Unmanned driving system, method and vehicle with roadside perception cloud planning under mixed traffic flow

Abstract

The application discloses a self-driving system of roadside perception cloud planning under mixed traffic flow, which comprises a self-driving vehicle, roadside intelligent infrastructure and a decision planning cloud platform; the self-driving vehicle comprises a positioning module, a vehicle-mounted network communication terminal and a drive-by-wire chassis; the roadside intelligent infrastructure comprises a radar, a camera, road traffic signs, traffic signal lamps, an environmental information fusion calculation module and a roadside communication module; the decision planning cloud platform comprises an information storage module, an information processing module and a cloud communication module; the perception system is uniformly planned to a road section, the decision planning system is arranged in the cloud, the perception sensor and the calculation unit are shared, and the cost of the self-driving vehicle is reduced; a communication safety redundancy mechanism, a network switching method and a signal interruption avoidance mechanism are provided; and the reliability and safety of the self-driving system are effectively improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of unmanned driving of automobiles, and particularly relates to an unmanned driving system and method of roadside perception cloud planning under mixed traffic flow. BACKGROUND

[0002] With the increasing use of automobiles, traffic congestion often reduces user travel efficiency, leading to situations where drivers often encounter each other in the driving process and cause serious traffic accidents. Vehicle-road cooperation and cloud computing technology implement dynamic real-time information interaction of vehicle-vehicle, vehicle-road, vehicle-cloud and road-cloud in all directions, and on the basis of full-space dynamic traffic information collection and fusion, develop vehicle active safety control and road cooperative management, fully realize effective cooperation of man, vehicle and road, ensure traffic safety, and improve traffic efficiency, thereby forming an active safety system that is safer, more efficient and more environmentally friendly.

[0003] An unmanned vehicle is a vehicle that does not need to rely on human consciousness to make decisions and can dynamically and autonomously avoid a series of obstacles in the environment during driving. The vehicle relies on vehicle-mounted sensors to complete the perception and detection of the surrounding environment, generate map data, vehicle position and obstacle information, and realize autonomous control according to information integration to plan the best driving path. At the same time, the vehicle relies on artificial intelligence to recognize various traffic signs and make the best driving decisions to complete the driving task of the vehicle with high intelligence, reduce the accident rate and improve driving efficiency. However, the high cost of sensors and computing units seriously hinders the popularization of unmanned vehicles. How to reduce the hardware cost of unmanned vehicles has become the most pressing problem in the current unmanned vehicle industry. SUMMARY

[0004] To solve the above problems, the present application provides an unmanned driving system and method of roadside perception cloud planning under mixed traffic flow.

[0005] In view of the above defects or improvement needs of the prior art, the present application relates to an unmanned driving system of roadside perception cloud planning under mixed traffic flow, comprising an unmanned vehicle, roadside intelligent infrastructure arranged at a road section, and a decision planning cloud platform arranged in the cloud.

[0006] The unmanned vehicle comprises a positioning module for acquiring positioning and attitude information, a vehicle-mounted network communication terminal for data interaction and control of a drive-by-wire chassis, and a drive-by-wire chassis.

[0007] The roadside intelligent infrastructure comprises an information perception module for perceiving road traffic environment information, an environment information fusion calculation module for fusion processing of the perceived information, and a roadside communication module for data interaction with vehicles within the communication range and the decision planning cloud platform.

[0008] The decision planning cloud platform comprises an information storage module for storing road traffic environment information and vehicle state information, a processing module for planning path information and updating road traffic environment information, and a cloud communication module for data interaction between the roadside intelligent infrastructure and the unmanned vehicle.

[0009] The wire control chassis is connected with the vehicle-mounted network communication terminal, one output end of the wire control chassis is connected with one input end of the positioning module, one output end of the positioning module is connected with one input end of the vehicle-mounted network communication terminal, the decision planning cloud platform is connected with the vehicle-mounted network communication terminal through 5G communication, and the roadside intelligent infrastructure is connected with the vehicle-mounted network communication terminal through PC5 communication.

[0010] Further, the positioning module acquires vehicle position information and driving posture information and sends the information to the vehicle-mounted network communication terminal, the vehicle-mounted network communication terminal is used for data interaction between the unmanned vehicle and other vehicles, the roadside intelligent infrastructure and the decision planning cloud platform, and can control the wire control chassis through the CAN bus.

[0011] Further, the information perception module perceives road traffic environment information within the range of the roadside intelligent infrastructure and sends the perceived information to the environment information fusion calculation module for fusion processing; the environment information fusion calculation module is used for processing road traffic environment information perceived by the radar and the camera and sending the information to the roadside communication module.

[0012] Further, the information storage module records and saves road traffic environment information uploaded by the roadside intelligent infrastructure and state information uploaded by the vehicle; the information processing module plans global driving path and local driving path of the unmanned vehicle in the mixed traffic flow and constructs a mathematical model to dynamically update the recorded road traffic environment information in the information storage module.

[0013] As another aspect of the application, the application also relates to an unmanned driving method based on roadside perception cloud planning in a mixed traffic flow, comprising the following steps:

[0014] Step one: start the unmanned driving mode of the vehicle, connect the vehicle-mounted network communication terminal and the decision planning cloud platform and match user information; meanwhile, send a signal of successful access to the vehicle-mounted network terminal and the roadside intelligent infrastructure;

[0015] Step two: after the connection is successful, the vehicle enters a self-checking mode to check whether each system of the vehicle is normally working and whether the connection between the network communication terminal and the decision planning cloud platform is stable.

[0016] Step 3: After the self-test mode ends, the user is prompted to enter the destination address. After the user enters the address and confirms it, the vehicle-mounted network communication terminal sends the vehicle's basic status information and driving status information to the decision-making and planning cloud platform.

[0017] Step 4: After the roadside intelligent infrastructure is built, it is connected to the decision-making and planning cloud platform. Depending on whether there is a path with full coverage of roadside intelligent infrastructure, the decision-making and planning cloud platform plans the global path of the autonomous vehicle based on the current location information and destination address of the autonomous vehicle.

[0018] Step 5: The decision-making and planning cloud platform prioritizes 5G communication and uses roadside intelligent infrastructure as a secondary priority to send control signals to the vehicle-mounted network communication terminal for autonomous vehicle control.

[0019] Step Six: During vehicle operation, the decision cloud platform works in conjunction with several roadside intelligent infrastructures along the vehicle's path to prevent loss of control signals in a short period of time.

[0020] Step 7: During vehicle operation, a safety redundancy braking mechanism ensures safety. In case of an unexpected event, the on-board network terminal controls the drive-by-wire chassis to pull over to the side of the road; otherwise, the remaining driving tasks are executed.

[0021] Furthermore, the specific method of step four is as follows: The decision planning cloud platform finds all possible paths between the current location information and the destination address of the autonomous vehicle as a path information set B(x). It then filters out all roadside intelligent infrastructures covered by the paths in B(x) and activates them. The roadside intelligent infrastructures upload real-time traffic environment information within their perception range. Next, it selects a path information set fully covered by the roadside intelligent infrastructure: B1(x) = {1, 2, ..., n}. If no path fully covered by the roadside intelligent infrastructure exists, the decision planning cloud platform remotely controls the vehicle to drive to the area covered by the roadside intelligent infrastructure and then returns to the autonomous driving mode to perform subsequent driving tasks. Alternatively, it reminds the passenger to drive the vehicle to the infrastructure coverage area and then return to the autonomous driving mode to perform subsequent driving tasks. Finally, the decision planning cloud platform remotely controls the autonomous vehicle to perform this driving task according to the optimal global path in B1(x).

[0022] Furthermore, the method for determining the optimal global path is as follows: the decision planning cloud platform segments all planned global paths according to roads based on user needs, and calculates the time cost f required to traverse the nth road in the path. n (n), that is, the total time cost of the nth travel route is D(n) = f n (1)+f n (2)+…+fn (n), the path with the minimum D(n) value is the optimal global path.

[0023] Further, on the road without unmanned vehicles passing through, the roadside intelligent infrastructure of the road is in a dormant state, and the roadside intelligent infrastructure of the road is awakened when the unmanned vehicle passes through the road.

[0024] Further, the time cost f n (n) is calculated according to the predicted road environment information when the vehicle travels to the road. The predicted road traffic environment information is obtained by the decision planning cloud platform collecting the road traffic and the traffic environment information of the road.

[0025] Further, the method of step five is specifically: the control signal of the decision cloud platform can also be forwarded to the vehicle-mounted network communication terminal through the roadside intelligent infrastructure; the linkage between the vehicle-mounted network communication terminal and the roadside intelligent infrastructure is controlled by the decision planning cloud platform as a whole, after the optimal driving path is selected by the path planning, the decision planning cloud platform numbers the IP address or other unique mark of the roadside intelligent infrastructure of the path according to the order of the vehicle passing through, and then forms the roadside intelligent infrastructure information set D(x) and sends it to the vehicle-mounted network communication terminal. When the 5G communication is normal, the received control signal of the decision cloud platform is executed, and when the control signal of the decision cloud platform is not received within the preset time ΔT, the control signal sent by the roadside intelligent infrastructure is executed.

[0026] Further, in step six, the method for avoiding the loss of control signals in a short time through the cooperation of the decision cloud platform and the several roadside intelligent infrastructures that the vehicle will pass through is: because the communication distance of the roadside intelligent infrastructure is limited, the vehicle-mounted network terminal will drive away from a certain roadside intelligent infrastructure coverage range during driving. In order to avoid the frequency change of the roadside intelligent infrastructure connected with the vehicle-mounted network communication terminal leading to the loss of control signals in a short time, the decision cloud platform will send the control signal to the several roadside intelligent infrastructures that the vehicle will pass through. The above roadside intelligent infrastructures will continuously broadcast the control signal sent by the decision cloud platform, and the vehicle-mounted network communication terminal can receive the control signal sent by the decision cloud platform during the whole driving process. When the vehicle drives away from the roadside intelligent infrastructure, the decision planning cloud platform identifies the vehicle position, and then controls the roadside intelligent infrastructure to stop sending the control signal. If there is no other unmanned vehicle passing through the roadside intelligent infrastructure, a dormant signal is sent to it to reduce the energy consumption of the system and improve the service life of the system.

[0027] Further, the safety redundancy braking mechanism of step seven is that the vehicle joins the safety redundancy braking mechanism, when a special accident occurs, the vehicle-mounted network terminal controls the chassis to stop by the roadside, the vehicle-mounted network terminal obtains the lateral distance of the vehicle from the roadside at this moment, the information roadside intelligent infrastructure monitors the lateral distance signal in real time through the sensor carried thereon, and then broadcasts the control signal of the decision planning cloud platform to the vehicle-mounted network communication terminal. When the communication terminal or other uncontrollable faults occur, the vehicle-mounted network communication terminal plans a driving path for the vehicle to drive to the roadside, controls the vehicle to stop by the roadside, and opens the danger alarm lamp and broadcasts externally before executing the roadside parking task, so as to avoid collision with the vehicle that is about to pass through the road section.

[0028] As another aspect of the present application, it also relates to a vehicle comprising the above-mentioned mixed traffic flow roadside perception cloud planning autonomous driving system.

[0029] Overall, compared with the prior art, the above technical solutions conceived by the present application can achieve the following beneficial effects:

[0030] (1) The mixed traffic flow roadside perception cloud planning autonomous driving system and method of the present application uses 5G communication, vehicle-road cooperation and cloud computing technology to uniformly plan the perception system to the road section, arranges the decision planning system in the cloud, realizes sharing of the perception sensor and the computing unit, and solves the problem that high sensor and computing unit costs seriously hinder the popularization of autonomous vehicles;

[0031] (2) The mixed traffic flow roadside perception cloud planning autonomous driving system and method of the present application uses 5G wireless communication and PC5 wireless communication to send control signals, adds a communication safety redundancy mechanism that gives priority to 5G communication and gives secondary priority to the retransmission of the roadside intelligent infrastructure, and gives a corresponding network switching method; effectively avoids traffic accidents of autonomous vehicles caused by network communication quality;

[0032] (3) The mixed traffic flow roadside perception cloud planning autonomous driving system and method of the present application avoids control signal interruption caused by limited coverage distance of the roadside intelligent infrastructure through cooperation of the decision cloud platform and the several roadside intelligent infrastructures that the vehicle driving path will pass through, and improves the reliability of the system;

[0033] (4) The mixed traffic flow roadside perception cloud planning autonomous driving system and method of the present application ensures the safety of autonomous vehicles through the safety redundancy braking mechanism. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1The overall architecture of the autonomous driving system under the roadside perception cloud planning in the mixed traffic flow according to the preferred embodiment of the present application;

[0035] Figure 2 The communication schematic diagram of the autonomous vehicle-decision planning cloud platform according to the preferred embodiment of the present application;

[0036] Figure 3 The management schematic diagram of the decision planning cloud platform to the roadside intelligent infrastructure (RSU) according to the preferred embodiment of the present application;

[0037] Figure 4 The schematic diagram of the RRT algorithm tree structure generation process according to the preferred embodiment of the present application;

[0038] Figure 5 The connection schematic diagram of the vehicle-mounted network communication terminal and the roadside intelligent infrastructure according to the preferred embodiment of the present application;

[0039] Figure 6 The management schematic diagram of the decision planning cloud platform to the vehicle-mounted network communication terminal and the roadside intelligent infrastructure according to the preferred embodiment of the present application;

[0040] Figure 7 The schematic diagram of the temporary parking of the vehicle under the safety redundancy mechanism according to the preferred embodiment of the present application;

[0041] Figure 8 The overall logic flow chart of the autonomous driving vehicle under the roadside perception cloud planning in the mixed traffic flow according to the preferred embodiment of the present application; DETAILED DESCRIPTION

[0042] In order to make the purpose, technical scheme and advantages of the present application clearer and more apparent, the present application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application. In addition, the technical features involved in each embodiment of the present application described below can be combined with each other as long as they do not conflict with each other.

[0043] Embodiment 1:

[0044] Please refer to Figure 1 The present embodiment relates to an autonomous driving system under the roadside perception cloud planning in the mixed traffic flow, which comprises three parts of autonomous vehicle, roadside intelligent infrastructure (Road Side Unit, abbreviated as RSU) and decision planning cloud platform (Central Service Unit, abbreviated as CSU, also known as central service unit).

[0045] The unmanned vehicle includes a positioning module, an on-board network communication terminal (OBU, also known as a vehicle-mounted unit), and a drive-by-wire chassis. The positioning module is used to obtain the vehicle position information and driving attitude information and send them to the on-board network communication terminal. The on-board network communication terminal is used for data interaction between the unmanned vehicle and other vehicles, roadside intelligent infrastructure, and a decision-making and planning cloud platform, and can control the drive-by-wire chassis through the CAN bus. The drive-by-wire chassis refers to a vehicle drive, gear, brake, steering, parking, and necessary indicator light that can be controlled through the CAN bus, and the vehicle can give correct and timely state feedback. The roadside intelligent infrastructure includes an information perception module (such as a radar and a camera), an environmental information fusion computing module, and a communication module. The radar and the camera are used to perceive the road traffic environment information within a certain range of the roadside intelligent infrastructure and send the perceived information to the environmental information fusion computing module for fusion processing. The environmental information fusion computing module is used to process the road traffic environment information perceived by the radar and the camera and send it to the communication module. The communication module is used for data interaction with vehicles and a decision-making and planning cloud platform within the communication range. The roadside intelligent infrastructure can also include road traffic signs, which refer to speed limit or warning signs, used to remind drivers of non-network function vehicles in a mixed traffic flow, and traffic lights are used to remind drivers of non-network function vehicles in a mixed traffic flow,

[0046] The decision-making and planning cloud platform includes an information storage module, an information processing module, and a communication module. The information storage module is used to record and save the road traffic environment information uploaded by the roadside intelligent infrastructure and the state information uploaded by the vehicle. The information processing module is used to plan the global driving path and the local driving path of the unmanned vehicle in a mixed traffic flow and to construct a mathematical model to dynamically update the road traffic environment information recorded in the information storage module. The communication module is used for data interaction with the roadside intelligent infrastructure and the unmanned vehicle.

[0047] The drive-by-wire chassis and the on-board network communication terminal are connected (such as through a CAN interface). One output end of the drive-by-wire chassis is connected to one input end of the positioning module. One output end of the positioning module is connected to one input end of the on-board network communication terminal. The decision-making and planning cloud platform is connected to the on-board network communication terminal through 5G communication. The roadside intelligent infrastructure is connected to the on-board network communication terminal through PC5 communication. The decision-making and planning cloud platform is connected to the on-board network communication terminal through hard-wire or 5G communication.

[0048] Based on the above system, the embodiment also relates to an unmanned driving method based on roadside perception and cloud planning in a mixed traffic flow, which specifically includes the following five stages.

[0049] Please refer to Figure 2, the first stage starts the vehicle unmanned mode, the vehicle-mounted networked communication terminal OBU applies to establish a connection with the decision planning cloud platform, sends the user's unique ID / password to the decision planning cloud platform CSU, the decision planning cloud platform CSU matches the received user information in the database, creates the user's session information after successful matching, and sends a signal to the vehicle-mounted networked terminal that the access is successful. After the connection is successful, the vehicle enters a self-checking mode, checks whether each system of the vehicle is working normally and whether the networked communication terminal is connected with the decision planning cloud platform stably (whether the communication quality is good), reminds the user to input the destination address after the self-checking is completed, and sends the basic state information and the driving state information of the vehicle to the decision planning cloud platform after the user inputs the address and confirms.

[0050] Please refer to Figure 3 , the second stage finds all possible paths between the current position information and the destination address of the unmanned vehicle as the total set B(x) of path information, and selects all road side intelligent infrastructures covered by the information set path in B(x) and wakes them up. Usually, after the road side intelligent infrastructure is built, it establishes a connection with the decision planning cloud platform. The connection mode is that the road side intelligent infrastructure first sends the user's unique ID / password to the decision planning cloud platform, the decision planning cloud platform matches the received RSU information in the database, and creates the session information of the road side intelligent infrastructure after successful matching, and sends a signal to the road side intelligent infrastructure that the access is successful.

[0051] As a preferred scheme, in order to reduce the energy consumption of the system and increase the service life of the system, when there is no unmanned vehicle passing through the road section, the road side intelligent infrastructure of the road section is usually in a dormant state, and when there is an unmanned vehicle passing through the road section, the road side intelligent infrastructure of the road section is woken up,

[0052] The road side intelligent infrastructure uploads the real-time traffic environment information in its sensing range. The path information set B1(x)={1, 2, …, n} covered by the road side intelligent infrastructure is selected. If there is no path covered by the road side intelligent infrastructure, the decision cloud platform controls the vehicle to drive to the road side intelligent infrastructure coverage area by remote driving and then restores to the unmanned mode to perform the subsequent driving task, or reminds the passenger to drive the vehicle to the infrastructure coverage area and then restores to the unmanned mode to perform the subsequent driving task. The decision planning cloud platform segments all global paths according to roads according to user demand, and calculates the time cost f n (n) paid by the nth road of the nth driving path calculated by the scheme. n(n) is not calculated according to the road environment information at the system planning moment, but is calculated according to the predicted road environment information when the vehicle travels to the road, the predicted road traffic environment information is summarized by the decision planning cloud platform in real time collecting the traffic flow of the road and the traffic environment information, and some objective factors (including but not limited to weather, whether it is commuting, season, whether there is a traffic accident, whether it is a holiday, etc.) affecting the traffic environment are considered, so as to further ensure the accuracy of the prediction, that is, the total time cost of the nth travel path is D(n) = f n (1) + f n (2) +…+ f n (n), the path with the minimum D(n) value is the optimal global path, and then the decision planning cloud platform remotely controls the unmanned vehicle to perform the travel task according to the path.

[0053] Please refer to Figure 4 As a preferred scheme: the above finding all possible paths between the current position information of the unmanned vehicle and the destination address and finding the optimal global path can adopt the rapid expansion random tree algorithm, the rapid expansion random tree (Rapidly-exploring Random Tree, RRT) algorithm continuously constructs a tree search structure in the multi-dimensional space according to the random increment to the unexplored blank area, each tree structure vertex is a state node, and the line segment between two adjacent nodes represents the connection process between the current state node and the last state node.

[0054] Figure 4 For the tree structure generation process of the RRT algorithm, first define the global path planning task space n represents the space dimension. The task space can be divided into a space with obstacles and a blank area Let the initial state point of the environment be the target point be and the obstacles in the environment be Figure 4 (black area).

[0055] Let P start be the root node of the entire tree structure, the algorithm generates a random point P rand in the unexplored area, takes P rand as the center, traverses all nodes on the tree structure, calculates the Euclidean distance between these nodes and P rand , and sorts them, and selects the node P near with the minimum distance as the nearest node.

[0056] Let P near be the nearest node, and P randThe direction of the point is the growth direction of the tree structure. By setting the growth step (denoted as l) of the algorithm, the search tree grows from P near The point grows a certain distance (the value of l needs to be selected through experiments. A too large growth step will make the algorithm "jump" over obstacles with small cross sections. A too short step will often weaken the search ability of the algorithm and reduce the planning efficiency of the algorithm) along the growth direction to obtain the next node on the tree structure, denoted as P1. It is judged whether P1 is within the range of the obstacle. If it collides with the obstacle, the node is removed, and random point sampling search is performed again. If no collision occurs, P1 is added to the tree. The algorithm is iterated until the newly generated node reaches the target position, or the distance between the nodes is less than a unit growth step l, indicating the end of the search process. The algorithm backtracks a series of parent nodes from the target point in order to obtain the final path.

[0057] Please refer to Figure 5 , the third stage decision planning cloud platform CSU sends control signals to the vehicle-mounted network communication terminal OBU for unmanned vehicle control. In order to avoid the problem of delay or loss of control signals due to fluctuations in 5G signals, which may cause the unmanned vehicle to lose control and collide, the control signals of the decision cloud platform can also be forwarded to the vehicle-mounted network communication terminal through the roadside intelligent infrastructure. To ensure the communication safety of the entire communication link, the link between the vehicle-mounted network communication terminal and the roadside intelligent infrastructure is controlled by the decision cloud platform. After the path planning is completed and the optimal driving path is selected, the decision planning cloud platform numbers the IP addresses or other unique marks of the roadside intelligent infrastructure of the path according to the order of vehicle passing, and then forms the roadside intelligent infrastructure information set D(x) and sends it to the vehicle-mounted network communication terminal. When the 5G communication is normal, the received control signals of the decision cloud platform are executed. When no control signals of the decision cloud platform are received for a certain time ΔT, the control signals sent by the roadside intelligent infrastructure are executed.

[0058] Please refer to Figure 6, the fourth stage when the vehicle is executing the control signal sent by the roadside intelligent infrastructure, since the communication distance of the roadside intelligent infrastructure RSU is the farthest limit (generally not more than 1000 meters), the vehicle will soon drive out of the coverage range of a roadside intelligent infrastructure during driving, in order to avoid the loss of control signals caused by the frequency handover of the roadside intelligent infrastructure connected with the vehicle-mounted networked terminal, the method provided in the scheme is that the decision cloud platform will issue the control signal to several roadside intelligent infrastructures that the vehicle will pass through, and the above roadside intelligent infrastructures will continuously broadcast the control signal issued by the decision cloud platform, so that the vehicle-mounted networked terminal can receive the control signal issued by the decision cloud platform CSU during the whole driving process, when the vehicle drives out of the roadside intelligent infrastructure, the decision planning cloud platform CSU can identify the position of the vehicle, and then control the roadside intelligent infrastructure to stop sending the control signal, if no other autonomous vehicle passes through the roadside intelligent infrastructure, a hibernation signal is sent to the roadside intelligent infrastructure to reduce the energy consumption of the system and improve the service life of the system.

[0059] Please refer to Figure 7 , the fifth stage is to further ensure the driving safety of the autonomous vehicle, the vehicle-mounted networked terminal can control the chassis-by-wire to park by the roadside in the case of special accidents, so as to realize the function, and the vehicle-mounted networked terminal needs to obtain the lateral distance of the vehicle from the roadside at this moment, so that the roadside intelligent infrastructure can monitor the information in real time through the sensor carried by the roadside intelligent infrastructure, and then broadcast the control signal of the decision planning cloud platform to the vehicle-mounted networked communication terminal, when the communication terminal or other uncontrollable faults, the vehicle-mounted networked communication terminal plans the driving path of the vehicle to the roadside, controls the vehicle to park by the roadside, and opens the danger alarm lamp and broadcasts externally before executing the task of parking by the roadside, so as to avoid the collision with the vehicle that will pass through the road section.

[0060] Embodiment 2 of the present application provides a vehicle comprising the autonomous driving system under the roadside perception cloud planning of mixed traffic flow provided in embodiment 1.

[0061] Embodiment 2 of the present application provides a vehicle comprising the autonomous driving system under the roadside perception cloud planning of mixed traffic flow provided in embodiment 1.

[0062] The technical scheme of the present application fully utilizes 5G communication, vehicle-road cooperation and cloud computing technology, etc., uniformly plans the perception system to the road section, arranges the decision planning system in the cloud, realizes the sharing of the perception sensor and the computing unit, and solves the problem that the high cost of the sensor and the computing unit seriously hinders the popularization of the autonomous vehicle.

[0063] Those skilled in the art can understand that the above description is only the preferred embodiment of the present application, and is not intended to limit the present application. Any modification, equivalent replacement and improvement made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. An autonomous driving method based on roadside perception and cloud-based planning under hybrid traffic flow, characterized in that, The steps include: Step 1: Activate the vehicle's autonomous driving mode, and connect the onboard network communication terminal and the decision-making and planning cloud platform. User information matching; simultaneously, a successful access signal is sent to the vehicle-mounted connected terminal and roadside intelligent infrastructure; Step 2: After successful connection, the vehicle enters self-test mode to check whether the vehicle's systems are working properly and whether the connection between the network communication terminal and the decision-making and planning cloud platform is stable. Step 3: After the self-test mode ends, the user is prompted to enter the destination address. After the user enters the address and confirms it, the vehicle-mounted network communication terminal sends the vehicle's basic status information and driving status information to the decision-making and planning cloud platform. Step 4: After the roadside intelligent infrastructure is built, it is connected to the decision-making and planning cloud platform. Depending on whether there is a path with full coverage of roadside intelligent infrastructure, the decision-making and planning cloud platform plans the global path of the autonomous vehicle based on the current location information and destination address of the autonomous vehicle. Step 5: The decision-making and planning cloud platform prioritizes 5G communication and uses roadside intelligent infrastructure as a secondary priority to send control signals to the vehicle-mounted network communication terminal for autonomous vehicle control. Step Six: During vehicle operation, the decision cloud platform works in conjunction with several roadside intelligent infrastructures along the vehicle's path to prevent loss of control signals in a short period of time. Step 7: During vehicle operation, a safety redundancy braking mechanism ensures safety. In case of an unexpected event, the on-board network terminal controls the drive-by-wire chassis to pull over to the side of the road; otherwise, the remaining driving tasks are executed. The specific method of step four is as follows: The decision planning cloud platform finds all possible paths between the current location information and the destination address of the autonomous vehicle as a path information set B(x). In B(x), it filters out all roadside intelligent infrastructures covered by the path information set and wakes them up. The roadside intelligent infrastructures upload real-time traffic environment information within their perception range. Then, it selects the path information set fully covered by the roadside intelligent infrastructure: B1(x) = {1, 2, ..., n}. If there is no path fully covered by the roadside intelligent infrastructure, the decision cloud platform controls the vehicle to drive to the area covered by the roadside intelligent infrastructure through remote driving and then restores the autonomous driving mode to perform subsequent driving tasks, or reminds the passenger to drive the vehicle to the area covered by the infrastructure and then restores the autonomous driving mode to perform subsequent driving tasks. Then, the decision planning cloud platform remotely controls the autonomous vehicle to execute the driving task in B1(x) according to the optimal global path.

2. The unmanned driving method based on roadside perception and cloud-based planning under mixed traffic flow as described in claim 1, characterized in that: The unmanned vehicle includes a positioning module for acquiring positioning and attitude information, an on-board network communication terminal for data interaction and control of the drive-by-wire chassis, and a drive-by-wire chassis. The roadside intelligent infrastructure includes an information sensing module for sensing road traffic environment information, an environmental information fusion computing module for fusing and processing the sensed information, and a roadside communication module for data interaction with vehicles within the communication range and the decision-making and planning cloud platform. The decision-making and planning cloud platform includes an information storage module for storing road traffic environment information and vehicle status information, a processing module for planning route information and updating road traffic environment information, and a cloud communication module for data interaction between roadside intelligent infrastructure and autonomous vehicles. The drive-by-wire chassis is connected to the vehicle-mounted network communication terminal. One output of the drive-by-wire chassis is connected to one input of the positioning module. One output of the positioning module is connected to one input of the vehicle-mounted network communication terminal. The decision-making and planning cloud platform is connected to the vehicle-mounted network communication terminal via 5G communication. The roadside intelligent infrastructure is connected to the vehicle-mounted network communication terminal via PC5 communication. The decision-making and planning cloud platform is connected to the vehicle-mounted network communication terminal via hardwired or 5G communication.

3. The unmanned driving method based on roadside perception and cloud-based planning under mixed traffic flow as described in claim 2, characterized in that, The positioning module acquires the vehicle's location information and driving posture information and sends them to the vehicle-mounted network communication terminal. The vehicle-mounted network communication terminal is used for data interaction between the unmanned vehicle and other vehicles, roadside intelligent infrastructure, and decision-making and planning cloud platform, and controls the drive-by-wire chassis.

4. The unmanned driving method based on roadside perception and cloud-based planning under mixed traffic flow as described in claim 2, characterized in that, The information sensing module senses the road traffic environment information within the scope of the roadside intelligent infrastructure and sends the sensed information to the environmental information fusion computing module for fusion processing; the environmental information fusion computing module is used to process the road traffic environment information sensed by radar and cameras and send it to the roadside communication module.

5. The unmanned driving method based on roadside perception and cloud planning under mixed traffic flow as described in claim 2, characterized in that, The information storage module records and saves road traffic environment information uploaded by roadside intelligent infrastructure and vehicle status information uploaded by vehicles; the information processing module plans the global and local driving paths of unmanned vehicles under mixed traffic flow and constructs mathematical models to dynamically update the road traffic environment information recorded in the information storage module.

6. The unmanned driving method based on roadside perception and cloud planning under mixed traffic flow as described in claim 5, characterized in that: The method for determining the optimal global path is as follows: The decision planning cloud platform divides all planned global paths into segments according to roads based on user needs, calculates the time cost fn(n) required to pass through the nth road in the path, that is, the total time cost of the nth travel path is D(n) = fn(1) + fn(2) + ... + fn(n), and the path with the smallest D(n) value is the optimal global path.

7. The unmanned driving method based on roadside perception and cloud-based planning under mixed traffic flow as described in claim 5, characterized in that: On roads where no autonomous vehicles are passing, the roadside intelligent infrastructure is in a dormant state, and is activated when an autonomous vehicle passes through the road.

8. The unmanned driving method based on roadside perception and cloud planning under mixed traffic flow as described in claim 6, characterized in that: The time cost fn(n) required for the nth route is calculated based on the predicted road environment information when the vehicle travels to that road. The predicted road traffic environment information is obtained by the decision planning cloud platform by collecting and summarizing road traffic flow and traffic environment information in real time.

9. The unmanned driving method based on roadside perception and cloud-based planning under mixed traffic flow as described in claim 4, characterized in that: The method in step five is as follows: the control signal of the decision cloud platform is forwarded to the vehicle-mounted network communication terminal through the roadside intelligent infrastructure; the connection between the vehicle-mounted network communication terminal and the roadside intelligent infrastructure is controlled by the decision cloud platform. After the optimal driving route is selected after the route planning is completed, the decision planning cloud platform numbers the IP address or other unique identifier of the roadside intelligent infrastructure along the route according to the order in which the vehicles pass, and then forms a roadside intelligent infrastructure information set D(x) and sends it to the vehicle-mounted network communication terminal. When 5G communication is normal, the received control signal from the decision cloud platform is executed. If no control signal from the decision cloud platform is received within a preset time ΔT, the control signal sent by the roadside intelligent infrastructure is executed.

10. The unmanned driving method based on roadside perception and cloud-based planning under mixed traffic flow as described in claim 4, characterized in that, The specific method in step six is ​​as follows: During the vehicle's journey, the on-board connected terminal will leave the coverage area of ​​a certain roadside intelligent infrastructure. The decision cloud platform will send control signals to several roadside intelligent infrastructures that the vehicle will pass through along its route. These roadside intelligent infrastructures will continuously broadcast the control signals sent by the decision cloud platform. Throughout the entire journey, the on-board connected communication terminal can receive the control signals sent by the decision cloud platform. When the vehicle leaves the roadside intelligent infrastructure, the decision planning cloud platform will identify the vehicle's location and then control the roadside intelligent infrastructure to stop sending control signals. If no other autonomous vehicles pass through the roadside intelligent infrastructure, a dormant signal will be sent to it.

11. The autonomous driving method for roadside perception and cloud-based planning under mixed traffic flow as described in claim 5, characterized in that... The safety redundancy braking mechanism in step seven is as follows: In the event of a special accident, the vehicle-mounted network terminal controls the drive-by-wire chassis to pull over to the side of the road. The vehicle-mounted network terminal obtains the lateral distance of the vehicle from the roadside at that moment. The information roadside intelligent infrastructure monitors the lateral distance signal in real time through its onboard sensors, and then broadcasts it to the vehicle-mounted network communication terminal along with the control signal from the forwarding decision-making and planning cloud platform. When encountering communication terminal or other uncontrollable faults, the vehicle-mounted network communication terminal plans the driving path of the vehicle to the roadside, controls the vehicle to pull over to the side of the road, and first turns on the hazard warning lights and broadcasts to the outside before performing the pulling over task.

12. A vehicle, characterized in that, The vehicle includes the autonomous driving method for roadside perception and cloud planning under mixed traffic flow as described in any one of claims 1-4.

Images