阅读说明：本技术 基于数字孪生和用户个性化需求的车辆路径规划方法 (Vehicle path planning method based on digital twin and user personalized requirements ) 是由 惠一龙 汪蔷蔷 承楠 肖潇 尹至胜 于 2021-09-16 设计创作，主要内容包括：本发明公开一种基于数字孪生和用户个性化需求的车辆路径规划方法,其步骤为：1)建立数字孪生体发送路径规划请求；2)计算用户个性化需求的待规划路径效用值；3)建立交通路口的状态和动作空间；4)利用Q-learning算法更新Q表；5)获取最佳驾驶路径；6)云服务器通过基站向车辆用户发送规划的最佳的驾驶路径。本发明能有效提高车流量调度效率,降低了路径规划的网络时延,可用于城市交通系统中车辆用户的路径规划。(The invention discloses a vehicle path planning method based on digital twin and user personalized requirements, which comprises the following steps: 1) establishing a digital twin sending path planning request; 2) calculating the utility value of the path to be planned of the user personalized demand; 3) establishing a state and an action space of a traffic intersection; 4) updating the Q table by using a Q-learning algorithm; 5) obtaining an optimal driving path; 6) and the cloud server sends the planned optimal driving path to the vehicle user through the base station. The invention can effectively improve the traffic flow scheduling efficiency, reduce the network time delay of path planning, and can be used for path planning of vehicle users in an urban traffic system.)

1. A vehicle path planning method based on digital twin and user personalized requirements is characterized in that a digital twin is established to send a path planning request, and a path utility value to be planned of the user personalized requirements is calculated; the method comprises the following steps:

step 1, establishing a digital twin sending path planning request:

a vehicle user to be planned with a path deploys a digital twin DTs on a cloud server and then sends a path planning request;

step 2, calculating the utility value of the path to be planned of the user personalized demand:

(2a) the cloud server adds the starting time of travel in the received path planning request to a service list of the cloud server;

(2b) calculating the driving time of each traffic intersection reached in the path to be planned according to the following formula:

wherein, tau_gRepresenting the phase of arrival from the last traffic intersection in the path to be plannedAdjacent g traffic crossing is at [ -1,1 [)]Time after range normalization, d_gRepresenting the distance from the last traffic crossing in the path to be planned to the adjacent g-th traffic crossing, v_minRepresenting the minimum driving speed of the road section limit from the last traffic intersection to the adjacent g-th traffic intersection in the path to be planned, max representing the maximum value operation, v_maxRepresenting the maximum speed of travel, k, limited by the section of road from the last traffic crossing to the adjacent g-th traffic crossing in the path to be planned_gAnd k_maxRespectively representing the traffic flow density and the maximum traffic flow density from one traffic intersection on the path to be planned to the adjacent g-th traffic intersection;

(2c) according to the following formula, the normalized toll of each adjacent traffic intersection road section in the path to be planned is designed to be collected according to the traffic flow density in the peak period:

wherein p is_gThe method indicates that the next g traffic intersection is reached at [ -1, 1] from the previous traffic intersection in the planned path to be collected]The toll after the range normalization is used, wherein cos represents cosine value taking operation;

(2d) calculating the cost performance of the driving time and the toll of the path to be planned according to the following formula:

wherein alpha represents the cost performance of the driving time and the toll of the path to be planned, and t_{Preparation of}Indicating the estimated time of arrival at the destination, t_minRepresents the shortest time to reach the destination;

(2e) calculating the utility value of each traffic intersection in the path to be planned according to the following formula;

r_g＝ατ_g+(1-α)p_g

wherein r is_gRepresenting the data to be plannedThe utility value of the adjacent g-th traffic intersection reached by the previous traffic intersection in the path;

step 3, establishing the state and action space of the traffic intersection:

taking each traffic intersection of the path to be planned as the state of the Q-learning algorithm from the selected traffic map, establishing the state space of all the traffic intersections, taking the selection of each traffic intersection of the path to be planned on the next adjacent traffic intersection as the action in the Q-learning algorithm, and establishing the action space of all the traffic intersections;

and 4, updating the Q table by using a Q-learning algorithm:

(4a) establishing a Q table, wherein the row of the Q table represents the state of each traffic intersection, the column represents the action selection of each traffic intersection, and the initialized Q value is set to be 0;

(4b) taking the starting point of the path to be planned as the current state;

(4c) in the column of the current state, selecting an epsilon-greedy strategy, setting epsilon to be 0.8, namely randomly selecting a decimal from 0 to 1, if the selected decimal is from 0 to 0.8, taking the column with the maximum Q value as a feasible action, and if the selected decimal is from 0.8 to 1, randomly selecting a certain column with the maximum Q value as a feasible action, executing the action, and entering the next state;

(4d) calculating the instantaneous reward value of the current traffic intersection to the next adjacent traffic intersection by using the utility formula of each traffic intersection reached in the path to be planned, and updating the Q value in the Q table according to the following formula:

Q(s,a)＝Q'(s,a)+l[R(s,a)+γmax Q(s',a')-Q'(s,a)]

wherein Q (s, a) represents the updated Q value after the current traffic intersection s reaches the next adjacent traffic intersection a, Q ' (s, a) represents the Q value before the Q (s, a) is updated, l represents the learning rate, the value range is [0,1], R (s, a) represents the utility value corresponding to the current traffic intersection s reaching the next adjacent traffic intersection a, gamma represents the discount factor, the value range is [0,1], max Q (s ', a ') represents the maximum Q value corresponding to the adjacent traffic intersection a ' after the next traffic intersection s ' is reached;

(4e) judging whether the next traffic intersection is the end point of the path to be planned or not, if so, executing the step (4f), and otherwise, executing the step (4c) after the traffic intersection is taken as the current state;

(4f) judging whether the updated Q value is consistent with the Q value before updating or not, if so, executing a step 5 after the Q table is updated, otherwise, executing a step (4 b);

step 5, obtaining an optimal driving path:

taking the starting point of the path to be planned as the current state, finding the row where the current state is located in the updated Q table, selecting the column with the maximum Q value in the row as the next traffic intersection to be reached, taking the traffic intersection to be reached as the current state, selecting the column with the maximum Q value in the row where the current state is located as the next traffic intersection to be reached until the end point of the path to be planned is reached, and obtaining the optimal driving path;

and 6, the cloud server sends the planned optimal driving path to the vehicle user through the base station.

2. The vehicle path planning method based on digital twin and user personalized needs according to claim 1, characterized in that the path planning request in step 1 comprises: starting point and end point of the path to be planned, predicted time for reaching the destination, and starting time of travel.

3. The vehicle path planning method based on the digital twin and the personalized demand of the user as claimed in claim 1, wherein the vehicle flow density in step (2b) refers to vehicle flow data at the time of the current request to be planned, which is sensed, collected and uploaded to a cloud server according to the internet of things device placed at the roadside in the path to be planned.

Technical Field

The invention belongs to the technical field of physics, and further relates to a vehicle path planning method based on digital twin and user personalized requirements in the technical field of path planning. The invention can plan the route of the vehicle to the destination in the traffic system.

Background

Vehicle path planning belongs to the field of vehicle navigation, intelligent traffic scheduling and the like which are widely applied to vehicles. The vehicle path planning refers to finding an optimal path which is from a starting point to a terminal point and meets certain evaluation indexes by combining traffic information of a road network so that a vehicle can smoothly reach a destination. On the one hand, the existing vehicle path planning method usually performs path planning by using the minimum time or the minimum distance as an evaluation index, for example:

the patent document "an intelligent vehicle path search method under environmental constraints" (application number: 202110611728.8, application publication number: CN113256013A) applied by the university of beijing studys discloses a vehicle path planning method under environmental constraints. The method provides a nested algorithm which is formed by target point clustering and path searching and aims at minimizing the completion time of the last task by constructing a proper number of target points, vehicle numbers and maps under the constraint conditions of various complex terrains, fluctuating meteorology and the like, allocates the target points to various vehicles on the constructed map, and automatically searches and calculates the optimal driving path for each vehicle. Although the method is a searching and well-adapted path planning method, the method still has the following defects: the planned vehicle path lacks comprehensive consideration on user travel cost and time, so that the problem of low vehicle scheduling efficiency in the planned path is caused.

The patent document "unmanned vehicle route planning method for special scenes under 5G environment" (application number: 202010789006.7, application publication number: CN112068548A) applied by Beijing aerospace university discloses a vehicle route planning method for special scenes under 5G environment. The method depends on a route planning system of the unmanned vehicle based on 5G communication, the route planning system comprises the unmanned vehicle, a road side unit, a cloud platform and a 5G network, and the method comprises the steps that the cloud platform establishes a backup route library based on a historical track route; sensing and monitoring road state in real time by a road side sensor in the road side unit; the vehicle requests a path planning scheme from the cloud platform through the 5G network; and the cloud platform calculates and provides a path planning scheme, and the vehicle receives and executes the path planning scheme through the 5G network, so that the decision of path planning and adjustment is completed. However, the method has the following disadvantages: the problem of long-time delay in determining and issuing scheduling decisions when planning paths for a large number of vehicle users is caused by network delay in requesting data transmission between vehicles and clouds.

Disclosure of Invention

The invention aims to provide a vehicle path planning method based on digital twin and user personalized requirements aiming at overcoming the defects in the prior art and solving the problems that the vehicle scheduling efficiency in a planned path is not high and the time delay of determining and releasing scheduling decision is long when a large number of users plan the path.

The idea for realizing the purpose of the invention is as follows: by deploying the digital twin of the vehicle user at the cloud end, sending the path planning request and mapping the digital twin and the real world vehicle users one by one, the real world vehicle users can be replaced to directly carry out real-time information interaction with the cloud server, the time delay of long-distance data transmission between the vehicle and the cloud is reduced, and the problem of long time delay of determining and issuing scheduling decision-making caused when a large number of vehicle users plan paths is solved. Because the vehicle path planning method in the prior art mainly provides the driver with the path selection with the least traffic flow, the probability that a plurality of vehicles select the same path at the same time and the sudden traffic flow increase of flash time is generated immediately. Therefore, on the premise that the traffic flow is kept unchanged in the planning request time interval, the invention designs a strategy for collecting the toll according to the proportional relation between the traffic flow and the maximum traffic flow in the time period, namely when the traffic flow in the time period exceeds half of the maximum traffic flow, the toll is started to guide the vehicle to select different paths, and the personalized demand utility value of the path to be planned is calculated by combining the traffic time and the path planning request to be used as the road index of the path to be planned, so that the driving path with good cost performance and low toll can be planned for the vehicle according to the utility value, and the problem of low vehicle scheduling efficiency in the planned path is solved.

The specific steps for realizing the purpose of the invention are as follows:

step 1, establishing a digital twin sending path planning request:

a vehicle user to be planned with a path deploys a digital twin DTs on a cloud server and then sends a path planning request;

the path planning request includes: starting and ending points of a path to be planned, estimated time for reaching a destination and starting time of travel;

step 2, calculating the utility value of the path to be planned of the user personalized demand:

(2a) the cloud server adds the starting time of travel in the received path planning request to a service list of the cloud server;

(2b) calculating the driving time of each traffic intersection reached in the path to be planned according to the following formula:

wherein, tau_gThe normalized value is shown to be [ -1, 1] after the normalization from the previous traffic intersection to the adjacent g-th traffic intersection in the path to be planned]Time of range, d_gRepresenting the distance from the last traffic crossing in the path to be planned to the adjacent g-th traffic crossing, v_minRepresenting the minimum driving speed of the road section limit from the last traffic intersection to the adjacent g-th traffic intersection in the path to be planned, max representing the maximum value operation, v_maxRepresenting the maximum speed of travel, k, limited by the section of road from the last traffic crossing to the adjacent g-th traffic crossing in the path to be planned_gAnd k_maxRespectively representing the traffic flow density and the maximum traffic flow density from one traffic intersection on the path to be planned to the adjacent g-th traffic intersection;

the traffic flow density refers to traffic flow data when the current request to be planned is sensed, collected and uploaded to a cloud server according to internet of things equipment arranged at the road side in the path to be planned;

(2c) according to the following formula, the normalized toll of each adjacent traffic intersection road section in the path to be planned is designed to be collected according to the traffic flow density in the peak period:

wherein p is_gThe method indicates that the next g traffic intersection is reached at [ -1, 1] from the previous traffic intersection in the planned path to be collected]The toll after the range normalization is used, wherein cos represents cosine value taking operation;

(2d) calculating the cost performance of the driving time and the toll of the path to be planned according to the following formula:

wherein alpha represents the cost performance of the driving time and the toll of the path to be planned, and t_{Preparation of}Indicating the estimated time of arrival at the destination, t_minRepresents the shortest time to reach the destination;

(2e) calculating the utility value of each traffic intersection in the path to be planned according to the following formula;

r_g＝ατ_g+(1-α)p_g

wherein r is_gRepresenting a utility value from a previous traffic intersection in the path to be planned to an adjacent g-th traffic intersection;

step 3, establishing the state and action space of the traffic intersection:

taking each traffic intersection of the path to be planned as the state of the Q-learning algorithm from the selected traffic map, establishing the state space of all the traffic intersections, taking the selection of each traffic intersection of the path to be planned on the next adjacent traffic intersection as the action in the Q-learning algorithm, and establishing the action space of all the traffic intersections;

and 4, updating the Q table by using a Q-learning algorithm:

(4a) establishing a Q table, wherein the row of the Q table represents the state of each traffic intersection, the column represents the action selection of each traffic intersection, and the initialized Q value is set to be 0;

(4b) taking the starting point of the path to be planned as the current state;

(4c) in the column of the current state, selecting an epsilon-greedy strategy, setting epsilon to be 0.8, namely randomly selecting a decimal from 0 to 1, if the selected decimal is from 0 to 0.8, taking the column with the maximum Q value as a feasible action, and if the selected decimal is from 0.8 to 1, randomly selecting a certain column with the maximum Q value as a feasible action, executing the action, and entering the next state;

(4d) calculating the instantaneous reward value of the current traffic intersection to the next adjacent traffic intersection by using the utility formula of each traffic intersection reached in the path to be planned, and updating the Q value in the Q table according to the following formula:

Q(s,a)＝Q'(s,a)+l[R(s,a)+γmaxQ(s',a')-Q'(s,a)]

wherein Q (s, a) represents the updated Q value after the current traffic intersection s reaches the next adjacent traffic intersection a, Q ' (s, a) represents the Q value before the Q (s, a) is updated, l represents the learning rate, the value range is [0,1], R (s, a) represents the utility value corresponding to the current traffic intersection s reaching the next adjacent traffic intersection a, gamma represents the discount factor, the value range is [0,1], maxQ (s ', a ') represents the maximum Q value corresponding to the adjacent traffic intersection a ' after the next traffic intersection s ' reaches;

(4e) judging whether the next traffic intersection is the end point of the path to be planned or not, if so, executing the step (4f), and otherwise, executing the step (4c) after the traffic intersection is taken as the current state;

(4f) judging whether the updated Q value is consistent with the Q value before updating or not, if so, executing a step 5 after the Q table is updated, otherwise, executing a step (4 b);

step 5, obtaining an optimal driving path:

taking the starting point of the path to be planned as the current state, finding the row where the current state is located in the updated Q table, selecting the column with the maximum Q value in the row as the next traffic intersection to be reached, taking the traffic intersection to be reached as the current state, selecting the column with the maximum Q value in the row where the current state is located as the next traffic intersection to be reached until the end point of the path to be planned is reached, and obtaining the optimal driving path;

and 6, the cloud server sends the planned optimal driving path to the vehicle user through the base station.

Compared with the prior art, the invention has the following advantages:

firstly, the digital twin body is established to send the path planning request, and the digital twin body and the real world vehicle users are mapped one by one, so that the real world vehicle users can be replaced to directly carry out real-time information interaction with the cloud server, and the problem of time delay of long-distance data transmission between vehicles and clouds in the path planned by the prior art is solved, so that the network time delay of path planning is reduced, the real-time performance of data interaction between vehicles and clouds is improved, and the time of path planning is shortened.

Secondly, the utility value of the path to be planned, which is required by the user in an individualized way, is calculated, and the toll guidance strategy is designed according to the traffic flow and is used as the road index of the path to be planned, so that the problem of low traffic flow scheduling efficiency of path planning by using the shortest time or shortest distance as an index of the vehicles planned in the prior art is solved, and the efficiency of traffic flow scheduling can be improved by the path planned by the method.

Drawings

FIG. 1 is a flow chart of the present invention;

FIG. 2 is a schematic diagram of the DT architecture of the present invention;

FIG. 3 is a simulation diagram of the present invention.

Detailed Description

The invention is described in further detail below with reference to the figures and examples.

The implementation steps of the present invention are described in further detail with reference to fig. 1.

Step 1, establishing a digital twin body transmission path planning request.

And the vehicle user to be planned sends a path planning request after deploying the digital twin DTs on the cloud server.

The path planning request includes: starting point and end point of the path to be planned, predicted time for reaching the destination, and starting time of travel.

And 2, calculating the utility value of the path to be planned of the user personalized demand.

And the cloud server adds the starting time of the trip in the received path planning request to a service list of the cloud server.

Calculating the driving time of each traffic intersection reached in the path to be planned according to the following formula:

wherein, tau_gRepresents the time between the arrival of the first traffic intersection in the path to be planned and the arrival of the adjacent g-th traffic intersection at [ -1,1]Time after range normalization, d_gRepresenting the distance from the last traffic crossing in the path to be planned to the adjacent g-th traffic crossing, v_minRepresenting the minimum driving speed of the road section limit from the last traffic intersection to the adjacent g-th traffic intersection in the path to be planned, max representing the maximum value operation, v_maxRepresenting the maximum speed of travel, k, limited by the section of road from the last traffic crossing to the adjacent g-th traffic crossing in the path to be planned_gAnd k_maxRespectively representing the traffic flow density and the maximum traffic flow density from one traffic intersection to the adjacent g-th traffic intersection on the path to be planned.

The traffic flow density refers to traffic flow data when the current request to be planned is sensed, collected and uploaded to a cloud server according to internet of things equipment arranged on the road side in the path to be planned.

According to the following formula, the normalized toll of each adjacent traffic intersection road section in the path to be planned is designed to be collected according to the traffic flow density in the peak period:

wherein p is_gIndicates a planned receiptThe upper traffic intersection in the path to be planned reaches the adjacent g-th traffic intersection at [ -1,1 [ -1 [ ]]And (4) carrying out range normalization on the toll, wherein cos represents cosine value taking operation.

Calculating the cost performance of the driving time and the toll of the path to be planned according to the following formula:

wherein alpha represents the cost performance of the driving time and the toll of the path to be planned, and t_{Preparation of}Indicating the estimated time of arrival at the destination, t_minIndicating the shortest time to reach the destination.

And according to the following formula, calculating the utility value of each traffic intersection reached in the path to be planned according to the driving time of each traffic intersection reached in the solved path to be planned, the normalized toll of each adjacent traffic intersection road section in the path to be planned and the cost performance of the toll and the driving time of the path to be planned.

The formula for calculating the utility value of each traffic intersection reached in the path to be planned is as follows:

r_g＝ατ_g+(1-α)p_g

wherein r is_gAnd the utility value of the adjacent g-th traffic intersection from the last traffic intersection in the path to be planned is shown.

And 3, establishing the state and the action space of the traffic intersection.

And taking each traffic intersection of the path to be planned as the state of the Q-learning algorithm from the selected traffic map, establishing the state space of all the traffic intersections, taking the selection of each traffic intersection of the path to be planned on the next adjacent traffic intersection as the action in the Q-learning algorithm, and establishing the action space of all the traffic intersections.

And 4, updating the Q table by using a Q-learning algorithm.

In the first step, a Q table is established, the rows of the table represent the state of each traffic intersection, the columns represent the action selection of each traffic intersection, and the initialized Q value is set to 0.

And secondly, taking the starting point of the path to be planned as the current state.

And thirdly, selecting an epsilon-greedy strategy in the row of the current state, setting epsilon to be 0.8, namely randomly selecting a decimal from 0 to 1, if the selected decimal is from 0 to 0.8, taking the row with the maximum Q value as a feasible action, and if the selected decimal is from 0.8 to 1, randomly selecting a certain row with the maximum Q value as a feasible action, executing the action and entering the next state.

Fourthly, calculating the instantaneous reward value of the current traffic intersection to the next adjacent traffic intersection by using the utility formula of each traffic intersection reached in the path to be planned, and updating the Q value in the Q table according to the following formula:

Q(s,a)＝Q'(s,a)+l[R(s,a)+γmaxQ(s',a')-Q'(s,a)]

wherein, Q (s, a) represents the updated Q value after the current traffic intersection s reaches the next adjacent traffic intersection a, Q ' (s, a) represents the Q value before Q (s, a) is updated, l represents the learning rate, the value range is [0,1], R (s, a) represents the utility value corresponding to the current traffic intersection s reaching the next adjacent traffic intersection a, γ represents the discount factor, the value range is [0,1], maxQ (s ', a ') represents the maximum Q value corresponding to the adjacent traffic intersection a ' after the next traffic intersection s ' is reached.

And fifthly, judging whether the next traffic intersection is the end point of the path to be planned, if so, executing the sixth step, otherwise, executing the third step after taking the traffic intersection as the current state.

Step six, judging whether the updated Q value is consistent with the Q value before updating or not, if so, finishing the updated Q table and then executing the step 5, otherwise, executing the second step of the step;

and 5, acquiring the optimal driving path.

And taking the starting point of the path to be planned as the current state, finding the row where the current state is located in the updated Q table, selecting the column with the maximum Q value in the row as the next traffic intersection to be reached, taking the traffic intersection to be reached as the current state, selecting the column with the maximum Q value in the row where the current state is located as the next traffic intersection to be reached until the end point of the path to be planned is reached, and obtaining the optimal driving path.

And 6, the cloud server sends the planned optimal driving path to the vehicle user through the base station.

The invention will be further described with reference to the embodiment of fig. 2.

For three vehicle users with path planning requests in a real traffic system, digital twins DT (digital twin) of the vehicle users are deployed in a cloud end, and one-to-one mapping of the vehicle users is completed. In fig. 2, DT of each of the three vehicle users may replace the vehicle user to send a path planning request to the cloud server, where the path planning request includes a start point and an end point of a path to be planned and a predicted time to reach a destination, and a starting time of a trip, the cloud server sequentially adds the starting times of the trip of the planning requests of the three vehicle users to the service list in sequence, and the cloud server may sequentially plan the paths of the three vehicle users respectively according to the starting time sequences of the three vehicle users in the service list. The process of planning the path of the vehicle by the cloud server is described in fig. 2 by taking the vehicle i as an example. The cloud server calculates the utility value of the path to be planned of the vehicle i according to the traffic flow data of the current request to be planned, sensed and collected by the Internet of things equipment arranged at the roadside in the path to be planned of the vehicle i, and the step 2 of the invention, obtains the updated Q table according to the step 4 of the invention, plans the optimal path of the vehicle i according to the Q table, and sends the planned optimal path to the vehicle i through the base station covering the action range of the vehicle i.

The effect of the present invention will be further described with reference to simulation experiments.

1. Simulation experiment conditions are as follows:

the platform of the simulation experiment of the invention is as follows: windows 10 operating system and Matlab 2016 b.

The map used in the simulation experiment of the invention is an urban traffic map of xi' an city in Shaanxi province.

2. Simulation content and result analysis thereof:

the simulation experiment of the invention is to respectively plan the path of an input automobile by adopting the invention and two prior arts (a path planning method based on shortest time and a path planning method based on shortest distance). The simulation experiment parameter settings are shown in table 1:

table 1 experimental parameter settings

Setting item	Value taking
		Total number of road sections	30
Maximum speed	110km/h
		Minimum velocity	10km/h
Maximum traffic density	80veh/km
		Exploration factor epsilon	0.8
Learning rate l	0.9
		Cost performance ratio alpha	[0,1]

In the simulation experiment, two prior arts are adopted:

the path planning method based on the shortest time in the prior art refers to: durdu, et al, put forward a Path planning method with Q-learning in Path planning of mobile robots, 2014, pp.2162-2165, abbreviated as the shortest time based Path planning method.

The path planning method based on the shortest distance in the prior art refers to: the path planning method proposed by Wooyoung KWon et al in "Fast relationship using stored short paths for a mobile robot,2007IEEE/RSJ International Conference on Intelligent Robots and Systems,2007, pp.82-87" is called the shortest distance-based path planning method for short.

The effect of the present invention will be further described with reference to the simulation diagram of fig. 3.

Fig. 3 is a comparison graph of the total utility value of the path planning from the starting point to the end point of one vehicle respectively by the three methods of the simulation experiment of the present invention. The total utility value refers to the sum of the utility values of all road sections forming the optimal path in the optimal planned path from the starting point to the end point of one vehicle.

The simulation experiment of the invention inputs different cost performance of the automobile to time and toll, the value range is [0,1], the value is taken at 0.02 intervals, 50 different cost performance ratios are totally taken, and three methods are adopted to respectively finish a comparison graph of the total utility value of the route planning from the starting point to the terminal point of the automobile.

In fig. 3, the abscissa represents the cost performance of the user of the vehicle to be planned for time and toll, and the ordinate represents the total utility value of the user for completing the route planning from the starting point to the ending point. The curves marked with stars in fig. 3 represent the user total utility curves obtained with the method of the present invention for different cost-performance parameters. And the curve marked by a circle represents the total utility curve of the user under different cost performance parameters obtained by adopting a shortest time-based path planning method. And the curve marked by the square represents the total utility curve of the user under different cost performance parameters obtained by adopting a path planning method based on the shortest distance.

As can be seen from fig. 3, for any fixed cost performance ratio, the total utility value obtained by the present invention can always bring the highest utility to the vehicle user compared with the total utility values obtained by the other two methods, mainly because the shortest time path planning method only considers one index factor to plan the path of the vehicle user, wherein the personalized demand of the user is ignored, and for the shortest distance, only the distance of each road section is considered to obtain the shortest path, and the consideration of time and toll of each road section is lacked.

The above simulation experiments show that: the method can plan the optimal path for the user by considering the individual requirements of the user on the road section indexes of the adjacent traffic intersections, including time and toll, solves the problem that the prior art only takes the minimum time or the shortest distance as an evaluation index to plan the path, so that the traffic flow scheduling efficiency is low, and is a very practical path planning method.