A decision-making and planning method for connected autonomous driving vehicles in expressway diversion areas

Abstract

The present invention discloses a decision-making and planning method for networked autonomous driving vehicles in expressway diversion areas, comprising: 1. establishing a basic scenario for the application of the method; 2. constructing a fuzzy logic reasoning system to output a lane change willingness coefficient and then formulate a decision plan; 3. using the remaining time to collision (TTC) for risk assessment; 4. introducing the concept of risk field to model the vehicle potential field, road boundary line potential field, lane dividing line potential field, and static roadblock potential field, and then constructing a multi-source superposition risk potential field model for the expressway diversion area scenario by superimposing the various risk potential fields; 5. using a quintic polynomial to generate a lane change trajectory curve; 6. constructing a multi-objective comprehensive evaluation function that considers safety, efficiency, and comfort, and screening the optimal lane change trajectory for the networked autonomous driving vehicle. The present invention can endow autonomous driving vehicles with intelligent decision-making power and accurate path planning capabilities in response to changing traffic conditions, thereby effectively alleviating traffic congestion and improving road capacity.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent traffic safety, and in particular to the field of decision-making and planning of networked autonomous driving vehicles in expressway diversion areas, and specifically to a decision-making and planning method for networked autonomous driving vehicles in expressway diversion areas. Background Art

[0002] With the rapid development of intelligent transportation systems, autonomous driving technology has become a hot topic in modern transportation research. In complex urban traffic environments, freeway diversion zones serve as critical nodes connecting different traffic flows, and their traffic conditions significantly impact the operational efficiency of the entire road network. However, existing decision-making and planning methods for autonomous vehicles in freeway diversion zones have several limitations.

[0003] Traditional autonomous vehicles rely primarily on pre-set routes and fixed rules for navigation, lacking the ability to dynamically respond to real-time traffic conditions. This is particularly evident in expressway diversion zones, where traffic flows fluctuate frequently. This prevents vehicles from making timely optimal decisions, thus impacting traffic efficiency. Furthermore, the traffic environment in expressway diversion zones is complex and ever-changing, including multiple lane changes, emergency response, and weather conditions. Existing autonomous vehicles often rely on complex algorithms and extensive computing resources to handle these complex situations, which not only increases system complexity but also limits the real-time nature of decision-making.

[0004] Therefore, the present invention proposes a decision-making and planning method for networked autonomous driving vehicles in expressway diversion areas, aiming to achieve rapid response to the complex traffic environment of expressway diversion areas and provide autonomous driving vehicles with a safer and more efficient driving experience. Summary of the Invention

[0005] In order to address the shortcomings of the above-mentioned existing technologies, the present invention proposes a decision-making and planning method for networked autonomous driving vehicles in expressway diversion areas, aiming to cope with changing traffic conditions and give autonomous driving vehicles intelligent decision-making power and precise path planning capabilities, thereby ensuring that autonomous driving vehicles can achieve safe and efficient driving even in complex traffic environments, thereby effectively alleviating traffic congestion and improving road capacity.

[0006] In order to achieve the above-mentioned object, the present invention adopts the following technical solutions:

[0007] A method for decision-making and planning for a networked autonomous driving vehicle in an expressway diverging area according to the present invention is characterized in that it is applied to a one-way three-lane expressway diverging area road, and the three lanes in the same direction on the expressway diverging area road are respectively designated as a first lane, a second lane, and a third lane. The first lane is provided with a static roadblock, and an exit ramp is provided outside the third lane. The one-way three-lane road contains at least one networked autonomous driving vehicle and one manually driven vehicle. A plane coordinate system is established with a point L0 upstream of the exit ramp as the origin, the vehicle's travel direction as the positive direction of the X-axis, and the direction perpendicular to the X-axis as the Y-axis. The method for planning lane change trajectories of the networked autonomous driving vehicle is performed in accordance with the following steps:

[0008] Step 1: Use fuzzy logic reasoning to make lane change decisions for connected autonomous vehicles;

[0009] Step 1.1: Construct a fuzzy logic system, including input layer, fuzzy layer, rule layer, and output layer.

[0010] Assume that the expected speed of the connected autonomous vehicle is , and the expected vehicle distance is , the input layer includes: speed difference coefficient membership function , membership function of vehicle distance expectation coefficient , as shown in formulas (1) and (2) respectively;

[0011]

[0012] In formula (1), is the speed of connected autonomous vehicles, The speed of manually driven vehicles;

[0013]

[0014] In formula (2), The distance between the connected autonomous vehicle and the manually driven vehicle;

[0015] The fuzzy layer includes: velocity difference coefficient fuzzy subset {small, smaller, medium, larger, large}, fuzzy subset of vehicle distance expectation coefficient {small, smaller, medium, larger, large}, fuzzy subset of lane change intention {weak, weaker, medium, stronger, strong};

[0016] The rule layer includes: lane change fuzzy rules ;

[0017] The output layer includes: lane change willingness coefficient ;

[0018] Step 1.2: Set a decision plan , , , the specific implementation decisions are as follows:

[0019] :If the lane change willingness coefficient satisfy: , then the decision plan is to keep going straight;

[0020] :If the lane change willingness coefficient satisfy: , then the decision plan is to wait for lane change and go to step 2;

[0021] :If the lane change willingness coefficient satisfy: , then the decision is to change lanes immediately and go to step 3. , , are all constants;

[0022] Step 2: Calculate the CAV of any connected autonomous vehicle according to formula (3): j Compared with manual driving vehicles HV h Remaining collision time , and conduct risk assessments;

[0023]

[0024] The risk assessment methodology described is as follows:

[0025] like , it is determined that there is a collision risk in the scene, and the decision is changed to keep going straight;

[0026] like , it is determined that there is no collision risk in the scenario, and the vehicle changes lanes and proceeds to step 3;

[0027] Step 3: Use formula (4) to construct the human-driven vehicle HV h , road boundary line, lane dividing line, static roadblock at any point in the plane coordinate system Multi-source superposition risk field model ;

[0028]

[0029] In formula (4), , , , Any point Vehicle potential field at , road boundary line potential field , lane dividing line potential field , static roadblock potential field The weight parameter of

[0030] Step 4: Assume any connected autonomous vehicle CAV j The starting point of lane change is A j , the lane change destination is F j , and the fifth-order polynomial curve fitting method is used to find the lane-changing starting point A j and lane change end point F j Generate a connected autonomous vehicle (CAV) j The lane-changing trajectory curve cluster is a cluster of m lane-changing trajectory curves, and the cluster contains m lane-changing trajectory curves. , , , let the jth lane change trajectory curve The polynomial fitting curve equation is , 1≤j≤m;

[0031] Step 5: Construct the j-th lane change trajectory curve using formula (5) Multi-objective comprehensive evaluation function ;

[0032]

[0033] In formula (5), , , The safety evaluation function , efficiency evaluation function , comfort evaluation function The weight parameter of

[0034] Step 6: Use formula (5) to solve the multi-objective comprehensive evaluation value of m trajectory curves, and select the lane change trajectory curve corresponding to the minimum comprehensive target evaluation value That is, connected autonomous vehicle CAV j Optimal lane change trajectory.

[0035] The method for decision-making and planning of a networked autonomous driving vehicle in a freeway diversion area according to the present invention is also characterized in that step 3 includes:

[0036] Step 3.1: Assume that any human-driven vehicle HV hThe centroid coordinates of , and the vehicle potential field model is performed according to formula (6), and the human-driven vehicle HV h Any point in the plane coordinate system The potential field strength generated at ;

[0037]

[0038] In formula (6), The direction of the vehicle potential field and the direction of the human-driven vehicle HV h The angle between the driving directions, R is the complexity coefficient of the operating environment, 0<R≤1, is the potential field parameter value of the manually driven vehicle;

[0039] Step 3.2: Let the center coordinates of the i-th road boundary line be According to formula (7), the potential field of the road boundary lines on both sides of the one-way three-lane road is modeled, and the road boundary line at any point in the plane coordinate system is obtained. The potential field strength generated at ;

[0040]

[0041] In formula (5), is the potential field parameter value of the road boundary line;

[0042] Step 3.3: Let the center coordinate of the jth lane separator be According to formula (8), the potential field of the lane divider of a one-way three-lane vehicle is modeled, and the lane divider at any point in the plane coordinate system is obtained. The magnitude of the potential field generated at ;

[0043]

[0044] Step 3.4: Assume that the center of mass coordinates of the static roadblock in the scene is , and the static roadblock potential field is modeled according to formula (9), and the static roadblock at any point in the plane coordinate system is obtained The potential field strength generated at ;

[0045] .

[0046] Furthermore, the step 5 includes:

[0047] Step 5.1: Construct the jth lane change trajectory curve according to formula (10) The safety evaluation function ;

[0048]

[0049] Step 5.2: Construct the jth lane change trajectory curve according to formula (11) The efficiency evaluation function ;

[0050]

[0051] Step 5.3: Construct the jth lane change trajectory curve according to formula (12) Comfort evaluation function ;

[0052]

[0053] In formula (12), is the number of lane-changing points on the j-th lane-changing trajectory curve, , The polynomial fitting curve equations are First and second order derivatives.

[0054] An electronic device of the present invention includes a memory and a processor, the characteristics of which are that the memory is used to store a program that supports the processor to execute the decision-making and planning method for a networked autonomous driving vehicle in an expressway diversion area, and the processor is configured to execute the program stored in the memory.

[0055] The present invention provides a computer-readable storage medium, and the computer-readable storage medium stores a computer program, which is characterized in that when the computer program is run by a processor, the steps of the method for decision-making and planning of a networked autonomous driving vehicle in an expressway diversion area are executed.

[0056] Compared with the existing technology, the beneficial technical effects of the present invention are embodied in:

[0057] 1. Traditional methods often overlook the uncertainty and complexity of environmental factors when making decisions. This invention utilizes fuzzy logic reasoning technology to process uncertain information, simulating the decision-making process of human drivers and improving system robustness. Fuzzy logic uses membership functions to represent the uncertainty of input variables, enabling the system to make reasonable decisions even with incomplete information.

[0058] 2. In terms of trajectory screening for autonomous vehicles, traditional screening methods often focus on a single performance indicator, such as safety or efficiency, while ignoring other key factors. In contrast, the present invention adopts a comprehensive approach when screening trajectories, taking into account the three dimensions of safety, efficiency, and comfort. By constructing a multi-objective optimization function, the present invention achieves the best balance between these objectives, ensuring optimal driving performance in different situations. In addition, the present invention also introduces the innovative concept of risk field as a key indicator for evaluating the safety of autonomous vehicles. The introduction of the risk field enables the system to more accurately predict and evaluate potential dangerous situations, thereby paying more attention to safety during the trajectory screening process. This method not only improves the safety of autonomous vehicles in various traffic environments, but also enhances the overall driving experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a scene schematic diagram of the present invention;

[0060] Figure 2 It is the overall flow chart of the present invention.

[0061] Figure 3 is a flow chart of the fuzzy logic system of the present invention;

[0062] Figure 4 Schematic diagram of the trajectory planning result of the present invention. DETAILED DESCRIPTION

[0063] In this case, within the ever-advancing field of autonomous driving, this innovative technology breaks through the limitations of traditional technologies and meticulously designs a decision-making and planning solution for connected autonomous vehicles in expressway diversion zones. This solution utilizes advanced decision-making techniques and path planning algorithms to factor in environmental uncertainty, enabling the autonomous driving system to make rational decisions. Furthermore, it comprehensively considers safety, efficiency, and comfort to ensure the stable operation of autonomous vehicles in complex and changing traffic environments. This technology's application goes beyond improving the performance of individual vehicles and focuses on optimizing the entire transportation network. By intelligently regulating traffic flow in diversion zones, it alleviates traffic pressure and improves road capacity. Specifically, the method is applied to a one-way three-lane expressway diversion area road, and the three lanes in the same direction on the expressway diversion area road are respectively recorded as the first lane, the second lane, and the third lane. The first lane is provided with a static roadblock, and an exit ramp is provided on the outside of the third lane. The one-way three-lane road contains at least one connected autonomous driving vehicle and one manually driven vehicle. The length of the upstream section of the exit ramp is L0 as the origin, the vehicle's driving direction is the positive direction of the X-axis, and the direction perpendicular to the X-axis is the Y-axis direction. A plane coordinate system is established. The scene diagram is shown as follows Figure 1 As shown; the flow chart of the lane change trajectory planning method for a connected autonomous driving vehicle is shown in Figure 2 As shown, it is carried out in the following steps;

[0064] Step 1: Use fuzzy logic reasoning to make lane change decisions for connected autonomous vehicles;

[0065] Step 1.1: Construct a fuzzy logic system. The flow chart of the system is as follows: Figure 3 As shown, it includes: input layer, fuzzy layer, rule layer, and output layer;

[0066] Assume that the expected speed of the connected autonomous vehicle is , and the expected vehicle distance is , the input layer includes: speed difference coefficient membership function , membership function of vehicle distance expectation coefficient , as shown in formulas (1) and (2) respectively;

[0067] (1)

[0068] In formula (1), is the speed of connected autonomous vehicles, The speed of manually driven vehicles;

[0069] (2)

[0070] In formula (2), It is the distance between connected autonomous vehicles and manually driven vehicles.

[0071] The fuzzy layer includes: velocity difference coefficient fuzzy subset {small, smaller, medium, larger, large}, fuzzy subset of vehicle distance expectation coefficient {small, smaller, medium, larger, large}, fuzzy subset of lane change intention {weak, weaker, medium, stronger, strong};

[0072] Rule layer, including: lane change fuzzy rules ,The fuzzy rules are shown in Table 1;

[0073] Table 1

[0074]

[0075] Output layer, including: lane change willingness coefficient ;

[0076] Step 1.2: Set a decision plan , , , the specific implementation decisions are as follows:

[0077] :If the lane change willingness coefficient satisfy: , then the decision plan is to keep going straight;

[0078] :If the lane change willingness coefficient satisfy: , then the decision plan is to wait for lane change and go to step 2;

[0079] :If the lane change willingness coefficient satisfy: , then the decision is to change lanes immediately and go to step 3. , , are all constants.

[0080] Step 2: Calculate the CAV of any connected autonomous vehicle according to formula (3): j Compared with manual driving vehicles HV h Remaining collision time , and conduct risk assessments;

[0081] (3)

[0082] The risk assessment method is as follows:

[0083] like , it is determined that there is a collision risk in the scene, and the decision is changed to keep going straight;

[0084] like , it is determined that there is no collision risk in the scene, and the vehicle changes lanes and proceeds to step 3.

[0085] Step 3: Construct a multi-source superposition risk field model, including vehicle potential field, road boundary line potential field, and lane dividing line potential field;

[0086] Step 3.1: Assume any human-driven vehicle HV in the scene h The centroid coordinates of , and the vehicle potential field model is performed according to formula (4), and the human-driven vehicle HV h Any point in the plane coordinate system The potential field strength generated at ;

[0087] (4)

[0088] In formula (4), The direction of the vehicle potential field and the direction of the human-driven vehicle HV hThe angle between the driving directions, R is the complexity coefficient of the operating environment, 0<R≤1, is the potential field parameter value of the manually driven vehicle.

[0089] Step 3.2: Let the center coordinates of the i-th road boundary line be According to formula (5), the potential field of the road boundary lines on both sides of the one-way three-lane road is modeled, and the potential field of any point of the road boundary line in the plane coordinate system is obtained. The potential field strength generated at ;

[0090] (5)

[0091] In formula (5), is the road boundary line potential field parameter value.

[0092] Step 3.3: Let the center coordinate of the jth lane separator be , the potential field of the lane divider of a one-way three-lane vehicle is modeled according to formula (6), and the lane divider at any point in the plane coordinate system is obtained The magnitude of the potential field generated at ;

[0093] (6)

[0094] Step 3.4: Assume the center of mass coordinates of the static roadblock in the scene is , and the static roadblock potential field is modeled according to formula (7), and the static roadblock at any point in the plane coordinate system is obtained The potential field strength generated at ;

[0095] (7)

[0096] Step 3.5: Obtain any point in the plane coordinate system by superimposing the vehicle potential field, road boundary line potential field, and lane dividing line potential field Multi-source risk field superposition potential field , as shown in formula (8);

[0097] (8)

[0098] In formula (8), , , , are the weight parameters of vehicle potential field, road boundary line potential field, lane dividing line potential field, and static roadblock potential field respectively.

[0099] Step 4: Assume any connected autonomous vehicle CAV j The starting point of lane change is Aj , the lane change destination is F j , and the fifth-order polynomial curve fitting method is used to find the lane-changing starting point A j and lane change end point F j Generate a connected autonomous vehicle (CAV) j The lane-changing trajectory curve cluster is a cluster of m lane-changing trajectory curves, and the cluster contains m lane-changing trajectory curves. , , , let the jth lane change trajectory curve The polynomial fitting curve equation is , 1≤j≤m.

[0100] Step 5: Construct a multi-objective comprehensive evaluation function for the lane change trajectory curve;

[0101] Step 5.1: Construct the jth lane change trajectory curve according to formula (9): The safety evaluation function ;

[0102] (9)

[0103] Step 5.2: Construct the j-th lane change trajectory curve according to formula (10) The efficiency evaluation function ;

[0104] (10)

[0105] Step 5.3: Construct the j-th lane change trajectory curve according to formula (11) Comfort evaluation function ;

[0106] (11)

[0107] In formula (11), is the number of lane-changing points on the j-th lane-changing trajectory curve, , The polynomial fitting curve equations are First and second order derivatives.

[0108] Step 5.4: Construct the j-th lane change trajectory curve using formula (12) Multi-objective comprehensive evaluation function ;

[0109] (12)

[0110] In formula (12), , , The safety evaluation function , efficiency evaluation function , comfort evaluation function The weight parameter of .

[0111] Step 6: Use formula (12) to solve the multi-objective comprehensive evaluation value of m trajectory curves, and select the lane change trajectory curve corresponding to the minimum comprehensive evaluation value That is, connected autonomous vehicle CAV j The optimal lane change trajectory and the trajectory planning result after screening are shown in the following figure: Figure 4 shown.

[0112] In this embodiment, an electronic device includes a memory and a processor. The memory is used to store a program that supports the processor to execute the above-mentioned expressway diversion area networked autonomous driving vehicle decision-making and planning method, and the processor is configured to execute the program stored in the memory.

[0113] In this embodiment, a computer-readable storage medium stores a computer program on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the above-mentioned expressway diversion area networked autonomous driving vehicle decision-making and planning method are executed.

Claims

1. A decision-making and planning method for connected autonomous driving vehicles in a freeway diversion area, characterized in that: The method is applied to a one-way three-lane expressway diversion zone road, wherein the three lanes in the same direction on the expressway diversion zone road are respectively designated as the first lane, the second lane, and the third lane. The first lane is provided with a static roadblock, and an exit ramp is provided outside the third lane. The one-way three-lane road contains at least one connected autonomous vehicle and one manually driven vehicle. A plane coordinate system is established with the length L0 of the upstream section of the exit ramp as the origin, the vehicle's travel direction as the positive direction of the X-axis, and the direction perpendicular to the X-axis as the Y-axis. The lane change trajectory planning method for the connected autonomous vehicle is performed in the following steps: Step 1: Use fuzzy logic reasoning to make lane change decisions for connected autonomous vehicles; Step 1.1: Construct a fuzzy logic system, including input layer, fuzzy layer, rule layer, and output layer. Assume that the expected speed of the connected autonomous vehicle is , and the expected vehicle distance is , the input layer includes: speed difference coefficient membership function , membership function of vehicle distance expectation coefficient , as shown in formulas (1) and (2) respectively; (1) In formula (1), is the speed of connected autonomous vehicles, The speed of manually driven vehicles; (2) In formula (2), The distance between the connected autonomous vehicle and the manually driven vehicle; The fuzzy layer includes: velocity difference coefficient fuzzy subset {small, smaller, medium, larger, large}, fuzzy subset of vehicle distance expectation coefficient {small, smaller, medium, larger, large}, fuzzy subset of lane change intention {weak, weaker, medium, stronger, strong}; The rule layer includes: lane change fuzzy rules ; The output layer includes: lane change willingness coefficient ; Step 1.2: Set a decision plan , , , the specific implementation decisions are as follows: :If the lane change willingness coefficient satisfy: , then the decision plan is to keep going straight; :If the lane change willingness coefficient satisfy: , then the decision plan is to wait for lane change and go to step 2; :If the lane change willingness coefficient satisfy: , then the decision is to change lanes immediately and go to step 3. , , are all constants; Step 2: Calculate the CAV of any connected autonomous vehicle according to formula (3): j Compared with manual driving vehicles HV h Remaining collision time , and conduct risk assessments; (3) The risk assessment method is as follows: like , it is determined that there is a collision risk in the scene, and the decision is changed to keep going straight; like , it is determined that there is no collision risk in the scene, and the vehicle changes lanes and proceeds to step 3; Step 3: Use formula (4) to construct the human-driven vehicle HV h , road boundary line, lane dividing line, static roadblock at any point in the plane coordinate system Multi-source superposition risk field model ; (4) In formula (4), , , , Any point Vehicle potential field at , road boundary line potential field , lane dividing line potential field , static roadblock potential field The weight parameter of Step 4: Assume any connected autonomous vehicle CAV j The starting point of lane change is A j , the lane change destination is F j , and the fifth-order polynomial curve fitting method is used to find the lane-changing starting point A j and lane change end point F j Generate a connected autonomous vehicle (CAV) j The lane-changing trajectory curve cluster is a cluster of m lane-changing trajectory curves, and the cluster contains m lane-changing trajectory curves. , , , let the jth lane change trajectory curve The polynomial fitting curve equation is , 1≤j≤m; Step 5: Construct the j-th lane change trajectory curve using formula (5) Multi-objective comprehensive evaluation function ; (5) In formula (5), , , The safety evaluation function , efficiency evaluation function , comfort evaluation function The weight parameter of Step 6: Use formula (5) to solve the multi-objective comprehensive evaluation value of m trajectory curves, and select the lane change trajectory curve corresponding to the minimum comprehensive target evaluation value That is, connected autonomous vehicle CAV j Optimal lane change trajectory.

2. The method for decision-making and planning of a networked autonomous driving vehicle in a freeway diversion area according to claim 1, characterized in that: The step 3 comprises: Step 3.1: Assume that any human-driven vehicle HV h The centroid coordinates of , and the vehicle potential field model is performed according to formula (6), and the human-driven vehicle HV h Any point in the plane coordinate system The potential field strength generated at ; (6) In formula (6), The direction of the vehicle potential field and the direction of the human-driven vehicle HV h The angle between the driving directions, R is the complexity coefficient of the operating environment, 0<R≤1, is the potential field parameter value of the manually driven vehicle; Step 3.2: Let the center coordinates of the i-th road boundary line be According to formula (7), the potential field of the road boundary lines on both sides of the one-way three-lane road is modeled, and the road boundary line at any point in the plane coordinate system is obtained. The potential field strength generated at ; (7) In formula (5), is the potential field parameter value of the road boundary line; Step 3.3: Let the center coordinate of the jth lane separator be According to formula (8), the potential field of the lane divider of a one-way three-lane vehicle is modeled, and the lane divider at any point in the plane coordinate system is obtained. The magnitude of the potential field generated at ; (8) Step 3.4: Assume that the center of mass coordinates of the static roadblock in the scene is , and the static roadblock potential field is modeled according to formula (9), and the static roadblock at any point in the plane coordinate system is obtained The potential field strength generated at ; (9)。 3. The method for decision-making and planning of a networked autonomous driving vehicle in a freeway diversion area according to claim 2, characterized in that: The step 5 comprises: Step 5.1: Construct the jth lane change trajectory curve according to formula (10) The safety evaluation function ; (10) Step 5.2: Construct the jth lane change trajectory curve according to formula (11) The efficiency evaluation function ; (11) Step 5.3: Construct the jth lane change trajectory curve according to formula (12) Comfort evaluation function ; (12) In formula (12), is the number of lane-changing points on the j-th lane-changing trajectory curve, , The polynomial fitting curve equations are First and second order derivatives.

4. An electronic device comprising a memory and a processor, characterized in that: The memory is used to store a program that supports the processor to execute the decision-making and planning method for a networked autonomous driving vehicle in an expressway diversion area as described in any one of claims 1-3, and the processor is configured to execute the program stored in the memory.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for decision-making and planning of a connected autonomous driving vehicle in an expressway diversion area as recited in any one of claims 1 to 3 are executed.

Images