CN111311959A

CN111311959A - Multi-interface cooperative control method and device, electronic equipment and storage medium

Info

Publication number: CN111311959A
Application number: CN202010087988.5A
Authority: CN
Inventors: 王鲁晗; 张宇恒; 温向明; 路兆铭
Original assignee: Beijing University of Posts and Telecommunications
Current assignee: Beijing University of Posts and Telecommunications
Priority date: 2020-02-12
Filing date: 2020-02-12
Publication date: 2020-06-19
Anticipated expiration: 2040-02-12
Also published as: CN111311959B

Abstract

The embodiment of the disclosure discloses a multi-interface cooperative control method and device, electronic equipment and a storage medium for networking automatic driving, wherein the method comprises the following steps: acquiring the current state information of the target intersection, wherein the state information comprises real-time traffic information of the target intersection, the vehicle state of a target vehicle driving to the target intersection and the occupation information of the target intersection; determining the maximum passing speed of the target vehicle according to the state information; determining a target driving sequence of the target vehicle passing through the target intersection under the state information by utilizing a deep Q learning network; planning the target speeds of a plurality of target vehicles passing through the target intersection according to the target running sequence according to the maximum passing speed; the target speed is the speed of the target vehicle passing through the target intersection under the condition that the maximum passing speed is not exceeded and the target vehicle does not collide with other vehicles; and sending the target speed to the target vehicle to control the target vehicle to run according to the target speed.

Description

Multi-interface cooperative control method and device, electronic equipment and storage medium

Technical Field

The disclosure relates to the technical field of wireless communication and automatic driving, in particular to a multi-interface cooperative control method and device for internet-oriented automatic driving, an electronic device and a storage medium.

Background

The road condition of the traffic intersection is complex, and the traffic intersection is an area with multiple traffic accidents. According to the traffic statistics result, the traffic accidents at the intersection account for about 40 percent of all the traffic accidents. Meanwhile, the time waste caused by traffic congestion in urban areas amounts to 69 hundred million hours each year. Traffic lights have been invented for over a hundred years. The most effective solution for intersection traffic flow control management at traffic lights has been in the past century. In recent years, a large variety of traffic light phase control methods have emerged, but the performances of road safety and traffic efficiency are still unsatisfactory. The cooperative intelligent driving technology has great potential in the aspect of improving the intersection control performance.

With the development of various technologies such as computer technology, 5G communication technology, sensor technology and the like, and the strong support of national policies, the development of internet-connected automatic driving vehicles will become a major trend in the future. Under the future scene of internet automatic driving, the existing traffic light intersection control scheme has a great optimization space. By installing road side facilities (including road side cameras and road side radars) and an automatic driving vehicle, detailed real-time traffic information and vehicle states can be obtained. Through the cooperation of a communication network, information interaction is realized between vehicles and between the vehicles and a base station, and in a future networked automatic driving system, the automatic vehicle can accurately measure and control the motion state of the automatic vehicle through a sensor. Therefore, the edge calculation server arranged near the intersection can issue instructions to the automatic driving vehicles, so that the passing of the automatic driving vehicles is coordinated and scheduled, and the vehicles can safely and efficiently pass through the intersection under the condition of no traffic lights.

The core of the automatic driving intersection control scheme without the traffic light is to design a set of automatic driving intersection collision avoidance and resource optimization distribution solution. Namely 1) predicting all possible collisions when no measures are taken by utilizing real-time road perception information, and taking corresponding measures to coordinate vehicles to avoid the collisions; 2) and optimizing the formed collision avoidance scheme and improving the crossing passing efficiency. Meanwhile, the passing fairness of all vehicles is considered, and a certain vehicle is prevented from waiting for a long time. In particular, urban traffic networks are areas where traffic congestion occurs more severely. Usually, the distance between each intersection is small, and the traffic flow of adjacent intersections has large relevance. And in most cases, the traffic flow in different areas is not balanced. At this time, the optimization of the collision avoidance scheme cannot be limited to a single intersection, and regional optimization needs to be made in consideration of the traffic states of surrounding intersections.

Disclosure of Invention

The embodiment of the disclosure provides a multi-interface cooperative control method and device for internet-oriented automatic driving, an electronic device and a computer-readable storage medium.

In a first aspect, an embodiment of the present disclosure provides a multi-intersection cooperative control method for internet-oriented automatic driving, including:

acquiring state information of a target intersection at the current moment, wherein the state information comprises real-time traffic information of the target intersection, a vehicle state of a target vehicle driving to the target intersection and occupation information of the target intersection; the real-time traffic information includes location information of the target vehicle; the vehicle state includes speed information of the target vehicle; the occupation information of the target intersection comprises the intersection condition of the vehicle track of the target vehicle which drives out of the target intersection and the vehicle track of other vehicles which drive out of the adjacent intersection;

determining the maximum passing speed of the target vehicle according to the state information;

determining a target driving sequence of the target vehicle through the target intersection under the state information by utilizing a deep Q learning network;

planning the target speeds of a plurality of target vehicles passing through the target intersection according to the target running sequence according to the maximum passing speed; the target speed is the speed of the target vehicle passing through the target intersection under the condition that the maximum passing speed is not exceeded and no collision occurs to other vehicles;

and sending the target speed to the target vehicle to control the target vehicle to run according to the target speed.

Wherein, still include:

and adding the state information into a memory base of the deep Q learning network, and storing the state information in a mode of covering the earliest record in the memory base under the condition that the length of the memory base is full.

Wherein determining a driving sequence of the target vehicle through the target intersection under the state information using a deep Q learning network comprises:

predicting Q values corresponding to different candidate driving sequences of a plurality of target vehicles by using the deep Q learning network;

determining the candidate travel speed at which the Q value is maximum as the target travel order.

Wherein, the deep Q learning network adopts a convolution neural network, and further comprises:

acquiring a target Q value corresponding to one or more adjacent intersections corresponding to the target intersection at the last moment, and acquiring a return value obtained by the target driving sequence by using the deep Q learning network;

and updating the neuron parameters of the convolutional neural network according to the target Q value and the return value.

Wherein the neuron parameters are updated by the following Q value update formula:

wherein, theta_iFor the neuron parameters of the convolutional neural network, α (t) and gamma are preset attenuation factors, theta_i' is the neuron parameter of the updated convolutional neural network, J is the set of all adjacent intersections, omega_i,jAnd (4) a conversion coefficient of the Q value of the adjacent intersection j to the Q value of the target intersection i, wherein omega is a penalty function.

Wherein determining a maximum traffic speed of the plurality of target vehicles according to the state information comprises:

adjusting a first speed when a first time is earlier than a second time so that the first time is later than the second time; the first time is the time when the current target vehicle passes through the target intersection according to the first speed under the state information, and the second time is the time when the previous target vehicle passes through the target intersection according to the second speed under the state information;

determining the adjusted first speed as the maximum passing speed.

The planning of the target speeds of the plurality of target vehicles passing through the target intersection according to the target driving sequence according to the maximum passing speed comprises the following steps:

when the speed planning of other current target vehicles running before the target vehicle is not finished, determining that the current target speed of the target vehicle is equal to the maximum passing speed;

determining that the current target speed of the target vehicle is equal to the maximum passing speed when all other target vehicles previously driven by the current target vehicle have completed the speed plan and the maximum passing speed is less than or equal to the minimum value;

when all other current target vehicles running before the target vehicle have finished the speed planning and the maximum passing speed is greater than the minimum value, determining whether the target vehicle conflicts with other vehicles after running to the target intersection at the initial planned speed; and when the conflict occurs, reducing the initial planning speed, and adjusting to the step of determining whether the reduced initial planning speed is less than or equal to the minimum value.

In a second aspect, the present disclosure provides a multi-intersection cooperative control device for internet-oriented automatic driving, including:

the system comprises an acquisition module, a display module and a control module, wherein the acquisition module is configured to acquire the current state information of a target intersection, and the state information comprises real-time traffic information of the target intersection, the vehicle state of a target vehicle driving to the target intersection and the occupation information of the target intersection; the real-time traffic information includes location information of the target vehicle; the vehicle state includes speed information of the target vehicle; the occupation information of the target intersection comprises the intersection condition of the vehicle track of the target vehicle which drives out of the target intersection and the vehicle track of other vehicles which drive out of the adjacent intersection;

a first determination module configured to determine a maximum transit speed of a target vehicle from the state information;

a second determination module configured to determine a target driving order of the target vehicle through the target intersection under the state information using a deep Q learning network;

a planning module configured to plan target speeds of a plurality of the target vehicles passing through the target intersections in the target driving order according to the maximum passing speed; the target speed is the speed of the target vehicle passing through the target intersection under the condition that the maximum passing speed is not exceeded and no collision occurs to other vehicles;

a transmitting module configured to transmit the target speed to the target vehicle to control the target vehicle to travel at the target speed.

The functions can be realized by hardware, and the functions can also be realized by executing corresponding software by hardware. The hardware or software includes one or more modules corresponding to the above-described functions.

In one possible design, the structure of the internet-oriented automatic driving-oriented multi-port cooperative control device includes a memory and a processor, the memory is used for storing one or more computer instructions for supporting the internet-oriented automatic driving-oriented multi-port cooperative control device to execute the method in the first aspect, and the processor is configured to execute the computer instructions stored in the memory. The multi-interface cooperative control device for the internet automatic driving can also comprise a communication interface, and the multi-interface cooperative control device for the internet automatic driving is used for communicating with other equipment or a communication network.

In a third aspect, an embodiment of the present disclosure provides an electronic device, including a memory and a processor; wherein the memory is configured to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method of any of the above aspects.

In a fourth aspect, the disclosed embodiments provide a computer-readable storage medium for storing computer instructions for use by any of the above-mentioned apparatuses, including computer instructions for performing the method according to any of the above-mentioned aspects.

The technical scheme provided by the embodiment of the disclosure can have the following beneficial effects:

the multi-intersection cooperative control method is specially used for automatically driving vehicles through the internet, the passing efficiency of intersections can be greatly improved, and traffic jam is relieved. In addition, the final planning solution is obtained by adopting an algorithm with extremely high computing resource computing speed, and the algorithm can enhance the emergency response capability of the algorithm and maintain the reliability of the algorithm in the problem of high real-time requirement of intersection control.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.

Drawings

Other features, objects, and advantages of the present disclosure will become more apparent from the following detailed description of non-limiting embodiments when taken in conjunction with the accompanying drawings. In the drawings:

fig. 1 shows a flowchart of a multi-intersection cooperative control method for internet-oriented automatic driving according to an embodiment of the present disclosure;

FIG. 2 illustrates an effect schematic diagram of vehicle trajectories at a traffic-light free intersection according to an embodiment of the present disclosure;

FIG. 3 illustrates a schematic diagram of a real-time vehicle status and status information at a target intersection according to an embodiment of the present disclosure;

FIG. 4 is a schematic representation of occupancy information for the target intersection shown in FIG. 3;

FIG. 5 illustrates a diagram of the visual effects of planning a target speed according to an embodiment of the present disclosure;

fig. 6 is a schematic structural diagram of an electronic device suitable for implementing a multi-intersection cooperative control method for internet-oriented automatic driving according to an embodiment of the present disclosure.

Detailed Description

Hereinafter, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily implement them. Also, for the sake of clarity, parts not relevant to the description of the exemplary embodiments are omitted in the drawings.

In the present disclosure, it is to be understood that terms such as "including" or "having," etc., are intended to indicate the presence of the disclosed features, numbers, steps, behaviors, components, parts, or combinations thereof, and are not intended to preclude the possibility that one or more other features, numbers, steps, behaviors, components, parts, or combinations thereof may be present or added.

It should be further noted that the embodiments and features of the embodiments in the present disclosure may be combined with each other without conflict. The present disclosure will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.

The details of the embodiments of the present disclosure are described in detail below with reference to specific embodiments.

Fig. 1 shows a flowchart of a multi-intersection cooperative control method for internet-oriented automatic driving according to an embodiment of the present disclosure. As shown in fig. 1, the multi-intersection cooperative control method for internet-oriented automatic driving includes the following steps:

in step S101, obtaining state information of a target intersection at a current time, where the state information includes real-time traffic information of the target intersection, a vehicle state of a target vehicle traveling to the target intersection, and occupancy information of the target intersection; the real-time traffic information includes location information of the target vehicle; the vehicle state includes speed information of the target vehicle; the occupation information of the target intersection comprises the intersection condition of the vehicle track of the target vehicle which drives out of the target intersection and the vehicle track of other vehicles which drive out of the adjacent intersection;

in step S102, determining the maximum passing speeds of a plurality of target vehicles according to the state information;

in step S103, determining a target traveling sequence of the target vehicle through the target intersection under the state information using a deep Q learning network;

in step S104, planning target speeds of a plurality of target vehicles passing through the target intersection in the target driving sequence according to the maximum passing speed; the target speed is the speed of the target vehicle passing through the target intersection under the condition that the maximum passing speed is not exceeded and no collision occurs to other vehicles;

in step S105, the target speed is sent to the target vehicle to control the target vehicle to travel at the target speed.

In the embodiment, aiming at the automatic driving scene, the automatic driving vehicles are coordinated to cooperate with a high-efficiency collision-free crossing solution by collecting the vehicle information and the road information, so that the high-efficiency orderly passing of the traffic-light-free crossing is realized.

The target intersection in this embodiment may be any intersection without a traffic light in an automatic driving scene. The multi-intersection cooperative control method for the internet-oriented automatic driving in the embodiment of the disclosure can be executed on the edge computing server, and the edge computing server plans the passing speed of each automatic driving vehicle in real time for a plurality of adjacent intersections in the intersection without the traffic light, so that the automatic driving vehicles passing through each adjacent intersection cannot collide when passing through the intersection without the traffic light, and each automatic driving vehicle passes through the intersection without the traffic light with the maximum passing efficiency.

The edge computing server collects the state information of each intersection at each moment in real time from the road side facilities and/or the automatic driving vehicles. In the following, how the edge computing server implements multi-intersection cooperative control will be described by taking an example of one intersection (called as a target intersection) of a plurality of adjacent intersections for an intersection without traffic lights. It should be noted that, the actual situations of the intersections without the traffic lights are different, the number of adjacent intersections for performing multi-intersection cooperative control and the intersection situations of the tracks between the adjacent intersections are also different, but the process of realizing the multi-intersection cooperative control by planning the passing speed of each target vehicle passing through the target intersection for each target intersection is consistent, so it can be understood that although the speed planning flow of the target vehicle related to the target intersection is designed below, other vehicles related to the target intersection and vehicles related to other intersections can all perform speed planning in the same flow, and finally, when each vehicle passes through the corresponding intersection at the planned passing speed, the vehicles do not collide with each other, and each vehicle passes through each intersection at the maximum passing efficiency.

For a target intersection, the state information of the current moment may include, but is not limited to, real-time traffic information of the target intersection, vehicle states of a plurality of target vehicles to pass through the target intersection, and occupancy information of the target intersection; the real-time traffic information may include, but is not limited to, location information of the target vehicle; the vehicle state may include, but is not limited to, speed information of the target vehicle; the occupancy information of the target intersection comprises occupancy conditions of one or more occupancy points involved in the target intersection, and the occupancy points represent intersection points between vehicle tracks involved in the target intersection. The occupancy of the occupancy point can be represented by the intersection of the vehicle trajectory of the target vehicle exiting the target intersection and the vehicle trajectories of other vehicles exiting the adjacent intersection, i.e., the situation where the target vehicle collides with other vehicles.

The status information of the target intersection is exemplified by the following figures.

Fig. 2 is a schematic diagram illustrating the effect of vehicle trajectories at an intersection without traffic lights according to an embodiment of the present disclosure. As shown in fig. 2, the intersection without traffic lights is an intersection, the intersection without traffic lights includes 4 intersections, the target intersection can be any one of the 4 intersections, and the other 3 intersections are adjacent intersections of the target intersection. As can be seen from fig. 2, the driving tracks of the vehicles entering or leaving the traffic light intersection from each intersection are shown by solid arrows in fig. 2, and the positions where the respective vehicle tracks intersect, i.e., the black solid origins in fig. 2, are the occupied points related to the target intersection. If any two vehicles pass the occupancy point at the same time, a collision occurs.

FIG. 3 illustrates a schematic diagram of a real-time vehicle status and status information at a target intersection according to an embodiment of the present disclosure. As shown in fig. 3, the target intersection corresponds to a two-way road, and the target vehicle on the two-way road is going to the stop line of the target intersection at the current time in the manner shown in fig. 3, and the stop line is the waiting position of the vehicle passing through the intersection without traffic lights on the actual road. And maintaining the distance corresponding to the data granularity L between the target vehicles.

The target intersection shown in fig. 3, the real-time traffic information and the vehicle status can be described as vectors shown in fig. 3. Assuming that the target intersection is the ith intersection, the real-time traffic information of the target intersection can be represented as Pⁱ＝{p₁,p₂…,p_M}^TM is a 0-1 state vector of dimension M, and M is preset and used for representing the position of each target vehicle; the vehicle state information of the target intersection can be represented as Vⁱ＝{v₁,v₂…,v_M}^TAnd is an M-dimensional normalized state vector for representing the traveling speed of each target vehicle. The target vehicle may include a plurality of vehicles traveling toward a stop-line of the target intersection.

Fig. 4 is a schematic representation of occupancy information for the target intersection shown in fig. 3. As shown in fig. 4, the resource occupation of an occupation point related to the target intersection is on the time axis. The target intersection usually involves a plurality of occupation points, so the occupation information of one target intersection can be represented as Oⁱ＝{o₁,o₂…,o_N}^TIs an N-dimensional normalized state vector. The occupation information of the target crossing comprises the occupation condition of the target crossing, wherein the occupation condition can be used for representing the situation that the target crossing is occupied at any moment, and if the target crossing is occupied at a certain moment, the situation shows that one or more occupation points related to the target crossing have the occupation conditions from the adjacent points at the momentVehicles driven out of the intersection pass through, and the occupied points represent intersection points between vehicle tracks related to the target intersection.

The state information of the target intersection i can be represented as Sⁱ＝(Pⁱ；Oⁱ；Vⁱ)。

After the edge computing server collects the state information of the target intersection at the current moment, the maximum passing speed of the target vehicle can be determined according to the state information. It should be noted that the target vehicle is one or more vehicles that will travel to the stop line of the target intersection.

The maximum traffic speed may be a speed at which the vehicle travels to a stop line of the target intersection without colliding with other vehicles according to the maximum traffic speed. In some embodiments, the time when the target vehicle travels to the stop-line of the target intersection at the maximum traffic speed may not be earlier than the time when a vehicle ahead of the target vehicle travels to the stop-line of the target intersection. That is, for any vehicle k for which no feasible passing scheme is found currently, the time for the vehicle k to pass through the intersection stop line in the current state is set as

The time when the previous vehicle k-1 of the vehicle k passes through the stop line of the intersection in the current state is

If it is

The speed of the vehicle k needs to be adjusted

The deep Q-learning network may be pre-trained offline. The edge computing server inputs the collected state information of the current time of the target intersection into the deep Q learning network, the deep Q learning network predicts the sequence relation of stop lines of a plurality of target vehicles corresponding to the target intersection, wherein the stop lines are driven to the target intersection, and the deep Q learning network can obtain the corresponding Q value according to any sequence relation. In some embodiments, the precedence relationship corresponding to the maximum Q value may be determined as the target driving order.

It should be noted that, the execution sequence of the steps of determining the maximum passing speed of the target vehicle and predicting the target driving sequence by using the deep Q learning network is not limited, that is, the step of determining the maximum passing speed of the target vehicle may be executed first, and then the step of predicting the target driving sequence by using the deep Q learning network may be executed, or the step of predicting the target driving sequence by using the deep Q learning network may be executed first, and then the step of predicting the target driving sequence by using the deep Q learning network may be executed, and both the steps may be executed simultaneously.

After the maximum passing speed of each target vehicle and the target running sequence among the target vehicles are determined, the maximum passing speed of each target vehicle can be adjusted, and the running speed of each target vehicle is enabled to be maximum under the condition that the target vehicles are prevented from colliding when running to the stop line of the target intersection according to the maximum passing speed. In the planning process, the target speed is less than or equal to the maximum traffic speed, and when the vehicle travels to the stop line of the target intersection according to the target speed, the vehicle does not collide with the front target vehicle. Therefore, a maximum speed at which the target vehicle does not collide with other target vehicles can be found from a range less than or equal to the maximum passing speed on the condition of the maximum passing speed and the target traveling order, and is taken as the target speed.

The edge calculation server, after determining the target speed of each target vehicle, may send the target speed to the control unit of the target vehicle, which is controlled by the control unit of the target vehicle to travel at the target speed.

The edge computing server can plan and control the speed of each target vehicle according to the flow. Because the planned target speed takes the speed planning condition of other adjacent intersections into consideration (namely the occupation condition of the target intersection, and the time information of the vehicles driven out from other adjacent intersections to the occupation point into account), the multi-intersection cooperative control can be realized through the embodiment of the disclosure.

At present, most of the existing multi-intersection cooperative control algorithms are control algorithms of traffic light intersections, and for an automatic driving traffic system in the absence of internet connection, the effect of improving traffic efficiency is limited. The multi-intersection cooperative control method is specially used for automatically driving vehicles through the internet, the passing efficiency of intersections can be greatly improved, and traffic jam is relieved.

The existing control methods for the network-connected automatic driving intersections are mostly single intersection control methods, and the road conditions of surrounding intersections are not considered cooperatively. The multi-intersection cooperative control method can be used for making an intelligent decision by integrating the regional traffic conditions, balancing the flow of each section of the traffic network and greatly relieving the local traffic jam condition.

The method adopts an algorithm with extremely high computing resource operation speed to obtain the final planning solution, and adopts the rapid algorithm to enhance the emergency capability of the algorithm to emergency and maintain the reliability of the algorithm in the problem of high real-time requirement of intersection control.

In an optional implementation manner of this embodiment, the multi-interface cooperative control method further includes: and adding the state information into a memory base of the deep Q learning network, and storing the state information in a mode of covering the earliest record in the memory base under the condition that the length of the memory base is full.

In this alternative implementation, the state information may be added to a memory bank of the deep Q learning network, the length of the memory bank being limited, so that each time the latest data is added, the oldest piece of data in the memory bank is overwritten. In the memory bank, the data at the current time t is stored in the form of

Wherein the content of the first and second substances,

the optimal decision sequence obtained by the edge calculation server at the time t through the deep Q learning network, namely the target row of each target vehicleAnd the driving sequence of the target vehicle corresponds to the driving sequence. A. theⁱ＝{a₁,a₂,...,a_HIs a sequence of length H. Sequence a₁～a_HThe sequence of (2) is the sequence of the edge calculation planning vehicle passing speed in the fourth step. a is_kK, may be a certain vehicle for which no feasible passing solution is currently found, and K is the number of vehicles for which no feasible passing solution is currently found. In a state

In the following, the first and second parts of the material,

the element in (A) needs to satisfy the following conditions

Then

Representing the reported value obtained when the edge calculation server makes the corresponding decision at time t.

In an optional implementation manner of this embodiment, step S103, namely, the step of determining, by using the deep Q learning network, a driving sequence of the target vehicle through the target intersection under the state information, further includes:

In this optional implementation manner, the deep Q learning network may predict, for the state information at the current time, a driving sequence between target vehicles corresponding to a target intersection, where each driving sequence may be a candidate driving sequence, and each candidate driving sequence may obtain a corresponding Q value. The candidate travel order having the largest Q value may be determined as the target travel order.

In an optional implementation manner of this embodiment, the deep Q learning network employs a convolutional neural network, and further includes:

In this optional implementation, the deep Q learning Neural network may adopt a Convolutional Neural Network (CNN), and is composed of a target network and an evaluation network. The target network evaluation can obtain a target driving sequence A' corresponding to the current maximum Q value, and the evaluation network obtains the Q value of the adjacent intersection at the last moment and the return of the operation

And then, updating the weights of all the neurons, and transmitting the updated weights to the target network. The convergence rate of the deep learning network is low, and in order to ensure the high efficiency and the flexibility of the algorithm, a mode of firstly training off-line data and then learning on line can be adopted.

Wherein the content of the first and second substances,

representing the return value obtained when the edge calculation server makes the corresponding decision at time t,

where K represents the set of all target vehicles for which no feasible passing solutions have been currently found. w is a_k,tThe waiting time of the target vehicle k to the target intersection i from the moment t, w_k,t+1The waiting time of the target vehicle k at the target intersection i is required until the moment t + 1.

In an optional implementation manner of the present embodiment, the neuron parameter is updated by the following Q value update formula:

In this alternative implementation, the adjacent intersection J is an adjacent intersection of the target intersection i, for example, as shown in fig. 2, when the target intersection i is one of the intersections, the other 3 intersections are adjacent intersections J, J being 3 of the target intersection i.

In an optional implementation manner of this embodiment, in step S102, the step of determining the maximum passing speeds of the plurality of target vehicles according to the state information further includes:

determining the adjusted first speed as the maximum passing speed.

In this optional implementation manner, for any one target vehicle k for which no feasible passing scheme is currently found, the time when the target vehicle k passes through the stop line of the target intersection in the current state is set as

The time when the previous target vehicle k-1 of the target vehicle k passes the stop line of the target intersection in the current state is

If it is

The speed of the vehicle k needs to be adjusted

The specific method comprises the following steps:

1) estimating

And

if it is

Skipping to the step 2), otherwise, ending;

2) let the velocity value v of the vehicle k_k＝v_kΔ v, where Δ v is a small speed variation value preset, and the process jumps back to step 1).

In an optional implementation manner of this embodiment, in step S104, the step of planning the target speeds of the plurality of target vehicles passing through the target intersection according to the target driving sequence according to the maximum passing speed includes:

when all other target vehicles which are driven by the current target vehicle have completed the speed planning and the maximum passing speed is greater than the minimum value, initializing a candidate passing speed to the maximum passing speed, and executing the following steps:

a. whether the target vehicle conflicts with other vehicles after driving out of the target intersection at the candidate passing speed or not;

b. when no conflict occurs, determining that the current target speed of the target vehicle is equal to the candidate passing speed;

c. when a conflict occurs, subtracting a preset adjustment value from the candidate passing speed, and jumping to b when the subsequent passing speed is greater than the minimum value; and ending the execution when the subsequent passing speed is less than or equal to the minimum value.

In this alternative implementation, after the maximum passing speed of each target vehicle and the target driving sequence of the respective target vehicles are determined, a target speed that is not greater than the maximum passing speed and does not collide with other target vehicles may be determined.

FIG. 5 illustrates a diagram of the intuitive effect of planning a target speed according to one embodiment of the present disclosure. As shown in FIG. 5, the minimum passing speeds v of all target vehicles may be predetermined_minWith a maximum transit speed v_maxThen the speeds of all target vehicles need to satisfy:

v_min≤v_k≤v′_k≤v_max

wherein, v'_kThe maximum traffic speed of the target vehicle. v. of_kIs the target speed of the target vehicle.

The embodiment finds the maximum traffic speed without collision with other vehicles by adjusting the running speed in the feasible speed interval. The method comprises the following specific steps:

1) judging whether other target vehicles before the stop line from the target vehicle to the target intersection have all completed speed planning (namely whether the target speed is found out through the embodiment of the disclosure), if so, enabling v to be_tmp＝v′_kAnd jumping to step 2), otherwise v_k＝v′_kAnd then, the process is ended.

2) Judging whether v is_tmp＞v_minIf yes, jumping to step 3), otherwise ending the planning (the target vehicle can be at presentSpeed travel).

3) Predicting when v_k＝v_tmpThe planning result of the target vehicle k. Comparing the existing conflict occupation situation, and if the planning result conflicts with the existing occupation situation, if so, v_tmp＝v_tmpΔ v, jump to step 2), otherwise v_k＝v_tmpAnd then, the process is ended.

The following are embodiments of the disclosed apparatus that may be used to perform embodiments of the disclosed methods.

According to the multi-interface cooperative control device for the internet-oriented automatic driving, the device can be realized to be a part or all of electronic equipment through software, hardware or a combination of the software and the hardware. The multi-intersection cooperative control device for the internet automatic driving comprises:

The multi-interface cooperative control device for internet-oriented automatic driving in this embodiment corresponds to the multi-interface cooperative control method for internet-oriented automatic driving, and specific details can be referred to the description of the multi-interface cooperative control method for internet-oriented automatic driving, which is not described herein again.

Fig. 6 is a schematic structural diagram of an electronic device suitable for implementing the multi-intersection cooperative control method for internet-oriented automatic driving according to the embodiment of the present disclosure.

As shown in FIG. 6, electronic device 600 includes a processing unit 601 that may be implemented as a CPU, GPU, FPAG, NPU, or other processing unit. The processing unit 601 may perform various processes in the embodiments of any one of the above-described methods of the present disclosure according to a program stored in a Read Only Memory (ROM)602 or a program loaded from a storage section 608 into a Random Access Memory (RAM) 603. In the RAM603, various programs and data necessary for the operation of the electronic apparatus 600 are also stored. The CPU601, ROM602, and RAM603 are connected to each other via a bus 604. An input/output (I/O) interface 605 is also connected to bus 604.

The following components are connected to the I/O interface 605: an input portion 606 including a keyboard, a mouse, and the like; an output portion 607 including a display such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker; a storage section 608 including a hard disk and the like; and a communication section 609 including a network interface card such as a LAN card, a modem, or the like. The communication section 609 performs communication processing via a network such as the internet. The driver 610 is also connected to the I/O interface 605 as needed. A removable medium 611 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 610 as necessary, so that a computer program read out therefrom is mounted in the storage section 608 as necessary.

In particular, according to embodiments of the present disclosure, any of the methods described above with reference to embodiments of the present disclosure may be implemented as a computer software program. For example, embodiments of the present disclosure include a computer program product comprising a computer program tangibly embodied on a medium readable thereby, the computer program comprising program code for performing any of the methods of the embodiments of the present disclosure. In such embodiments, the computer program may be downloaded and installed from a network through the communication section 609, and/or installed from the removable medium 611.

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

The units or modules described in the embodiments of the present disclosure may be implemented by software or hardware. The units or modules described may also be provided in a processor, and the names of the units or modules do not in some cases constitute a limitation of the units or modules themselves.

As another aspect, the present disclosure also provides a computer-readable storage medium, which may be the computer-readable storage medium included in the apparatus in the above-described embodiment; or it may be a separate computer readable storage medium not incorporated into the device. The computer readable storage medium stores one or more programs for use by one or more processors in performing the methods described in the present disclosure.

The foregoing description is only exemplary of the preferred embodiments of the disclosure and is illustrative of the principles of the technology employed. It will be appreciated by those skilled in the art that the scope of the invention in the present disclosure is not limited to the specific combination of the above-mentioned features, but also encompasses other embodiments in which any combination of the above-mentioned features or their equivalents is possible without departing from the inventive concept. For example, the above features and (but not limited to) the features disclosed in this disclosure having similar functions are replaced with each other to form the technical solution.

Claims

1. A multi-interface cooperative control method for automatic internet driving comprises the following steps:

2. The method of claim 1, further comprising:

3. The method of claim 1 or 2, wherein determining the order of travel of the target vehicle through the target intersection under the state information using a deep Q learning network comprises:

4. The method of claim 1 or 2, wherein the deep Q learning network employs a convolutional neural network, further comprising:

5. The method of claim 4, wherein the neuron parameters are updated by a Q-value update formula as follows:

wherein, theta_iThe neuron parameters, α (t) and gamma, of the convolutional neural network are preset attenuation factors, theta'_iNeuron parameters for updated convolutional neural networkNumber, J is the set of all adjacent intersections, ω_i,jAnd (4) a conversion coefficient of the Q value of the adjacent intersection j to the Q value of the target intersection i, wherein omega is a penalty function.

6. The method of any of claims 1-2, 5, wherein determining a maximum transit speed for a plurality of the target vehicles from the status information comprises:

determining the adjusted first speed as the maximum passing speed.

7. The method of any of claims 1-2, 5, wherein planning a target speed for a plurality of the target vehicles to pass through the target intersection in the target driving order based on the maximum transit speed comprises:

8. A multi-interface cooperative control device for automatic internet driving comprises:

9. An electronic device, comprising a memory and a processor; wherein the content of the first and second substances,

the memory is to store one or more computer instructions, wherein the one or more computer instructions are to be executed by the processor to implement the method of any one of claims 1-7.

10. A computer readable storage medium having computer instructions stored thereon, wherein the computer instructions, when executed by a processor, implement the method of any of claims 1-7.