the route of waypoint 1->waypoint 2, and continue to cyclically solve the ideal arrival time and solve the speed according to the road traffic model of the unmanned vehicle.

Specifically, it is judged whether the unmanned vehicle has reached the end point, and if the judgment result is "No", take 8:20 as the travel time of waypoint 1, and calculate according to the road speed limit and the trajectory planning of the unmanned vehicle, and obtain the arrival time of waypoint 2. The time range is 8:25-8:35. According to the road traffic model of the unmanned vehicle, the trajectory cost Cost _{2 of the D2 road section is calculated. According to the minimum value of the trajectory cost Cost 2} _, the travel time point corresponding to the ideal road point 2 is obtained as At 8:30, it can drive smoothly without stopping; take the arrival time of 8:20 as the constraint condition of speed trajectory optimization, bring it into formulas (3)-(5) to solve the speed, and get the arrival time at 8:30. 2. The driving scheme that meets the planning requirements.

S507 , repeating S506 to perform trajectory planning for the next road section, until it is finally determined that the unmanned vehicle has reached the end point, and the trajectory planning is canceled. Complete the planning of this unmanned vehicle driving plan.

The above-mentioned embodiments plan the driving scheme according to the road traffic model based on the unmanned vehicle. When the unmanned vehicle is planned based on the model, the model data can be fully utilized to avoid the waiting time of the red light and the waiting of the congested road section, so as to improve the traffic efficiency and improve the traffic efficiency. driving experience.

Embodiments of the present application also provide a device for planning a vehicle driving scheme, the device can be deployed or integrated on a vehicle, is a part of an on-board system, and can be an on-board control unit, such as an ECU, DCU, or MDC, etc., or a Semiconductor chips installed in in-vehicle systems, etc.

As shown in FIG. 7 , the device obtains the navigation route from the starting point to the destination at the specified time through the route determination module 71, the navigation route includes one or more road segment units, and each road segment unit in the one or more road segment units is two respectively The road section between the waypoints; the historical traffic data information of the navigation route is acquired by the data acquisition module 72; the road traffic model of each road section unit in the one or more road section units is determined by the model evaluation module 73 based on the historical traffic data information with time as The trajectory cost evaluation is performed on the baseline to obtain an evaluation result that meets the planning requirements; the time planning module 74 determines the travel time and arrival time corresponding to the two waypoints of each road segment unit according to the evaluation results that meet the planning requirements; and the solution planning module 75 According to the travel time and arrival time respectively corresponding to the two waypoints in each road segment unit, the driving plan on the navigation route that meets the planning requirements is determined.

In an apparatus for planning a vehicle driving scheme provided by an embodiment of the present application, the time planning module includes: a speed calculation unit and a scheme planning unit; the device uses the speed calculation unit according to the travel times corresponding to the two road points of each road section unit respectively and the arrival time to determine the speed of the vehicle on each road segment unit; the solution planning unit determines the navigation based on the travel time and arrival time corresponding to the two waypoints in each road segment unit and the vehicle's driving speed on each road segment unit. The driving plan on the route that meets the planning requirements.

Optionally, the historical traffic data information of the navigation route includes one or more of the following: the vehicle is on one or more road segment units, and the vehicle control information, itinerary information, and vehicle traffic at multiple different times in a historical time period Information about surrounding traffic participants and traffic information.

In an apparatus for planning a vehicle driving scheme provided in an embodiment of the present application, one or more road segment units are obtained by dividing the navigation route according to the length of the road segment, or one or more road segment units are obtained by dividing the navigation route according to traffic flow or traffic elements route obtained.

In an apparatus for planning a vehicle driving scheme provided by an embodiment of the present application, a model building module is further included, and a road traffic model of each road segment unit in one or more road segment units is established by the model building module according to the historical traffic data information. .

Optionally, building a model module includes using a feature extraction unit to extract one or more traffic features on each road section unit from historical traffic data information with time as a baseline; or multiple traffic features, and establish multiple road traffic models based on time on each road segment unit.

Optionally, the modeling subunit accumulates one or more traffic features on each road segment unit to establish multiple time-based road traffic models on each road segment unit.

Optionally, the modeling subunit is configured to use one or more traffic features on each road segment unit to establish a time-based road traffic model through an RNN recurrent neural network.

In an apparatus for planning a vehicle driving scheme provided by an embodiment of the present application, the apparatus further includes an evaluation system establishment module, and the evaluation system establishment module establishes an evaluation system of a road traffic model according to the historical traffic data information. The evaluation system includes trajectory cost scores corresponding to one or more traffic features at different times; set the trajectory cost scores corresponding to one or more traffic features at different times to establish an evaluation system for the road traffic model.

Optionally, the traffic characteristics include one or more of the following: vehicle control characteristics, travel characteristics, traffic participant characteristics, traffic characteristics, arrival time characteristics, traffic light waiting time characteristics, and green light passing time characteristics.

Optionally, the model evaluation module includes a calculation unit and an evaluation unit; the device uses the calculation unit to calculate the trajectory cost score corresponding to a plurality of the traffic features of each road traffic model according to the road traffic model evaluation system as The time is the baseline summation or weighted summation to obtain the comprehensive trajectory cost value of each road traffic model of each road segment unit at different times; and the evaluation unit is used to compare the comprehensive trajectory cost value with the time as the baseline to obtain Evaluation results that meet planning requirements.

In an apparatus for planning a vehicle driving scheme provided in an embodiment of the present application, a transit time planning module of the apparatus obtains time information corresponding to a road passing model of each road segment unit according to an evaluation result that meets the planning requirements, and determines each road segment The travel time of the starting waypoint of the unit; according to the traveling time of the starting waypoint of each road segment unit, combined with the road speed limit, specify the arrival time of the ending waypoint of each road segment unit; obtain from the starting waypoint to the ending waypoint travel time and arrival time.

Optionally, the device obtains the time information corresponding to the road traffic model of each road segment unit through the time planning module according to the evaluation result that meets the planning requirements, and determines the travel time of each road segment unit at the starting waypoint; The travel time of the starting waypoint of the road segment unit, combined with the road speed limit, calculates the interval of the arrival time of the ending waypoint of each road segment unit; The travel time of the next road segment unit of each road segment unit; taking the travel time of the next road segment unit of each road segment unit as the arrival time of each road segment unit, obtain the corresponding road points of each road segment unit respectively Travel time and arrival time.

In an apparatus for planning a vehicle driving scheme provided by an embodiment of the present application, the speed calculation unit is used for:

According to the travel time and arrival time corresponding to the two waypoints in each road segment unit, the speed optimization formula is used to determine the driving speed of the vehicle on each road segment unit. The speed optimization formula is as formula (3):

represents the acceleration bias weight;

Represents the planned arrival time when traveling from the i-th waypoint.

Optionally, the velocity calculation unit is also used to:

Set waypoint trip _si , speed

acceleration

the upper and lower limit values;

Set waypoint trip _si , speed

acceleration

The relationship between the constraints is as formula (4):

set speed

acceleration

The relationship between the constraints is as formula (5):

set acceleration

and the speed of acceleration change

The relational constraints of are as formula (6):

As shown in FIG. 8 , an embodiment of the present application provides an electronic device 1100, including a processor 1101 and a memory 1102; the processor 1101 is configured to execute computer-executed instructions stored in the memory 1102, and the processor 1101 runs The computer executes the instructions to execute the method for unmanned driving planning described in any of the foregoing embodiments.

This embodiment of the present application provides a storage medium 1103, including a readable storage medium and a computer program stored in the readable storage medium, where the computer program is used to implement the unmanned driving planning described in any of the foregoing embodiments. method.

Embodiments of the present application further provide a vehicle, where at least one device for planning a vehicle driving scheme is deployed or integrated in the vehicle, and the device is a part of an in-vehicle system, which may be an in-vehicle control unit, such as an ECU, DCU, or MDC, etc., It may also be a semiconductor chip or the like provided in an in-vehicle system. The vehicle can travel according to the driving plan planned by the device according to the method of any one of the above embodiments; the device includes at least: a route determination module, configured to obtain a navigation route from a starting point to a destination at a specified time, the navigation route including one or A plurality of road section units, each of which is a road section between two road points in one or more road section units; a data acquisition module for acquiring historical traffic data information of a navigation route; a model evaluation module for historically based The traffic data information is used to evaluate the trajectory cost of the road traffic model of each road segment unit in one or more road segment units with time as the baseline, and obtain the evaluation results that meet the planning requirements; the time planning module is used to determine the evaluation results that meet the planning requirements. The travel time and arrival time corresponding to the two waypoints in each road section unit respectively; and a scheme planning module, which is used to determine the travel time on the navigation route that meets the planning requirements according to the travel time and arrival time respectively corresponding to the two waypoints in each road section unit Program.

Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Experts may use different methods for each specific application to implement the described functions, but such implementation should not be considered beyond the scope of the embodiments of the present application.

Furthermore, various aspects or features of the embodiments of the present application may be implemented as a method, apparatus or article of manufacture using standard programming and/or engineering techniques. The term "article of manufacture" as used in this application encompasses a computer program accessible from any computer-readable device, carrier or medium. For example, computer readable media may include, but are not limited to: magnetic storage devices (eg, hard disks, floppy disks, or magnetic tapes, etc.), optical disks (eg, compact discs (CDs), digital versatile discs (DVDs) etc.), smart cards and flash memory devices (eg, erasable programmable read-only memory (EPROM), card, stick or key drives, etc.). Additionally, various storage media described herein can represent one or more devices and/or other machine-readable media for storing information. The term "machine-readable medium" may include, but is not limited to, wireless channels and various other media capable of storing, containing, and/or carrying instructions and/or data.

It should be understood that, in various embodiments of the embodiments of the present application, the size of the sequence numbers of the above-mentioned processes does not mean the sequence of execution, and the execution sequence of each process should be determined by its functions and internal logic, and should not be The implementation process of the embodiments of the present application constitutes any limitation.

Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described systems, devices and units may refer to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

In the several embodiments provided in this application, it should be understood that the disclosed system, apparatus and method may be implemented in other manners. For example, the apparatus embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or Can be integrated into another system, or some features can be ignored, or not implemented. On the other hand, the shown or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, indirect coupling or communication connection of devices or units, and may be in electrical, mechanical or other forms.

The units described as separate components may or may not be physically separated, and components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution in this embodiment.

The functions, if implemented in the form of software functional units and sold or used as independent products, may be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present application can be embodied in the form of software products in essence, or the parts that make contributions to the prior art or the parts of the technical solutions, and the computer software products are stored in a storage medium , including several instructions to cause a computer device (which may be a personal computer, a server, or an access network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the embodiments of this application. The aforementioned storage medium includes: U disk, mobile hard disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes .

The above are only specific implementations of the embodiments of the present application, but the protection scope of the embodiments of the present application is not limited thereto. Any changes or substitutions should be included within the protection scope of the embodiments of the present application.

Claims

A method for planning a vehicle driving scheme, characterized in that the method comprises:

Obtain a navigation route from a starting point to a destination at a specified time, the navigation route includes one or more road segment units, and each road segment unit in the one or more road segment units is respectively between two waypoints on the navigation route section of the road;

Obtain historical traffic data information of the navigation route;

Based on the historical traffic data information, a trajectory cost evaluation is performed on the road traffic model of each of the one or more road segment units with time as the baseline, and an evaluation result that meets the planning requirements is obtained;

Determine the travel time and the arrival time corresponding to the two waypoints of each road segment unit according to the evaluation result that meets the planning requirements;

According to the travel time and the arrival time respectively corresponding to the two waypoints in each road segment unit, a driving plan on the navigation route that meets the planning requirement is determined.
The method according to claim 1, characterized in that, determining a travel plan on the navigation route that meets the planning requirements according to travel times and arrival times corresponding to two waypoints in each road segment unit, comprising: :

determining the travel speed of the vehicle on each road segment unit according to the travel time and the arrival time corresponding to the two waypoints in each road segment unit respectively;

According to the travel time and arrival time respectively corresponding to the two waypoints in each road segment unit and the driving speed of the vehicle on each road segment unit, a driving plan on the navigation route that meets the planning requirement is determined.
The method according to claim 1, wherein the historical traffic data information of the navigation route includes one or more of the following:

On the one or more road segment units, the vehicle control information, travel information, and traffic participant information and traffic information around the vehicle at multiple different times in a historical time period.
The method according to claim 1, wherein the one or more road segment units are obtained by dividing the navigation route according to the length of the road segment, or the one or more road segment units are obtained by dividing the navigation route according to the traffic flow or traffic element obtained by dividing the navigation route.
The method according to claims 1-3, wherein the method further comprises establishing a road traffic model of each of the one or more road segment units according to the historical traffic data information.
The method according to claim 5, wherein the establishing a road traffic model of each of the one or more road segment units according to the historical traffic data information comprises:

extracting one or more traffic features on each road segment unit from the historical traffic data information using time as a baseline;

According to one or more traffic characteristics on each road segment unit, a plurality of time-based road traffic models on each road segment unit are established.
The method according to claim 6, wherein, according to one or more traffic characteristics on each road segment unit, establishing a plurality of time-based road traffic models on each road segment unit , which includes: accumulating one or more traffic characteristics on each road segment unit to establish multiple road traffic models on each road segment unit with time as the baseline.
The method according to claim 6, wherein, according to one or more traffic characteristics on each road segment unit, establishing a plurality of time-based road traffic models on each road segment unit ,include:

One or more traffic features on each road segment unit are used to establish a time-based road traffic model through an RNN cyclic neural network.
The method according to any one of claims 6-8, wherein the method further comprises: establishing an evaluation system of the road traffic model according to the historical traffic data information, and the evaluation system of the road traffic model includes: The trajectory cost scores corresponding to the one or more traffic features at different times; the trajectory cost scores corresponding to the one or more traffic features at different times are set to establish an evaluation system for the road traffic model.
The method according to any one of claims 6 to 9, wherein the traffic characteristics include one or more of the following: vehicle control characteristics, travel characteristics, traffic participant characteristics, traffic characteristics, arrival time characteristics, traffic lights Waiting time characteristics and green light transit time characteristics.
The method according to claim 9, characterized in that, performing trajectory cost evaluation on the road traffic model of each road segment unit based on historical traffic data information with time as a baseline, and obtaining an evaluation result that meets planning requirements, comprising:

According to the road traffic model evaluation system, the trajectory cost scores corresponding to the plurality of traffic features of each road traffic model are summed or weighted with the time as the baseline, and each road segment unit at different times is obtained. The comprehensive trajectory cost value of the road traffic model;

The comprehensive trajectory cost value is compared with time as the baseline to obtain an evaluation result that meets the planning requirements.
The method according to any one of claims 1-2, wherein the two waypoints of each road segment unit are a starting waypoint and an ending waypoint respectively, and the evaluation result according to the planning requirement Determining the travel time and arrival time respectively corresponding to the two waypoints of each road segment unit, including:

According to the evaluation result that meets the planning requirements, obtain time information corresponding to the road traffic model of each road segment unit, and determine the travel time of the starting waypoint of each road segment unit;

According to the travel time of the starting waypoint, in combination with the road speed limit, the arrival time of the ending waypoint of each road segment unit is designated.
The method according to any one of claims 1-2, wherein the two waypoints of each road segment unit are a starting waypoint and an ending waypoint respectively, and the evaluation result according to the planning requirement Determining the travel time and arrival time respectively corresponding to the two waypoints of each road segment unit, including:

According to the evaluation result that meets the planning requirements, obtain time information corresponding to the road traffic model of each road segment unit, and determine the travel time of each road segment unit at the starting waypoint;

According to the travel time of the starting waypoint of each road segment unit, combined with the road speed limit, calculate the interval of the arrival time of the ending waypoint of each road segment unit;

planning the travel time of the next road segment unit of each road segment unit according to the interval of the arrival time of the terminating waypoint of each road segment unit;

Taking the travel time of the next road segment unit of each road segment unit as the arrival time of each road segment unit, the travel time and the arrival time respectively corresponding to the two waypoints of each road segment unit are obtained.
The method according to claim 2, wherein the driving speed of the vehicle on each road segment unit is determined according to the travel time and the arrival time corresponding to the two waypoints in each road segment unit respectively, include:

According to the travel time and arrival time respectively corresponding to the two waypoints of each road segment unit, the speed optimization formula is used to determine the driving speed of the vehicle on each road segment unit, and the speed optimization formula is:

Among them, f is the optimization result of the optimization function, and the optimization goal is to minimize f when planning the speed of the unmanned vehicle; si represents the travel of the ith actual waypoint;
Represents the i -th planned waypoint trip; ws represents the position deviation weight;
represents the acceleration bias weight;
Indicates the acceleration of the vehicle, and the subscript i is the planned waypoint of the ith;
Indicates the weight of the deviation of the speed value of the vehicle in the acceleration change;
represents the speed of the acceleration change, the subscript i→i+1 is the planned waypoint from the ith to the ith+1; w t represents the arrival time deviation weight; t i represents the travel from the ith waypoint to the destination the estimated time;
Represents the planned arrival time when traveling from the i-th waypoint.
The method according to claim 14, wherein the determining the driving speed of the vehicle on each road segment unit by using the speed optimization formula further comprises:

Set waypoint trip si , speed
acceleration
the upper and lower limit values;

Set waypoint trip si , speed
acceleration
The relationship between the constraints is:

Among them, Δt is the difference between the travel time and the arrival time planned from the ith waypoint;

set speed
acceleration
relationship constraints;

set acceleration
and the speed of acceleration change
The relational constraints are:
A device for planning a vehicle driving scheme, characterized in that the device comprises:

A route determination module is used to obtain a navigation route from a starting point to a destination at a specified time, the navigation route includes one or more road segment units, and each road segment unit in the one or more road segment units is the difference between the two waypoints. the road section between;

a data acquisition module for acquiring historical traffic data information of the navigation route;

A model evaluation module, configured to perform trajectory cost evaluation on the road traffic model of each road segment unit in the one or more road segment units based on the historical traffic data information with time as the baseline, and obtain an evaluation result that meets the planning requirements;

a time planning module, configured to determine the travel time and the arrival time corresponding to the two waypoints of each road segment unit respectively according to the evaluation result that meets the planning requirements; and

A scheme planning module, configured to determine a travel scheme on the navigation route that meets the planning requirements according to the travel time and the arrival time respectively corresponding to the two waypoints in each road segment unit.
The apparatus according to claim 16, wherein the time planning module comprises:

a speed calculation unit, configured to determine the travel speed of the vehicle on each road segment unit according to the travel time and arrival time corresponding to the two waypoints in each road segment unit; and

A scheme planning unit, configured to determine that the navigation route satisfies the requirements according to the travel time and arrival time corresponding to the two waypoints in each road segment unit and the driving speed of the vehicle on each road segment unit. Plan the required driving scenarios.
The device according to claim 16, wherein the historical traffic data information of the navigation route includes one or more of the following:

The vehicle control information, itinerary information, and traffic participant information and traffic information around the vehicle at a plurality of different times in a historical time period on the one or more road section units.
The device according to claim 16, wherein the one or more road segment units are obtained by dividing the navigation route according to the length of the road segment, or the one or more road segment units are obtained by dividing the navigation route according to the traffic flow or traffic element obtained by dividing the navigation route.
The apparatus according to any one of claims 16-18, characterized in that the apparatus further comprises a model building module, configured to: build each road section unit in the one or more road section units according to the historical traffic data information road traffic model.
The apparatus according to claim 20, wherein the building model module comprises:

a feature extraction unit for extracting one or more traffic features on each road segment unit from the historical traffic data information using time as a baseline; and

The modeling unit is configured to establish a plurality of time-based road traffic models on each road segment unit according to one or more traffic characteristics on each road segment unit.
The apparatus according to claim 21, wherein the modeling sub-unit is configured to accumulate one or more traffic characteristics on each road section unit to establish a time-based Baseline multiple road traffic models.
The device according to claim 21, wherein the modeling sub-unit is configured to use one or more traffic features on each road segment unit to establish a time-based road traffic model through an RNN recurrent neural network .
The device according to any one of claims 21-23, characterized in that, the device further comprises an evaluation system establishment module, and the evaluation system establishment module is configured to establish an evaluation of the road traffic model according to the historical traffic data information The evaluation system of the road traffic model includes the trajectory cost scores corresponding to the one or more traffic features at different times; set the trajectory cost scores corresponding to the one or more traffic features at different times, and establish The evaluation system of the road traffic model.
The device according to any one of claims 21-23, wherein the traffic characteristics include one or more of the following: vehicle control characteristics, travel characteristics, traffic participant characteristics, traffic characteristics, arrival time characteristics, traffic light waiting Time characteristics and green light transit time characteristics.
The apparatus according to claim 24, wherein the model evaluation module is used for:

The calculation unit is used to sum or weight the trajectory cost scores corresponding to a plurality of the traffic features of each road traffic model according to the road traffic model evaluation system with the time as the baseline, and obtain each road segment unit at the time. the composite trajectory cost values for each road traffic model at different times; and

An evaluation unit, configured to compare the comprehensive trajectory cost value with time as a baseline to obtain an evaluation result that meets the planning requirements.
The device according to any one of claims 16-17, wherein the time planning module is used for:

According to the evaluation result that meets the planning requirements, obtain time information corresponding to the road traffic model of each road segment unit, and determine the travel time of the starting waypoint of each road segment unit;

According to the travel time of the starting waypoint, in combination with the road speed limit, the arrival time of the ending waypoint of each road segment unit is designated.
The device according to any one of claims 16-17, wherein the two waypoints of each road segment unit are a starting waypoint and an ending waypoint respectively, and the time planning module is used for:

According to the evaluation result that meets the planning requirements, the time information corresponding to the road traffic model of each road segment unit is obtained, and the travel time of each road segment unit at the starting waypoint is determined; The travel time of the starting waypoint of the unit, combined with the road speed limit situation, calculates the interval of the arrival time of the ending waypoint of each road segment unit; according to the interval of the arrival time of the ending waypoint of each road segment unit, Planning the travel time of the next road segment unit of each road segment unit; taking the travel time of the next road segment unit of each road segment unit as the arrival time of each road segment unit, obtain two waypoints for each road segment unit Corresponding travel time and arrival time, respectively.
The device according to claim 17, wherein the speed calculation unit is used for:

Among them, f is the optimization result of the optimization function, and the optimization goal is to minimize f when planning the speed of the unmanned vehicle; si represents the travel of the ith actual waypoint;
Represents the i -th planned waypoint trip; ws represents the position deviation weight;
represents the acceleration bias weight;
Indicates the acceleration of the vehicle, and the subscript i is the planned waypoint of the ith;
Indicates the weight of the deviation of the speed value of the vehicle in the acceleration change;
represents the speed of the acceleration change, the subscript i→i+1 is the planned waypoint from the ith to the ith+1; w t represents the arrival time deviation weight; t i represents the travel from the ith waypoint to the destination the estimated time;
Represents the planned arrival time when traveling from the i-th waypoint.
The device according to claim 29, wherein the speed calculation unit is further configured to:

Set waypoint trip si , speed
acceleration
the upper and lower limit values;

Set waypoint trip si , speed
acceleration
The relationship between the constraints is:

Among them, Δt is the difference between the travel time and the arrival time planned from the ith waypoint;

set speed
acceleration
relationship constraints;

set acceleration
and the speed of acceleration change
The relational constraints are:
An electronic device, comprising a memory and a processor; the processor is configured to execute computer-executable instructions stored in the memory, and the processor executes the computer-executable instructions to execute the plan described in any one of claims 1-15 Method of vehicle driving scheme.
A vehicle, characterized in that the vehicle comprises the device according to any one of claims 16 to 31, the device being used for planning a driving scheme for the vehicle.
A storage medium, comprising a readable storage medium and a computer program stored in the readable storage medium, the computer program being used to implement the method for planning a vehicle driving scheme according to any one of claims 1-15.