A method, system, device and medium for controlling a left turn of a connected autonomous vehicle

Abstract

The application provides a network-connected automatic driving vehicle left turn control method, system, device and medium, comprising: acquiring road traffic information, calculating the interaction intensity coefficient between the ego vehicle and the oncoming vehicle and the predicted collision probability between the ego vehicle and all traffic participants according to the road traffic information; determining the traffic environment index according to the road traffic information, and then calculating the risk exposure index and the communication confidence of the ego vehicle according to the traffic environment index; determining the decision state of the ego vehicle according to the risk exposure index, the interaction intensity coefficient, the predicted collision probability and the communication confidence, and controlling the ego vehicle to turn left according to the decision state. The application solves the problem that the automatic driving decision system in the prior art is difficult to quantify road risks and efficiently cooperate with other vehicles when controlling the network-connected automatic driving vehicle to make unprotected left turns.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, and in particular to a method, system, device and medium for controlling left turns of connected autonomous vehicles. Background Technology

[0002] Unprotected left turns at unsignalized intersections or under damaged traffic lights have long been a technical challenge of "high complexity and high risk" in the field of autonomous driving. Therefore, existing autonomous driving decision-making systems face many problems when controlling connected autonomous vehicles to make left turns. Due to the lack of clear priority rules and the need to handle complex interactions between multiple traffic participants from multiple directions, autonomous driving decision-making systems not only lack global traffic situation awareness in unprotected left turn scenarios, but also struggle to quantify road risks, limiting the accuracy and safety of decision-making. In addition, autonomous driving decision-making systems often use rule-based decision-making logic based on simple threshold judgments (such as distance, speed, and TTC), which cannot efficiently coordinate with other vehicles and is difficult to flexibly adapt to the highly dynamic and uncertain multi-agent interaction behavior in unprotected left turn scenarios. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a left-turn control method for connected autonomous vehicles, which solves the problem that existing autonomous driving decision-making systems struggle to quantify road risks and efficiently coordinate with other vehicles when controlling connected autonomous vehicles to make unprotected left turns.

[0004] According to an embodiment of the present invention, a left-turn control method for a connected autonomous vehicle includes: Obtain road traffic information and calculate the interaction intensity coefficient between the vehicle and oncoming vehicles, as well as the predicted collision probability between the vehicle and all traffic participants based on the road traffic information. Based on road traffic information, traffic environment indicators are determined, and then the risk exposure index and communication confidence level of the vehicle are calculated based on the traffic environment indicators. The decision state of the vehicle is determined based on the risk exposure index, interaction intensity coefficient, predicted collision probability, and communication confidence level, and the vehicle is controlled to turn left based on the decision state.

[0005] Preferably, the method for calculating the interaction intensity coefficient between a vehicle and oncoming vehicles based on road traffic information includes: Vehicle driving information of the vehicle and oncoming vehicles is extracted from road traffic information. Then, based on the vehicle driving information, multiple predicted trajectories of the vehicle and oncoming vehicles are generated using a trajectory prediction model. Set the sampling interval, and then sample each predicted trajectory of the vehicle based on the sampling interval to obtain multiple corresponding trajectory points and corresponding coordinates. Then, combine all the trajectory points of the vehicle into a corresponding set of trajectory points. Based on the sampling interval, each oncoming vehicle predicted trajectory is sampled to obtain multiple oncoming vehicle trajectory points and their corresponding coordinates. Then, all oncoming vehicle trajectory points are combined into a corresponding oncoming vehicle trajectory point set. The trajectory conflict points are determined based on the set of self-vehicle trajectory points and the set of vehicle trajectory points. The risk collision distance between the self-vehicle and the trajectory conflict point is calculated. Then, the interaction intensity coefficient is calculated based on the vehicle driving information and the risk collision distance.

[0006] Preferably, the method for determining trajectory conflict points based on the set of self-vehicle trajectory points and the set of vehicle trajectory points includes: Combine all sets of trajectory points of autonomous vehicles with all sets of trajectory points of oncoming vehicles to obtain multiple pairs of trajectory points; For each pair of trajectory points, calculate the first distance between the trajectory point of the vehicle and the trajectory point of the oncoming vehicle, and take all trajectory point pairs whose first distance is less than a preset value as candidate trajectory point pairs. Then, form a candidate point set from all candidate trajectory point pairs. The most frequent candidate trajectory point pair in the entire candidate point set is identified, and the midpoint between the two trajectory points in the candidate trajectory point pair is then taken as the trajectory conflict point.

[0007] Preferably, if the distance between all the trajectory points of the self-vehicle and all the trajectory points of the oncoming vehicles is greater than a preset value, then the distance between the self-vehicle and the oncoming vehicles is taken as the risk collision distance.

[0008] Preferably, the method for calculating the predicted collision probability between the vehicle and all traffic participants based on road traffic information includes: Traffic participation data of vehicles and all traffic participants are extracted from road traffic information. Then, based on the traffic participation data, multiple predicted trajectories of vehicles and multiple predicted trajectories of each traffic participant are generated using a trajectory prediction model. Combine all the predicted trajectories of the vehicles with the predicted trajectories of all participants into multiple trajectory pairs, and then calculate the second distance between the two trajectories in each trajectory pair; The ratio of the number of trajectory pairs whose second distance is less than a preset value to the total number of trajectory pairs is used as the predicted collision probability.

[0009] Preferably, the method for calculating the risk exposure index of a vehicle based on road traffic information includes: Based on road traffic information, determine the type and location of all traffic participants, and assign a type weight to each type of traffic participant; The road type is determined based on road traffic information, and a road risk coefficient is assigned to the current road. The relative distance between each traffic participant and the vehicle is calculated based on the location of each traffic participant. Then, the risk exposure index of the vehicle is calculated based on the relative distance, type weight, and road risk coefficient.

[0010] Preferably, the decision states include a strong interaction waiting state, a medium interaction cautious left turn state, a weak interaction cooperative left turn state, and a no-interaction free left turn state. When the vehicle is in a strong interactive waiting state, the first target driving speed is calculated based on the distance between the vehicle and the trajectory conflict point, and then the vehicle drives at the first target driving speed. When the vehicle is in a cautious left turn state, the potential conflict participants are identified based on the minimum second distance, the third distance between the vehicle and the potential conflict participants is calculated, and the vehicle's speed limit is determined based on the third distance. Then the vehicle travels at a speed that does not exceed the speed limit. When the vehicle is in a weak interactive cooperative left turn state, the second target driving speed is calculated based on the speed of the oncoming vehicle and the distance between the vehicle and the oncoming vehicle, and then the vehicle drives at the second target driving speed. When the vehicle is in a free left turn state without interaction, the vehicle travels at its current speed.

[0011] On the other hand, according to embodiments of the present invention, a left-turn control system for a connected autonomous vehicle is also provided. This system uses the aforementioned left-turn control method for a connected autonomous vehicle, comprising: A road information acquisition module, wherein the road information acquisition module is used to acquire road traffic information; The calculation module is used to calculate the interaction intensity coefficient between the vehicle and oncoming vehicles and the predicted collision probability between the vehicle and all traffic participants based on road traffic information, as well as to determine traffic environment indicators, and then calculate the risk exposure index and communication confidence of the vehicle based on the traffic environment indicators. The judgment module is used to determine the decision state of the vehicle based on the risk exposure index, interaction intensity coefficient, predicted collision probability and communication confidence. The control module is used to control the vehicle to turn left based on the decision state.

[0012] On the other hand, according to an embodiment of the present invention, a computer is also provided, including at least one processor and a memory, the memory storing a computer program configured to be executed by the processor to implement the above-described method for controlling left turns of a connected autonomous vehicle.

[0013] On the other hand, according to embodiments of the present invention, a storage medium is also provided, which is a computer-readable storage medium storing a computer program that can be executed by one or more processors to implement the above-described method for controlling left turns of a connected autonomous vehicle.

[0014] Compared with the prior art, the present invention has the following beneficial effects: This invention quantifies road traffic information into interaction intensity coefficients and predicted collision probabilities between the vehicle and other traffic participants. Simultaneously, it determines traffic environment indicators based on this information and then quantifies these indicators into risk exposure indices and communication confidence levels. This comprehensive approach quantifies road risks, providing specific data for the autonomous driving decision-making system and improving decision accuracy. Furthermore, this quantified data includes interactions between the vehicle and other traffic participants, as well as interactions between the vehicle and its environment. By integrating all traffic information for comprehensive judgment, the system enables efficient collaboration between the vehicle and other traffic participants, further enhancing the decision-making accuracy of the autonomous driving system. Attached Figure Description

[0015] Figure 1 This is a diagram illustrating a method for controlling left turns of a connected autonomous vehicle according to an embodiment of the present invention.

[0016] Figure 2 This is a schematic diagram of trajectory conflict points in an embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0018] like Figure 1 As shown in the figure, an embodiment of the present invention proposes a left-turn control method for connected autonomous vehicles, including: Obtain road traffic information and calculate the interaction intensity coefficient between the vehicle and oncoming vehicles, as well as the predicted collision probability between the vehicle and all traffic participants based on the road traffic information. This invention fully utilizes the global perception capability of the Roadside Cooperative Control Unit (RCCU) to provide autonomous vehicles with real-time traffic situation information that is beyond line of sight, has no blind spots, and is highly accurate and has low latency through V2X communication. This compensates for the blind spots of onboard perception and accurately identifies and predicts the motion state and intentions of all traffic participants (especially manually driven vehicles, HDVs). After collecting road traffic information from all traffic participants (including autonomous vehicles, other vehicles, and pedestrians) on the road, the Roadside Cooperative Control Unit sends it to the decision-making system of each connected autonomous vehicle. The decision-making system uses this road traffic information to make decisions and control the vehicle to turn left.

[0019] In this invention, the Roadside Cooperative Control Unit (RCCU) plays the role of a global dispatch center and enabler, providing autonomous vehicles with key real-time traffic information and dispatch instructions through V2X communication, thereby optimizing the overall traffic efficiency and safety of intersections.

[0020] In this invention, the Interaction Intensity Index (ISI) is used to quantify the intensity of instantaneous interaction between the self-vehicle and a potential conflicting vehicle. It is primarily based on relative kinematic states and considers factors that may exacerbate the interaction. Its core influencing factors include: (1) The time urgency brought about by speed difference: the greater the speed difference between the vehicle and the oncoming vehicle, the faster they enter the conflict zone and the higher the interaction intensity.

[0021] (2) Difference in motion trends due to acceleration differences: The greater the difference in acceleration between the vehicle and the oncoming vehicle, the greater the difference in their future motion trends, which is more likely to lead to uncertainty and potential conflict.

[0022] (3) Uncertainty of intention reflected by yaw angle: The change in the vehicle's heading angle (or angular velocity) reflects its possible turning intention. In a left turn scenario, the change in the heading angle of an oncoming vehicle may indicate that it will continue to go straight or make a turn, increasing the complexity of the vehicle's decision-making.

[0023] like Figure 2 As shown, based on the aforementioned influencing factors, it is first necessary to determine the potential conflict points between the vehicle and oncoming vehicles (here, oncoming vehicles are vehicles in other lanes not in the same lane as the vehicle). Therefore, this invention is based on the current state (location) of the vehicle and the target vehicle in the acquired road traffic information. and heading angle ,speed Add Gaussian noise perturbation, where the heading angle perturbation follows a normal distribution. ,in (Initial values, based on driving behavior uncertainty) Using existing trajectory prediction models (such as GAT and Transform deep learning models) to predict the trajectories of the self-vehicle and the oncoming vehicle respectively, multiple predicted trajectories of the self-vehicle and multiple predicted trajectories of the oncoming vehicle are generated.

[0024] Set time step Prediction window (Based on human reaction time and vehicle dynamic calibration), total number of steps For each time step : The vehicle's trajectory points can be extracted from the predicted trajectory of the vehicle: The trajectory points of oncoming vehicles can be extracted from the predicted trajectories of oncoming vehicles: in For the first The perturbation heading angle of the step.

[0025] All the self-vehicle trajectory points are combined into a corresponding self-vehicle trajectory point set, and all the oncoming vehicle trajectory points are combined into a corresponding self-vehicle trajectory point set. Then, all the self-vehicle trajectory point sets and all the oncoming vehicle trajectory point sets are combined to obtain multiple sets of trajectory point pairs. All trajectory point pairs are traversed, and the first distance between the self-vehicle trajectory point and the oncoming vehicle trajectory point in each set of trajectory point pairs is calculated.

[0026] If distance (Initial safety threshold), record the trajectory point pair as a candidate trajectory point pair, and then form a candidate point set from all candidate trajectory point pairs.

[0027] The pair of candidate trajectory points that appears most frequently in the entire candidate point set is identified, and the midpoint between the two trajectory points in the candidate trajectory is then taken as the trajectory conflict point. Then, the risk collision distance is calculated based on the coordinates of the vehicle's front center and the coordinates of the trajectory conflict point. If the distance between all the trajectory points of the self-vehicle and all the trajectory points of the oncoming vehicles is greater than the preset value, then the distance between the self-vehicle and the oncoming vehicles will be used as the risk collision distance.

[0028] Finally, the risk collision distance between the vehicle and the point of trajectory conflict is calculated using the following formula: in, In time The current longitudinal speed of the vehicle and the oncoming vehicle. In time The current longitudinal acceleration of the vehicle and the oncoming vehicle. To represent the risky collision distance from the center of the vehicle's front end to the point of trajectory conflict at time t, In time The angular velocity of the oncoming vehicle reflects its lateral movement trend or intention change.

[0029] Predicting collision probability Used to assess the future of autonomous vehicles The probability of colliding with any other traffic participant within a given time is a quantitative indicator based on the predicted trajectory. Therefore, this invention extracts traffic participation data of the vehicle and all traffic participants from road traffic information. Then, based on the traffic participation data, a trajectory prediction model is used to generate multiple predicted trajectories for the vehicle and multiple predicted trajectories for each traffic participant. After that, all the predicted trajectories of the vehicle and the predicted trajectories of all participants are combined into multiple trajectory pairs. Then, the second distance between the two trajectories in each trajectory pair is calculated, which is the minimum distance between the two trajectories.

[0030] If the second distance is less than a preset value (usually set to 2m), it is determined that a collision (trajectory overlap) will occur if the vehicle and other traffic participants move along the predicted trajectory of the vehicle and the expected trajectory of the other participants, respectively, of the trajectory pair. Therefore, the ratio of the number of trajectory pairs with the second distance less than the preset value to the total number of trajectory pairs is used as the predicted collision probability of the vehicle. The calculation formula is as follows: In the formula: For the future Within the time window, the vehicle has the highest probability of being safe (i.e., at any given moment, the distance to any other road user is greater than the preset safe distance), and the range of values ​​is [not specified]. ; The default value is used to determine the length of the prediction time window. ; In time The second distance between the vehicle and all other road users; This is a preset safe distance threshold. When the second distance is less than this value, a potential collision risk is considered to exist. (Default value) ; This indicates that the second distance is less than a preset value ( The probability estimate is formed by the ratio of the number of corresponding trajectory pairs to the total number of trajectory pairs.

[0031] This indicator quantifies the key information of "whether it is safe and passable in the future" in probabilistic form by predicting the safety of all possible trajectories over a period of time.

[0032] Based on road traffic information, traffic environment indicators are determined, and then the risk exposure index and communication confidence level of the vehicle are calculated based on the traffic environment indicators. Both the Risk Exposure Index (ERI) and the Communication Confidence Index (CVI) assess the impact of the road and communication environment on the vehicle itself. The ERI quantifies the overall potential risk level faced by the vehicle in the current scenario. It comprehensively considers the risk coefficients of traffic environment indicators around the vehicle, including the type weights of traffic participants and the risk coefficients of the road environment. Different types of traffic participants (such as trucks, cars, and pedestrians) pose different potential dangers to the vehicle. For example, trucks, due to their large size and mass, have more severe collision consequences; pedestrians, due to their higher uncertainty in movement, also pose a greater risk. At the same time, road environmental conditions (such as dry or slippery roads) directly affect the vehicle's grip and braking performance, thereby changing the overall risk level.

[0033] The formula for calculating the Exposure Risk Index (ERI) is as follows: in, In time The risk exposure index faced by a vehicle is typically normalized to a range of values. ; To include all nearby road users who may potentially conflict with the vehicle. Summation, For traffic participants The current speed; For bicycles and traffic participants The current relative distance, This is the road environmental risk weight, used to adjust the degree of impact of environmental factors on the overall risk (default value: 1.0).

[0034] For traffic participants The type weight reflects the relative risk posed by that type of traffic participant to the vehicle. In this invention, the truck is set as... Car is set as Pedestrian set as Non-motorized vehicles (Cyclist) are set as These weights are empirically set based on the impact of vehicle type on collision severity and behavioral uncertainty, with pedestrians having the highest risk weight due to their high degree of uncertainty.

[0035] The road environmental risk coefficient reflects the degree of risk of current road conditions. In this invention, the dry road surface is... Slippery road surface is set as Icy road surface is set as These coefficients are empirically set based on changes in road surface adhesion coefficient and driving difficulty.

[0036] ERI comprehensively assesses the risk environment currently in which a vehicle is located by superimposing the speed / distance ratio (reflecting urgency) of all traffic participants with potential conflicts, combining it with their type weights, and then correcting it with an environmental risk item.

[0037] Communication confidence Used to evaluate the reliability and content validity of V2X communication, it directly affects the system's adoption of V2X information and the depth of vehicle-to-infrastructure (V2I) cooperation. The communication confidence model comprehensively considers physical layer indicators such as V2X signal strength, signal-to-noise ratio, packet loss rate, and communication latency, as well as the type of received information, the credibility of the sending source (whether it comes from the certified RSU), and the freshness of the information. In this invention, if the vehicle can receive a clear yield signal or cooperative passage instruction (highest confidence level), then If the vehicle receives scheduling suggestions or predictive information from the roadside cooperative control unit (with medium confidence level, requiring verification in conjunction with its own perception), then =0.7, if the vehicle's communication is interrupted or the signal is weak (low reliability, for reference only), then =0.3, if the vehicle receives a conflict warning or abnormal information (extremely low confidence, requiring immediate evasive action, or indicating communication abnormality), the measurement... =0.

[0038] By quantifying the confidence level of communication, the system can intelligently weigh the priorities of onboard perception and external V2X information, avoiding erroneous decisions caused by unstable communication or incorrect information.

[0039] This invention quantifies road traffic information into interaction intensity coefficients between the vehicle and other traffic participants and predicted collision probabilities. Simultaneously, road traffic information is used to determine traffic environment indicators, which are then quantified into risk exposure indices and communication confidence levels. This comprehensive approach quantifies road risks, providing specific judgment data for autonomous driving decision-making systems and improving decision-making accuracy.

[0040] The decision state of the vehicle is determined based on the risk exposure index, interaction intensity coefficient, predicted collision probability, and communication confidence level, and the vehicle is controlled to turn left based on the decision state.

[0041] After obtaining the risk exposure index, interaction intensity coefficient, predicted collision probability, and communication confidence level, the autonomous driving decision-making system on the vehicle uses these four indicators to determine the vehicle's driving decisions. This aims to fully cover the entire spectrum of left-turn scenarios, from high-risk to risk-free, enabling autonomous vehicles to adaptively select the safest control strategy based on the real-time environment and risk assessment results.

[0042] (1) Strong interactive waiting state If ISI or ERI or or The vehicle is identified as being in a high-interaction waiting state. This state is applicable to high-risk, high-interaction scenarios. The vehicle needs to slow down to a low speed and drive until it comes to a complete stop and wait for a safe opportunity.

[0043] When the interaction intensity is extremely high, or the overall environmental risk is extremely high, or the probability of future conflict exceeds 50%, a waiting state must be entered. In particular, when the confidence level of roadside communication is low (e.g., the received instructions are unreliable or missing), even if other risk indicators temporarily allow passage, vehicles should enter a waiting state to ensure safety.

[0044] At this point, the vehicle will employ a deceleration and stopping strategy until it comes to a complete stop or waits for a safe opportunity.

[0045] In the formula: The first target speed for the vehicle in the next moment. The vehicle's current speed. The distance between the current vehicle and the point of conflict with the trajectory. The distance required for a vehicle to park safely, or a safe distance, default value. , The velocity decay distance determines the gradient at which the velocity decreases; the default value is... .

[0046] This formula indicates that when the distance from the point of conflict with the trajectory... Much larger At that time, the speed was close to ;when near At that point, the speed will drop sharply and eventually come to a stop.

[0047] (2) Cautious left turn state in the middle interaction like and and and At this time, the vehicle is judged to be in a medium-interaction cautious left turn state. This state is applicable to medium-risk, medium-interaction scenarios. The vehicle turns left cautiously at a limited speed and is ready to stop at any time.

[0048] When the interaction intensity and environmental risk are at a moderate level, and the probability of future conflict is within an acceptable range, a left turn can be made cautiously. In this case, if the confidence level of roadside communication is moderate, the vehicle will refer to roadside information but mainly rely on its own perception to make a cautious decision.

[0049] At this point, the vehicle identifies potential conflict participants based on the minimum second distance and then makes a left turn at a speed not exceeding the speed limit, which takes into account the TTC time between the vehicle and the traffic participants as well as the safe braking distance.

[0050] In the formula, This is the maximum speed limit for the vehicle. This is the maximum permissible speed for a vehicle turning left; the default value is... , The current distance between the vehicle and other road users. The desired minimum TTC time is the time it would take for the vehicle and other road users to collide if the vehicle continues in its current state. (Default value) .

[0051] This method ensures that the vehicle maintains sufficient safe braking distance and reaction time when turning left, avoiding excessive acceleration.

[0052] (3) Weak interaction cooperative left turn state If ISI And ERI and And received confirmation from roadside coordination ( The vehicle is judged to be in a weak interactive collaborative left turn state. This state is suitable for low-risk, weak-interaction but collaborative scenarios. The vehicle can perform collaborative left turns with V2X collaborative commands or HDV predicted intentions.

[0053] A coordinated left turn can be made when the risk is low and clear roadside guidance or coordinated passage instructions have been received via V2X. Please note that... This represents the confidence level of communication. A value of 1.0 indicates that communication is reliable and explicit coordination instructions have been received, but the final determination of secure passage still depends on... Risk indicators, etc.

[0054] At this point, the vehicle calculates the second target speed based on the speed of the oncoming vehicle and the distance between the vehicle and the oncoming vehicle, and then plans the coordinated turning speed according to the coordinated instructions.

[0055] In the formula, The second target speed for the vehicle during a coordinated left turn. The speed of oncoming vehicles, The default value represents the desired deceleration rate for oncoming vehicles when deceleration is required. This represents the current distance between your vehicle and oncoming vehicles.

[0056] This method calculates the speed required for the vehicle to safely complete a turn when the oncoming vehicle slows down to yield, based on the motion state of the oncoming vehicle. It ensures that the vehicle can pass at the most reasonable speed when the oncoming vehicle cooperates in yielding.

[0057] (4) Free left turn state without interaction If ISI And ERI and The vehicle is judged to be in a free left-turn state without interaction. This state is suitable for scenarios with almost no risk and no interaction. The vehicle can turn left freely according to the preset ideal trajectory.

[0058] When the interaction intensity is extremely low, the environmental risk is extremely low, and the probability of future conflict is extremely small (almost safe), the vehicle can turn left freely. In this state, the communication confidence level can be disregarded. At this time, the vehicle can turn left according to the predetermined route, speed, and heading angle.

[0059] In this state, the risk is extremely low, and vehicles can prioritize efficiency and comfort, passing through at the optimal path and speed.

[0060] The four decision states mentioned above use quantitative data that includes the interaction between the vehicle and other traffic participants, as well as the interaction between the vehicle and the environment. By integrating all traffic information, a comprehensive judgment can be made to enable efficient collaboration between the vehicle and other traffic participants, thereby further improving the decision-making accuracy of the autonomous driving decision-making system.

[0061] On the other hand, embodiments of the present invention also provide a left-turn control system for a connected autonomous vehicle, which uses the aforementioned left-turn control method for a connected autonomous vehicle, including: A road information acquisition module, wherein the road information acquisition module is used to acquire road traffic information; The calculation module is used to calculate the interaction intensity coefficient between the vehicle and oncoming vehicles and the predicted collision probability between the vehicle and all traffic participants based on road traffic information, as well as to determine traffic environment indicators, and then calculate the risk exposure index and communication confidence of the vehicle based on the traffic environment indicators. The judgment module is used to determine the decision state of the vehicle based on the risk exposure index, interaction intensity coefficient, predicted collision probability and communication confidence. The control module is used to control the vehicle to turn left based on the decision state.

[0062] On the other hand, embodiments of the present invention also provide a computer, including at least one processor and a memory, wherein the memory stores a computer program configured to be executed by the processor to implement the above-described method for controlling left turns of a connected autonomous vehicle.

[0063] On the other hand, embodiments of the present invention also provide a storage medium, which is a computer-readable storage medium, on which a computer program is stored. The computer program can be executed by one or more processors to implement the above-described method for controlling left turns of a connected autonomous vehicle.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for controlling left turns in a connected autonomous vehicle, characterized in that: include: Obtain road traffic information and calculate the interaction intensity coefficient between the vehicle and oncoming vehicles based on the road traffic information, including: Vehicle driving information of the vehicle and oncoming vehicles is extracted from road traffic information. Then, based on the vehicle driving information, multiple predicted trajectories of the vehicle and oncoming vehicles are generated using a trajectory prediction model. Set the sampling interval, and then sample each predicted trajectory of the vehicle based on the sampling interval to obtain multiple corresponding trajectory points and corresponding coordinates. Then, combine all the trajectory points of the vehicle into a corresponding set of trajectory points. Based on the sampling interval, each oncoming vehicle predicted trajectory is sampled to obtain multiple oncoming vehicle trajectory points and their corresponding coordinates. Then, all oncoming vehicle trajectory points are combined into a corresponding oncoming vehicle trajectory point set. The trajectory conflict points are determined based on the set of self-vehicle trajectory points and the set of vehicle trajectory points. The risk collision distance between the self-vehicle and the trajectory conflict point is calculated. Then, the interaction intensity coefficient is calculated based on the vehicle driving information and the risk collision distance. Calculate the predicted collision probability between the vehicle and all other road users based on road traffic information, including: Traffic participation data of vehicles and all traffic participants are extracted from road traffic information. Then, based on the traffic participation data, multiple predicted trajectories of vehicles and multiple predicted trajectories of each traffic participant are generated using a trajectory prediction model. Combine all the predicted trajectories of the vehicles with the predicted trajectories of all participants into multiple trajectory pairs, and then calculate the second distance between the two trajectories in each trajectory pair; The ratio of the number of trajectory pairs whose second distance is less than the preset value to the total number of trajectory pairs is used as the predicted collision probability. Based on road traffic information, traffic environment indicators are determined, and then the risk exposure index and communication confidence level of the vehicle are calculated based on the traffic environment indicators. The decision state of the vehicle is determined based on the risk exposure index, interaction intensity coefficient, predicted collision probability, and communication confidence level, and the vehicle is controlled to turn left based on the decision state.

2. The left-turn control method for a connected autonomous vehicle as described in claim 1, characterized in that: Methods for determining trajectory conflict points based on the set of self-vehicle trajectory points and the set of vehicle trajectory points include: Combine all sets of trajectory points of autonomous vehicles with all sets of trajectory points of oncoming vehicles to obtain multiple pairs of trajectory points; For each pair of trajectory points, calculate the first distance between the trajectory point of the vehicle and the trajectory point of the oncoming vehicle, and take all trajectory point pairs whose first distance is less than a preset value as candidate trajectory point pairs. Then, form a candidate point set from all candidate trajectory point pairs. The most frequent candidate trajectory point pair in the entire candidate point set is identified, and the midpoint between the two trajectory points in the candidate trajectory point pair is then taken as the trajectory conflict point.

3. The left-turn control method for a connected autonomous vehicle as described in claim 2, characterized in that: If the distance between all the trajectory points of the autonomous vehicle and all the trajectory points of the oncoming vehicle is greater than the preset value, then the distance between the autonomous vehicle and the oncoming vehicle will be used as the risk collision distance.

4. The left-turn control method for a connected autonomous vehicle as described in claim 1, characterized in that: Methods for calculating the risk exposure index of a vehicle based on road traffic information include: Based on road traffic information, determine the type and location of all traffic participants, and assign a type weight to each type of traffic participant; The road type is determined based on road traffic information, and a road risk coefficient is assigned to the current road. The relative distance between each traffic participant and the vehicle is calculated based on the location of each traffic participant. Then, the risk exposure index of the vehicle is calculated based on the relative distance, type weight, and road risk coefficient.

5. The left-turn control method for a connected autonomous vehicle as described in claim 1, characterized in that: The decision states include a strong interaction waiting state, a medium interaction cautious left turn state, a weak interaction cooperative left turn state, and a no-interaction free left turn state. When the vehicle is in a strong interactive waiting state, the first target driving speed is calculated based on the distance between the vehicle and the trajectory conflict point, and then the vehicle drives at the first target driving speed. When the vehicle is in a cautious left turn state, the potential conflict participants are identified based on the minimum second distance, the third distance between the vehicle and the potential conflict participants is calculated, and the vehicle's speed limit is determined based on the third distance. Then the vehicle travels at a speed that does not exceed the speed limit. When the vehicle is in a weak interactive cooperative left turn state, the second target driving speed is calculated based on the speed of the oncoming vehicle and the distance between the vehicle and the oncoming vehicle, and then the vehicle drives at the second target driving speed. When the vehicle is in a free left turn state without interaction, the vehicle travels at its current speed.

6. A left-turn control system for a connected autonomous vehicle, characterized in that: The system uses a left-turn control method for a connected autonomous vehicle as described in any one of claims 1-5, comprising: A road information acquisition module, wherein the road information acquisition module is used to acquire road traffic information; The calculation module is used to calculate the interaction intensity coefficient between the vehicle and oncoming vehicles and the predicted collision probability between the vehicle and all traffic participants based on road traffic information, as well as to determine traffic environment indicators, and then calculate the risk exposure index and communication confidence of the vehicle based on the traffic environment indicators. The judgment module is used to determine the decision state of the vehicle based on the risk exposure index, interaction intensity coefficient, predicted collision probability and communication confidence. The control module is used to control the vehicle to turn left based on the decision state.

7. A computer, characterized in that: It includes at least one processor and a memory, the memory storing a computer program configured to be executed by the processor to implement the left-turn control method for a connected autonomous vehicle as described in any one of claims 1-5.

8. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, on which a computer program is stored. The computer program can be executed by one or more processors to implement a left-turn control method for a connected autonomous vehicle as described in any one of claims 1-5.

Images