A method for generating traffic light-road topology control relations based on four-dimensional vehicle trajectories

Abstract

The present invention relates to a method for generating a control relationship between a signal light and a road topology based on a four-dimensional trajectory of a vehicle, and relates to the field of autonomous driving. The method comprises: obtaining status information of a target vehicle and position information of the target vehicle; inputting the position information and status information into a lane topology model, and outputting a set of vehicle passable road topologies corresponding to the target vehicle. By utilizing the four-dimensional trajectory of the vehicle at the intersection, a binding relationship between the signal light and the road topology is established, thereby obtaining a passable road topology based on the current signal light perception. The passable road topology calculated based on the current signal light status can help the autonomous driving vehicle to adjust its driving strategy in a timely manner based on the current signal light status and the offline road topology control relationship, which helps to make correct decisions in a complex and changing traffic environment and improve safety and driving efficiency.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving, and in particular to a method for generating a traffic light-road topology control relationship based on a four-dimensional vehicle trajectory. Background Art

[0002] A vehicle's perception of surrounding road and traffic signal information is a crucial component of autonomous driving. Enhancing this understanding of the environment can help autonomous driving systems make better decisions in complex scenarios.

[0003] Traffic lights, the most common traffic signal on roads, guide vehicles through orderly lanes. Traffic lights embody the trafficability relationships between lanes. The relationship between traffic lights and vehicles determines the lane controlled by each light, the duration of the traffic light cycle, and the topological relationship between lanes.

[0004] However, there is currently a lack of methods for collecting and obtaining the relationship between traffic lights and road topology, and it is also impossible to obtain drivable roads based on the relationship between traffic lights. Summary of the Invention

[0005] The present invention relates to a method for generating a traffic light-road topology control relationship based on a vehicle's four-dimensional trajectory, which can improve the safety and efficiency of autonomous driving. The method is applied to a computer device and includes:

[0006] Obtain the status information and location information of the target vehicle. The location information is used to indicate the traffic light position of the target vehicle, and the status information is used to indicate the start / stop status of the target vehicle.

[0007] The position information and status information are input into the lane topology model, and the output is the vehicle passable road topology set corresponding to the target vehicle. The lane topology model is a model generated based on historical vehicle four-dimensional trajectory data. The vehicle passable road topology set is used to indicate the expected trajectory of the target vehicle.

[0008] In an optional embodiment, the method further includes:

[0009] Obtain historical vehicle traffic information, which corresponds to the target intersection;

[0010] Generate vehicle four-dimensional trajectory information based on historical vehicle traffic information;

[0011] Combined with the vehicle's four-dimensional trajectory information, the traffic light cycle data corresponding to the intersection is determined based on a statistical voting algorithm;

[0012] Generate a lane topology model based on traffic light cycle data and vehicle four-dimensional trajectory information.

[0013] In an optional embodiment, generating vehicle four-dimensional trajectory information based on historical vehicle traffic information includes:

[0014] Divide the target intersection into lanes to obtain at least two independent lanes;

[0015] Obtaining lane vehicle traffic data corresponding to each independent lane;

[0016] Generate vehicle four-dimensional trajectory information based on the vehicle traffic data in the lane.

[0017] In an optional embodiment, combining the four-dimensional trajectory information of the vehicle and determining the traffic light cycle data corresponding to the target intersection based on a statistical voting algorithm includes:

[0018] Determining vehicle motion information corresponding to the independent lane based on the vehicle's four-dimensional trajectory information, the vehicle motion information including at least one of vehicle stop time information, vehicle start time information, and vehicle speed information;

[0019] Based on the vehicle motion information and in combination with an interval threshold, the signal light cycle data corresponding to the target intersection is determined. The interval threshold is used to merge the signal lights corresponding to at least two independent lanes of the target intersection.

[0020] In an optional embodiment, a lane topology model is generated based on traffic light cycle data and vehicle four-dimensional trajectory information, including:

[0021] Obtain the number of lanes, number of traffic lights, and lane associations corresponding to the target intersection;

[0022] Based on the number of lanes and their associations, combined with the vehicle's four-dimensional trajectory information, lane change information corresponding to the target intersection is generated;

[0023] Based on the lane change information and combined with the traffic light cycle data, a connected road set corresponding to the target intersection is generated. The connected road set includes connected road mode information and connected road time interval information;

[0024] Generate a lane topology model based on a set of connected roads.

[0025] In an optional embodiment, after obtaining the number of lanes, the number of traffic lights, and the lane association relationship corresponding to the target intersection, the following steps are included:

[0026] Based on the number of lanes and the lane association relationship, a road topology matrix corresponding to the target intersection is determined. The road topology matrix is ​​used to indicate the connectivity status of the target intersection.

[0027] In an optional embodiment, the lane change information is implemented as a triple data set;

[0028] The ternary data set includes lane change time data, initial lane data, and lane change data.

[0029] In an optional embodiment, the lane topology model is implemented as an offline model.

[0030] The beneficial effects brought about by the technical solution provided by the present invention include at least:

[0031] By leveraging the four-dimensional trajectories of vehicles at intersections, a binding relationship between traffic lights and road topology is established, allowing the navigable road topology to be derived based on the current traffic light perception. This navigable road topology, calculated based on the current traffic light status, enables autonomous vehicles to adjust their driving strategies in real time based on the current traffic light status and offline road topology control relationships. This helps autonomous vehicles make informed decisions in complex and changing traffic environments, improving safety and driving efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0033] Figure 1 A flow chart of a method for generating a traffic light-to-road topology control relationship based on a four-dimensional vehicle trajectory provided by an exemplary embodiment of the present invention is shown.

[0034] Figure 2 A schematic diagram of a lane topology model construction process provided by an exemplary embodiment of the present invention is shown.

[0035] Figure 3 A schematic diagram of a construction process of another lane topology model provided by an exemplary embodiment of the present invention is shown. DETAILED DESCRIPTION

[0036] To make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0037] Figure 1 A flow chart of a method for generating a traffic light-to-road topology control relationship based on a four-dimensional vehicle trajectory, provided by an exemplary embodiment of the present invention, is shown. This method is described using a computer device as an example. The method includes:

[0038] Step 101: Acquire the status information and location information of a target vehicle.

[0039] In the embodiment of the present invention, the position information is used to indicate the position of the traffic light where the target vehicle is located, and the status information is used to indicate the start / stop status of the target vehicle.

[0040] In an embodiment of the present invention, the target vehicle is a vehicle with an automatic driving function, and the computer device is implemented as a vehicle central control device equipped with an offline lane topology model. The vehicle central control device can receive status information and position information corresponding to the target vehicle through on-board sensors.

[0041] Step 102: input the position information and the state information into the lane topology model, and output a vehicle-passable road topology set corresponding to the target vehicle.

[0042] In the embodiment of the present invention, the lane topology model is a model generated based on historical vehicle four-dimensional trajectory data, and the vehicle-passable road topology set is used to indicate the expected passing trajectory of the target vehicle.

[0043] In one example, the passable road topology set is used to indicate the roads that a vehicle can travel under the current traffic light status. In another example, the passable road topology set is used to generate a prompt indicating that the vehicle's current status does not match the traffic light status. In another example, the passable road topology set is used to generate a prediction of the traffic light content when the vehicle travels from the current intersection to the next intersection. The present invention does not limit the form in which the passable road topology set is displayed on the visual interface of the vehicle's central control device.

[0044] It should be noted that the present invention also involves a process for generating a lane topology model. Figure 2 A schematic diagram of a lane topology model construction process provided by an exemplary embodiment of the present invention is shown. The method is described by taking the application of the method in a computer device as an example. The method includes:

[0045] Step 201: Obtain historical vehicle traffic information.

[0046] In the embodiment of the present invention, the historical vehicle traffic information corresponds to the target intersection. It should be noted that, in the embodiment of the present invention, the target intersection indicates the intersection of the one-way traffic process in the intersection. In the target intersection, vehicles usually have three driving states: left turn, straight ahead, and right turn.

[0047] Step 202: Generate vehicle four-dimensional trajectory information based on historical vehicle traffic information.

[0048] In the embodiment of the present invention, the four-dimensional trajectory information of the vehicle is realized as a set of the three-dimensional coordinates of the vehicle and the current time. In one example, N cThe vehicle's four-dimensional trajectory data set in where x i 、y i 、z i and t i It is a collection of three-dimensional coordinates and time, recording the three-dimensional coordinates and current time at every Δt time.

[0049] Step 203 : Determine the traffic light cycle data corresponding to the intersection based on a statistical voting algorithm in combination with the four-dimensional trajectory information of the vehicle.

[0050] In an embodiment of the present invention, a computer device can determine the speed of a vehicle based on the four-dimensional information of the vehicle, and further determine the driving state of the vehicle. In this case, based on the driving states of multiple vehicles, a voting algorithm can be used to determine the traffic light cycle at the target intersection.

[0051] Step 204 : Generate a lane topology model based on the traffic light cycle data and the vehicle's four-dimensional trajectory information.

[0052] In the embodiment of the present invention, the traffic light cycle data and the four-dimensional trajectory information of the vehicle are combined to generate a vehicle topology model suitable for single vehicle matching through information fusion of multiple intersections in multiple time periods.

[0053] Figure 3 A schematic diagram of a construction process of another lane topology model provided by an exemplary embodiment of the present invention is shown. The method is described by taking the application of the method in a computer device as an example. The method includes:

[0054] Step 301: Obtain historical vehicle traffic information.

[0055] This process corresponds to the process shown in step 201 and will not be described in detail here.

[0056] Step 302: perform lane division on the target intersection to obtain at least two independent lanes.

[0057] Step 303: Acquire lane vehicle traffic data corresponding to each independent lane.

[0058] Step 304: Generate vehicle four-dimensional trajectory information based on the vehicle traffic data in the lane.

[0059] The above process corresponds to the above embodiment, in which N c The vehicle's four-dimensional trajectory data set in where x i 、y i 、z i and t iIt is a collection of three-dimensional coordinates and time, recording the three-dimensional coordinates and current time at every Δt time.

[0060] Step 305 : Determine vehicle motion information corresponding to the independent lane based on the vehicle four-dimensional trajectory information.

[0061] In the embodiment of the present invention, v i Is the speed of the vehicle predicted by the model. If the predicted speed exists, use the provided speed. If the predicted speed does not exist, use the formula Calculated. When the speed at time t Less than the threshold v min When the vehicle is considered to be in a stopped state; when the speed at time t Greater than or equal to the threshold v min When , the vehicle is considered to be in the starting state. Based on this, the vehicle's stop time point set Start time point collection For the set of stop time points s r and the starting time point set G r Grouping is performed, and the interval threshold θ1 is set. Two adjacent time points less than θ1 are grouped into the same group. The time points in the same group are the average of all time points. Each group is arranged in the order of time points, and the red light time interval set Δt is obtained by subtracting the previous adjacent stop time from the start time. red , subtract the previous adjacent start time from the stop time to get the green light time interval set Δt green .

[0062] In the embodiment of the present invention, the vehicle motion information includes at least one of vehicle stop time information, vehicle start time information, and vehicle speed information.

[0063] Step 306 : Based on the vehicle motion information and the interval threshold, determine the traffic light cycle data corresponding to the target intersection.

[0064] In the embodiment of the present invention, based on two time interval sets, the final time interval is obtained by statistical voting. The longest cycle time is set to t max , with a smaller Δt v Divide into intervals intervals, and obtain the interval set The i-th interval is denoted as I i =[(i-1)Δt v ,min(iΔt v ,t max )) is an interval with a closed left and an open right. Each time interval in the time interval set is assigned to an interval of corresponding length, and the interval with the highest frequency is counted. As the cycle obtained by the current statistical vote. According to the red light time interval set Δt at different times of the day red and the green light time interval set Δt green By obtaining the traffic light cycle information for each time period, we can estimate the traffic light status of the current lane based on real-time information. Repeating the above operation for each lane can provide an estimate of the red light status of all lanes. By setting an interval threshold θ2, lanes with red light cycle information differences less than θ2 are grouped together, thus merging the red lights controlling multiple lanes.

[0065] That is, in the embodiment of the present invention, the interval threshold is used to merge the traffic lights corresponding to at least two independent lanes of the target intersection.

[0066] Step 307: Obtain the number of lanes, the number of traffic lights, and the lane association relationship corresponding to the target intersection.

[0067] Step 308 : Based on the number of lanes and the lane association relationship, and in combination with the vehicle's four-dimensional trajectory information, generate lane change information corresponding to the target intersection.

[0068] Step 309 : Based on the lane change information and combined with the traffic light cycle data, a connected road set corresponding to the target intersection is generated.

[0069] In the embodiment of the present invention, the connectable road set includes connectable road mode information and connectable road time interval information.

[0070] Optionally, in an embodiment of the present invention, a road topology matrix corresponding to the target intersection is determined based on the number of lanes and the lane association relationship, and the road topology matrix is ​​used to indicate the connectivity status of the target intersection.

[0071] In the embodiment of the present invention, the lane change information is implemented as a ternary data set; the ternary data set includes lane change time data, initial lane data, and lane data after the change. That is, in one example, suppose there are N intersections. r lanes There is N l Traffic lights According to the existing lane topology model, the size of N r ×N r The road topology matrix A, where A i，j Represents lane r i and r j Is it connected: If it is connected, A i，j =1; if disconnected, A i，j = 0. According to the vehicle four-dimensional trajectory data, lane change information is extracted. For P vehicles at the intersection in a certain period of time, we can extract the lane change information triple set C = {(r11 , r 12 , t1), (r 21 , r 22 , t2),…,(r P1 , r P2 , t P )}. Among them, the triple (r i1 , r i2 , t i ) represents the i There is a car coming from the driveway i1 Turned into lane r i2 Synchronize the triplet with the previously obtained signal light cycle to obtain the relationship set of Q cycles in this time period in is the time interval I in the i-th cycle i The set of connected roads within

[0072] Step 310: Generate a lane topology model based on the connected road set.

[0073] In the embodiment of the present invention, the lane topology model is a model generated by integrating multiple time periods.

[0074] In summary, the method provided by the embodiments of the present invention establishes a binding relationship between traffic lights and road topology by utilizing the four-dimensional trajectories of vehicles at an intersection, thereby deriving a navigable road topology based on the current traffic light perception. This navigable road topology calculated based on the current traffic light status can help autonomous vehicles promptly adjust their driving strategies based on the current traffic light status and offline road topology control relationships, facilitating accurate decision-making in complex and changing traffic environments, and improving safety and driving efficiency.

[0075] The above are only optional embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for generating a traffic light-road topology control relationship based on a vehicle's four-dimensional trajectory, characterized in that: The method is applied to a computer device, and the method includes: Obtaining status information of a target vehicle and position information of the target vehicle, wherein the position information is used to indicate the traffic light position of the target vehicle, and the status information is used to indicate the start / stop status of the target vehicle; Inputting the position information and the state information into a lane topology model, and outputting a vehicle passable road topology set corresponding to the target vehicle, wherein the lane topology model is a model generated based on historical vehicle four-dimensional trajectory data, and the vehicle passable road topology set is used to indicate the expected passing trajectory of the target vehicle; The method further comprises: Obtaining historical vehicle traffic information, wherein the historical vehicle traffic information corresponds to a target intersection; generating vehicle four-dimensional trajectory information based on the historical vehicle traffic information; Determining the traffic light cycle data corresponding to the intersection based on a statistical voting algorithm in combination with the four-dimensional trajectory information of the vehicle; generating the lane topology model based on the signal light cycle data and the four-dimensional trajectory information of the vehicle; The generating of vehicle four-dimensional trajectory information based on the historical vehicle traffic information includes: Performing lane division on the target intersection to obtain at least two independent lanes; Obtaining lane vehicle traffic data corresponding to each independent lane; generating four-dimensional trajectory information of the vehicle based on the vehicle traffic data in the lane; The step of determining the traffic light cycle data corresponding to the target intersection based on a statistical voting algorithm in combination with the four-dimensional vehicle trajectory information includes: determining vehicle motion information corresponding to the independent lane based on the vehicle four-dimensional trajectory information, the vehicle motion information including at least one of vehicle stop time information, vehicle start time information, and vehicle speed information; Based on the vehicle motion information and in combination with an interval threshold, traffic light cycle data corresponding to the target intersection is determined, and the interval threshold is used to merge traffic lights corresponding to at least two independent lanes of the target intersection.

2. The method for generating a traffic light-to-road topology control relationship based on a vehicle's four-dimensional trajectory according to claim 1, characterized in that: The generating the lane topology model based on the traffic light cycle data and the four-dimensional vehicle trajectory information includes: Obtain the number of lanes, number of traffic lights, and lane associations corresponding to the target intersection; Based on the number of lanes and the lane association relationship, combined with the four-dimensional trajectory information of the vehicle, generating lane change information corresponding to the target intersection; generating a connected road set corresponding to the target intersection based on the lane change information and in combination with the traffic light cycle data, wherein the connected road set includes connectable road mode information and connectable road time interval information; The lane topology model is generated based on the connected road set.

3. The method for generating a traffic light-to-road topology control relationship based on a vehicle's four-dimensional trajectory according to claim 2, characterized in that: After obtaining the number of lanes, the number of traffic lights, and the lane association relationship corresponding to the target intersection, the method includes: Based on the number of lanes and the lane association relationship, a road topology matrix corresponding to the target intersection is determined, where the road topology matrix is ​​used to indicate a connectivity state of the target intersection.

4. The method for generating a traffic light-to-road topology control relationship based on a vehicle's four-dimensional trajectory according to claim 2, characterized in that: The lane change information is implemented as a three-dimensional data set; The three-dimensional data set includes lane change time data, initial lane data, and lane change data.

5. The method for generating a traffic light-to-road topology control relationship based on a vehicle's four-dimensional trajectory according to claim 1, characterized in that: The lane topology model is implemented as an offline model.

Images