WO2022133775A1 - 轨迹数据处理方法、装置、计算机设备和存储介质 - Google Patents
轨迹数据处理方法、装置、计算机设备和存储介质 Download PDFInfo
- Publication number
- WO2022133775A1 WO2022133775A1 PCT/CN2020/138520 CN2020138520W WO2022133775A1 WO 2022133775 A1 WO2022133775 A1 WO 2022133775A1 CN 2020138520 W CN2020138520 W CN 2020138520W WO 2022133775 A1 WO2022133775 A1 WO 2022133775A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- trajectory
- data
- frame
- planning parameters
- simulated
- Prior art date
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 12
- 238000004088 simulation Methods 0.000 claims abstract description 137
- 238000000034 method Methods 0.000 claims description 51
- 238000004364 calculation method Methods 0.000 claims description 28
- 239000000284 extract Substances 0.000 claims description 25
- 238000012545 processing Methods 0.000 claims description 8
- 230000001131 transforming effect Effects 0.000 claims 1
- 230000008569 process Effects 0.000 description 20
- 238000004422 calculation algorithm Methods 0.000 description 18
- 230000006870 function Effects 0.000 description 12
- 230000007613 environmental effect Effects 0.000 description 11
- 230000001133 acceleration Effects 0.000 description 7
- 230000036461 convulsion Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- 230000006399 behavior Effects 0.000 description 4
- 239000000306 component Substances 0.000 description 3
- 238000013398 bayesian method Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 239000008358 core component Substances 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000012634 fragment Substances 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
Definitions
- the present application relates to a trajectory data processing method, apparatus, computer equipment and storage medium.
- trajectory planning is the core component of driving behavior.
- the trajectory planning algorithm used by the trajectory planner will be A large number of parameters are set in the network. Due to the complex balance between the parameters, it is extremely difficult to adjust the parameters to adapt to diverse traffic scenarios.
- the traditional method is to adjust the parameters for diverse traffic scenarios by learning the trajectory planning parameters in the trajectory planning algorithm, such as the inverse reinforcement learning method.
- the trajectory planning algorithm after parameter adjustment is used for trajectory planning.
- the traditional method is only suitable for a specific trajectory planning algorithm.
- the parameters of the trajectory planning algorithm will be inaccurate.
- a trajectory data processing method, apparatus, computer device and storage medium capable of improving the parameter accuracy of a trajectory planning algorithm are provided.
- a trajectory data processing method comprising:
- Adjust the current trajectory planning parameters according to the cost value to obtain the adjusted trajectory planning parameters use the adjusted trajectory planning parameters as the current trajectory planning parameters, and return to the data simulation unit through the data
- Trajectory planning parameters are used to obtain the target vehicle.
- a trajectory data processing device comprising:
- a data acquisition module is used to acquire the trajectory data of the target vehicle within a preset time period, road condition data within a preset distance from the target vehicle, road information corresponding to the target vehicle within the preset time period, and route data of the target vehicle within the preset time period;
- a data simulation module for inputting the trajectory data, the road condition data, the road information and the route data into a data simulation unit, where the data simulation unit includes current trajectory planning parameters;
- the unit generates a simulated trajectory of the target vehicle according to the trajectory data, the road condition data, the road information, the route data and the current trajectory planning parameters;
- a cost calculation module for performing cost calculation on the simulated trajectory to obtain a cost value corresponding to the simulated trajectory
- a parameter adjustment module configured to adjust the current trajectory planning parameters according to the cost value, obtain adjusted trajectory planning parameters, use the adjusted trajectory planning parameters as the current trajectory planning parameters, and return to the The step of generating, by the data simulation unit, a simulated trajectory of the target vehicle according to the trajectory data, the road condition data, the road information, the route data and the current trajectory planning parameters, until a preset condition is met, Stop the parameter adjustment to get the target trajectory planning parameters.
- a computer device comprising a memory and one or more processors, the memory having computer-readable instructions stored therein, the computer-readable instructions, when executed by the processor, cause the one or more processors to execute The following steps:
- Adjust the current trajectory planning parameters according to the cost value to obtain the adjusted trajectory planning parameters use the adjusted trajectory planning parameters as the current trajectory planning parameters, and return to the data simulation unit through the data
- Trajectory planning parameters are used to obtain the target vehicle.
- One or more non-volatile computer-readable storage media storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the following steps:
- Adjust the current trajectory planning parameters according to the cost value to obtain the adjusted trajectory planning parameters use the adjusted trajectory planning parameters as the current trajectory planning parameters, and return to the data simulation unit through the data
- Trajectory planning parameters are used to obtain the target vehicle.
- FIG. 1 is an application environment diagram of a trajectory data processing method in one or more embodiments.
- FIG. 2 is a schematic flowchart of a method for processing trajectory data in one or more embodiments.
- FIG. 3 is a schematic flowchart of a step of adjusting current trajectory planning parameters in one or more embodiments.
- FIG. 4 is a block diagram of a trajectory data processing apparatus in one or more embodiments.
- FIG. 5 is a block diagram of a computer device in one or more embodiments.
- the trajectory data processing method provided in this application can be applied to the application environment shown in FIG. 1 .
- the terminal 102 communicates with the server 104 through a network.
- the terminal 102 can send a parameter adjustment request to the server 104.
- the server 104 parses the parameter adjustment request to obtain the trajectory data of the target vehicle and the target vehicle within the preset time period.
- the server 104 issues a parameter adjustment instruction to the terminal 102, and the terminal 102 obtains the trajectory data of the target vehicle within the preset time period, the road condition data within the preset distance range from the target vehicle, and the location of the target vehicle according to the parameter adjustment instruction.
- the road information and the route data of the target vehicle within the preset time period are sent to the server 104 .
- the server 104 inputs the trajectory data, road condition data, road information and route data into the data simulation unit, the data simulation unit includes the current trajectory planning parameters, and the data simulation unit uses the trajectory data, road condition data, road information, route data and current trajectory according to the trajectory data, road condition data, road information, route data and current trajectory.
- the planning parameters generate the simulated trajectory of the target vehicle, so that the server 104 performs cost calculation on the simulated trajectory to obtain the cost value corresponding to the simulated trajectory, and then adjusts the current trajectory planning parameters according to the cost value to obtain the adjusted trajectory planning parameters.
- the trajectory planning parameters are used as the current trajectory planning parameters, and return to the step of generating the simulated trajectory of the target vehicle according to the trajectory data, road condition data, road information, route data and current trajectory planning parameters through the data simulation unit, until the preset conditions are met, and the parameters are stopped. Adjust to get the target planning parameters.
- the terminal 102 can be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers and portable wearable devices.
- the server 104 can be implemented by an independent server or a server cluster composed of multiple servers.
- a method for processing trajectory data is provided, which is described by taking the method applied to the server in FIG. 1 as an example, including the following steps:
- Step 202 Acquire trajectory data of the target vehicle within a preset time period, road condition data within a preset distance from the target vehicle, road information where the target vehicle is located, and route data of the target vehicle within the preset time period.
- trajectory planning is the core component of driving behavior.
- the trajectory planning algorithm used by the trajectory planner will be A large number of parameters are set in.
- the driving route for a period of time can be planned based on location information, surrounding environment information, etc., and specific actions can be performed to change the driving state of the autonomous vehicle, such as parking, going straight, changing lanes, and turning.
- the trajectory data processing method provided in this application can be applied to any trajectory planning algorithm to determine the optimal trajectory planning parameters of any complex and diverse traffic scene.
- the trajectory planning algorithm may include a rapidly expanding random tree (rapidly -exploring random tree (RRT) method, visibility map, probability roadmap method (PRM), etc.
- Traffic scenarios can include lane changing scenarios, turning scenarios, lane keeping scenarios, parking scenarios, etc. .
- the trajectory data of the target vehicle within the preset time period refers to the trajectory data of the target vehicle during the actual driving process
- the road condition data within the preset distance from the target vehicle refers to the environment within the preset distance from the vehicle. Trajectory data during driving. Both track data and road condition data are recorded historical data.
- the preset time period is the time period selected for adjusting the trajectory planning parameters.
- the target vehicle refers to any vehicle used for automatic driving, and its trajectory data is used to adjust the trajectory planning parameters.
- the road condition data within the preset distance range from the target vehicle refers to the trajectory data of the environmental vehicles around the target vehicle.
- the route data refers to the destination data that the target vehicle needs to reach.
- the server may obtain the parameter adjustment request sent by the terminal, parse the parameter adjustment request, and obtain the trajectory data of the target vehicle within the preset time period, the road condition data within the preset distance from the target vehicle, and the road where the target vehicle is located. information and route data of the target vehicle over a preset time period.
- the server can also send parameter adjustment instructions to the terminal, and the terminal obtains the trajectory data of the target vehicle within the preset time period, the road condition data within the preset distance from the target vehicle, the road information and the target vehicle where the target vehicle is located according to the parameter adjustment instructions.
- the route data of the vehicle within a preset time period, and the obtained data is sent to the server.
- the trajectory data of the target vehicle may include the position, velocity, acceleration, jerk, etc. of the target vehicle at each moment.
- the road condition data within the preset distance range from the target vehicle includes the position, speed, acceleration, jerk, etc. of the environmental vehicle within the preset distance range from the target vehicle at each moment.
- the road information where the target vehicle is located may include traffic lanes where the target vehicle is located, road boundary points, lane line information, etc., and the road information may be obtained from a high-precision map.
- the route data may include the destination identification of the target vehicle, the location of the destination, and the like.
- the preset time period may be 40s.
- the preset distance range may be a range with the location of the target vehicle as the center and 10m as the distance radius.
- the trajectory data of the target vehicle within the preset time period, the road condition data within the preset distance from the target vehicle, the road information where the target vehicle is located, and the route data of the target vehicle within the preset time period can be obtained from the terminal from the on-board computer equipment. extracted.
- the on-board computer equipment is installed in the self-driving vehicle.
- the on-board computer equipment can convert the trajectory data of the target vehicle and the road condition data within the damaged distance from the target vehicle, the road information where the target vehicle is located, and the target vehicle in a preset time period.
- the route data is stored in it.
- the server can also obtain the parameter adjustment request sent by the terminal, parse the parameter adjustment request, obtain the request parameters, and extract the trajectory data of the target vehicle in the preset time period from the on-board computer equipment according to the request parameters, and the target vehicle is in the pre-defined time period. Set the road condition data within the distance range, the road information where the target vehicle is located, and the route data of the target vehicle within a preset time period.
- Step 204 input the trajectory data, road condition data, road information and route data into the data simulation unit, the data simulation unit includes the current trajectory planning parameters through the data simulation unit according to the trajectory data, road condition data, road information, route data and current trajectory.
- the planning parameters generate the simulated trajectory of the target vehicle.
- the data simulation unit is used for planning a trajectory and generating a simulated trajectory according to the planned trajectory.
- the simulated trajectory refers to the trajectory of the target vehicle running in the data simulation unit during the data simulation process.
- the server After acquiring the trajectory data, road condition data, road information and route data, the server calls the data simulation unit, and inputs the trajectory data, road condition data, road information and route data into the data simulation unit, which includes the current trajectory planning parameters , so that the trajectory of the target vehicle running in the preset time period can be simulated by the data simulation unit according to the trajectory data, road condition data, road information, route data and current trajectory planning parameters.
- the server can process the trajectory data, road condition data, road information and route data into frames through the data simulation unit to obtain multi-frame data.
- Each frame of data includes trajectory data, road condition data, road information and route data corresponding to the frame.
- a simulated trajectory segment corresponding to each frame of data is generated by the data simulation unit according to the current trajectory planning parameters.
- the current trajectory planning parameters may be pre-specified parameters that need to be adjusted.
- the simulated trajectory segment refers to the trajectory that controls the target vehicle to run in each frame during the data simulation process.
- the simulated track segments corresponding to each frame of data are sequentially generated by the data simulation unit according to the time sequence between each frame of data. Then, the simulated trajectory of the target vehicle is generated by the data simulation unit according to the simulated trajectory segments corresponding to the multi-frame data.
- the simulated trajectory can include position, velocity, acceleration, jerk, etc. at each moment.
- Step 206 Perform cost calculation on the simulated trajectory to obtain a cost value corresponding to the simulated trajectory.
- the server can calculate the cost of the simulated trajectory.
- the cost can be calculated by calculating the similarity between the simulated trajectory and the trajectory data of the target vehicle. During the trajectory simulation process, the parameter values of each driving parameter in the target vehicle and the environmental vehicle within the preset distance range are compared to calculate the cost.
- the method of cost calculation is not limited, and a corresponding cost calculation method can be adopted according to actual needs.
- performing cost calculation on the simulated trajectory to obtain the cost value corresponding to the simulated trajectory includes: calculating the similarity between the trajectory data and the simulated trajectory; and calculating the cost value corresponding to the simulated trajectory according to the similarity.
- the server can use the imitation cost method to calculate the distance between the trajectory data and the simulated trajectory, and use the distance as the similarity.
- the server determines multiple cost values to be calculated according to the cost function in the imitation cost method, and then calculates the cost value corresponding to the simulated trajectory according to the calculated similarity, cost function, and cost weight corresponding to each cost value.
- the cost weights corresponding to multiple cost values in the cost function can be used as trajectory planning parameters.
- the cost function is the weighted sum of efficiency cost, safety cost, and comfort cost
- the corresponding weights of these three costs can be The weights are used as trajectory planning parameters.
- the cost calculation is performed by calculating the similarity between the simulated trajectory of the trajectory data and the trajectory data of the target vehicle, so that a driving trajectory that is more in line with human driving habits can be obtained in the subsequent automatic driving process, thereby improving the trajectory planning efficiency. flexibility.
- performing cost calculation on the simulated trajectory to obtain the cost value corresponding to the simulated trajectory includes: determining the driving trajectory data corresponding to the preset driving parameters in the simulated trajectory; determining the driving trajectory data corresponding to the preset driving parameters in the road condition data Driving road condition data; calculate the cost value corresponding to the simulated trajectory according to the driving trajectory data and the driving road condition data.
- the driving trajectory data corresponding to the preset driving parameters can be determined in the simulated trajectory, and
- the driving road condition data corresponding to the preset driving parameters is determined in the road condition data.
- the preset driving parameters may include distance, speed, acceleration, jerk, and the like.
- the driving trajectory data may include parameter values of each driving parameter corresponding to the target vehicle at each moment.
- the road condition data includes trajectory data of environmental vehicles within a preset distance from the target vehicle.
- the driving track data includes parameter values of each driving parameter corresponding to the environmental vehicle at each moment.
- the server uses the self-supervised cost method to calculate the cost value corresponding to the simulated trajectory according to the parameter values of each driving parameter of the target vehicle and the environmental vehicle at each moment.
- the cost value corresponding to the simulated trajectory can be a weighted sum of multiple cost values.
- the server determines a plurality of cost values to be calculated according to the cost function in the self-supervised cost method, and then according to the parameter values, cost functions, and the corresponding cost values of each driving parameter of the target vehicle and the environmental vehicle at each moment
- the cost weight is calculated to obtain the cost value corresponding to the simulated trajectory.
- the cost weights corresponding to multiple cost values in the cost function can be used as trajectory planning parameters.
- the cost function is the weighted sum of the distance cost and the speed cost
- the weights corresponding to the distance cost and the speed cost need to be automatically adjusted
- the weights corresponding to the two costs can be used as trajectory planning parameters.
- the accuracy of trajectory planning ensures the safety of autonomous driving.
- Step 208 Adjust the current trajectory planning parameters according to the cost value, obtain the adjusted trajectory planning parameters, use the adjusted trajectory planning parameters as the current trajectory planning parameters, and return to the data simulation unit according to the trajectory data, road condition data, and road information.
- the data planning unit includes multiple trajectory planning parameters.
- the trajectory planning parameters can be determined according to the trajectory planning algorithm used by the data simulation unit, and the trajectory planning algorithm needs to use the cost function to calculate the cost. Therefore, the trajectory planning parameters can be calculated according to the cost function. to be sure.
- the trajectory planning parameters corresponding to different cost functions can be different.
- the server can set the data planning unit according to the trajectory planning parameters in the cost function, so that the subsequent data planning unit can perform trajectory planning according to the trajectory planning parameters.
- the server may pre-designate the trajectory planning parameters and preset the parameter adjustment range of the trajectory planning parameters in a large number of trajectory planning parameters of the data planning unit. Take the pre-specified trajectory planning parameters as the current trajectory planning parameters.
- the server can adjust the current trajectory planning parameters in the data simulation unit according to the cost value and the trajectory planning range, and use the adjusted trajectory planning parameters as the current trajectory Planning parameters, return to the step of generating the simulated trajectory of the target vehicle according to the trajectory data, road condition data, road information, route data and current trajectory planning parameters by the data simulation unit, and repeat the parameter adjustment until the preset conditions are met.
- the preset condition may be that the current trajectory planning parameter has reached the optimal value within the parameter adjustment range.
- the optimal value is pre-stored in the server so that the server can adjust the parameters.
- the server may use a Bayesian method to generate adjusted trajectory planning parameters according to the cost value, so as to use the adjusted trajectory planning parameters as the current trajectory planning parameters and input them into the data simulation unit.
- the server stops the parameter adjustment, and uses the trajectory planning parameters at this time as the target trajectory planning parameters.
- the server can iteratively adjust the weight of the two costs within the parameter adjustment range according to the cost value.
- the weight of the cost reaches the optimal value, the parameter adjustment is stopped, and the weight of the efficiency cost and the weight of the safety cost at this time are used as the target trajectory planning parameters.
- the server may store the target trajectory planning parameters in the data simulation unit.
- the data simulation unit can use the target trajectory planning parameters to plan the trajectory of the autonomous vehicle. Since the target trajectory planning parameters are pre-adjusted and accurate trajectory planning parameters, the accuracy of trajectory planning can be improved.
- the route data within the preset time period is input into the data simulation unit, and the simulated trajectory of the target vehicle is generated by the data simulation unit according to the trajectory data, road condition data, road information, route data and current trajectory planning parameters, so as to calculate the cost of the simulated trajectory , obtain the cost value corresponding to the simulated trajectory, and then adjust the current trajectory planning parameters according to the cost value, take the adjusted trajectory planning parameters as the current trajectory planning parameters, and return to the data simulation unit according to the trajectory data, road condition data, road information,
- trajectory planning parameters In the process of adjusting the trajectory planning parameters, it does not pay attention to the trajectory planning method of the trajectory planner, and can adjust the parameters independently of the trajectory planner. Therefore, it can be applied to any trajectory planning method in diverse traffic scenarios, effectively improving trajectory planning.
- the parameter accuracy of the algorithm At the same time, only the current trajectory planning parameters and the setting of preset conditions are needed to realize the automatic adjustment of the parameters, which improves the adjustment efficiency of the parameters of the trajectory planning algorithm.
- the step of generating the simulated trajectory of the target vehicle according to the trajectory data, road condition data, road information, route data and current trajectory planning parameters by the data simulation unit includes: extracting each frame from the trajectory data by the data simulation unit trajectory data, extract the road condition data of each frame from the road condition data, extract the road information of each frame from the road information, and extract the route data of each frame from the route data; according to the trajectory data of each frame, each frame The road condition data of the frame, the road information of each frame, the route data of each frame and the current trajectory planning parameters generate the simulated trajectory segment of each frame; after obtaining the simulated trajectory segment of the previous frame, according to the trajectory data of the next frame , the road condition data of the next frame, the road information of the next frame, the route data of the next frame and the current trajectory planning parameters to generate the simulated trajectory segment corresponding to the next frame until the simulated trajectory segment of the last frame is obtained.
- the simulated trajectory segment obtains the simulated trajectory of the target vehicle.
- the trajectory data may include the position, velocity, acceleration, jerk, etc. of the target vehicle at each moment.
- the road condition data includes the position, speed, acceleration, jerk, etc. of the environmental vehicle within a preset distance from the target vehicle at each moment.
- the road information may include traffic lanes where the target vehicle is located, road boundary points, lane line information, etc.
- the road information may be obtained from a high-precision map.
- the route data may include the destination identification of the target vehicle, the location of the destination, and the like.
- the preset time period may be 40s.
- the preset distance range may be a range with the location of the target vehicle as the center and 10m as the distance radius.
- the data simulation unit can divide the trajectory data, road condition data, road information and route data into frames process to obtain multiple frames of data.
- Each frame of data includes trajectory data, road condition data, road information and route data corresponding to the frame. Therefore, the data simulation unit sequentially extracts the data of each frame according to the time sequence between the data of each frame, and performs trajectory planning on the data of each frame according to the current trajectory planning parameters to obtain the trajectory planning result.
- the data simulation unit controls the target vehicle to run for a period of time according to the trajectory planning result until the time stamp corresponding to the next frame is reached, and the simulation trajectory segment corresponding to the frame is obtained.
- the simulated trajectory segment refers to the trajectory that controls the target vehicle to run in each frame during the data simulation process.
- the data simulation unit After the data simulation unit generates the simulation track segment of the previous frame, it obtains the data of the next frame, and generates the simulation track segment corresponding to the next frame according to the data of the next frame, until the simulation track segment of the last frame is obtained, and then the data simulation is performed.
- the unit obtains the simulated trajectory of the target vehicle according to the simulated trajectory segments of multiple frames.
- the data of the next frame includes trajectory data of the next frame, road condition data of the next frame, road information of the next frame, and route data of the next frame.
- the simulated trajectory can include position, velocity, acceleration, jerk, etc. at each moment.
- the number of simulation frames of the data simulation unit may be fixed, so the number of cycles for generating simulation segments is also fixed.
- the simulated trajectory segment of each frame is generated according to the trajectory data of each frame, the road condition data of each frame, the road information of each frame, the route data of each frame and the current trajectory planning parameters.
- the simulated trajectory corresponding to the next frame is generated according to the trajectory data of the next frame, the road condition data of the next frame, the road information of the next frame, the route data of the next frame and the current trajectory planning parameters.
- segment until the simulated trajectory segment of the last frame is obtained, and the simulated trajectory of the target vehicle is obtained according to the simulated trajectory segment of multiple frames. It is beneficial to subsequently calculate the cost value according to the simulated trajectory, so as to realize the adjustment of the trajectory planning parameters.
- the data simulation unit includes a simulator and a trajectory planner
- the method further includes: sending trajectory data and road condition data to the simulator in the data simulation unit, and sending road information and route data to the data simulation
- the trajectory planner in the unit, the trajectory planner includes the current trajectory planning parameters; the trajectory data of each frame is extracted from the trajectory data by the simulator and the road condition data of each frame is extracted from the road condition data.
- the trajectory data and the road condition data of each frame are sent to the trajectory planner; the road information of each frame is extracted from the road information by the trajectory planner, and the route data of each frame is extracted from the route data.
- the controller According to the trajectory data of each frame , the road condition data of each frame, the road information of each frame, the route data of each frame and the current trajectory planning parameters to generate the trajectory planning result corresponding to the corresponding frame, and send the trajectory planning result corresponding to the corresponding frame to the simulator;
- the controller generates simulated trajectory segments of the corresponding frames according to the trajectory planning results.
- the data simulation unit includes a simulator and a trajectory planner, and the simulator is used for generating a simulated trajectory according to the trajectory planning result sent by the trajectory planner.
- the trajectory planner is used to generate the trajectory planning result corresponding to the corresponding frame according to the trajectory data of each frame, the road condition data of each frame, the road information of each frame, the route data of each frame and the current trajectory planning parameters sent by the simulator.
- the simulator sequentially extracts the trajectory data of each frame from the trajectory data and the road condition data of each frame from the road condition data in chronological order, and sends the extracted data to the trajectory planner.
- the trajectory planner extracts the road information of the corresponding frame from the road information and the route data of the corresponding frame from the route data according to the received data, and can use the extracted trajectory data, road condition data, road information and route data of the same frame as one. frame data.
- the trajectory planner can generate a trajectory planning result corresponding to the corresponding frame according to the frame data and the current trajectory planning parameters, and send the trajectory planning result corresponding to the corresponding frame to the simulator.
- the simulator thus controls the target vehicle to drive according to the trajectory planning results until the time stamp corresponding to the next frame is reached.
- the extracted trajectory data of the next frame and the road condition data of the next frame are sent to the trajectory planner.
- the trajectory planner generates the trajectory planning result corresponding to the next frame according to the data of the next frame and the current trajectory planning parameters, and sends the trajectory planning result corresponding to the next frame to the simulator, and the simulator controls the target vehicle to drive according to the trajectory planning result. until the timestamp corresponding to the next frame is reached.
- the simulator and the trajectory planner repeat the above steps of generating the simulated trajectory segment until the simulated trajectory segment of the last frame is obtained.
- the final simulator outputs the simulated trajectory of the target vehicle according to the simulated trajectory segments of multiple frames.
- the trajectory planner is not concerned.
- the specific trajectory planning method of the target vehicle only needs the trajectory planning result output by the trajectory planner to generate the simulated trajectory of the target vehicle.
- the parameter adjustment is realized independently of the trajectory planner, which can be applied to any trajectory planning method in diverse traffic scenarios, and effectively improves the parameter accuracy of the trajectory planning algorithm.
- the step of adjusting the current trajectory planning parameters includes:
- Step 302 Input the current trajectory planning parameters and the cost value into the non-gradient optimizer, and adjust the current trajectory planning parameters according to the cost value within the parameter adjustment range through the non-gradient optimizer, obtain the adjusted trajectory planning parameters, and adjust the parameters.
- the resulting trajectory planning parameters are sent to the data simulation unit.
- Step 304 using the adjusted trajectory planning parameters as the current trajectory planning parameters by the data simulation unit, and returning to the process of generating the simulated trajectory of the target vehicle according to the trajectory data, road condition data, road information, route data and current trajectory planning parameters by the data simulation unit. Steps until the preset conditions are met, the parameter adjustment is stopped, and the target trajectory planning parameters are obtained.
- the current trajectory planning parameters may be parameters preset by the server and need to be adjusted.
- the server acquires the current trajectory planning parameters of the data simulation unit.
- the data simulation unit includes a simulator and a trajectory planner, and the current trajectory planning parameters are the parameters of the trajectory planner.
- the server can call the non-gradient optimizer, input the current trajectory planning parameters and cost value into the non-gradient optimizer, and use the Bayesian method to adjust the current trajectory planning parameters within the parameter adjustment range through the non-gradient optimizer.
- the non-gradient optimizer inputs the adjusted trajectory planning parameters into the data simulation unit, and the data simulation unit uses the adjusted trajectory planning parameters as the current trajectory planning parameters, and returns to the data simulation unit according to the trajectory data, road conditions
- the preset condition may be that the current trajectory planning parameter has reached the optimal value within the parameter adjustment range.
- a new trajectory planning parameter can be quickly calculated, so that a simulated trajectory is generated again according to the new trajectory planning parameters, and the simulated trajectory is Perform cost calculation to obtain the cost value corresponding to the simulated trajectory, adjust the current trajectory planning parameters within the parameter adjustment range through the non-gradient optimizer, and obtain the target trajectory planning parameters through an iterative loop.
- a trajectory data processing device including: a data acquisition module 402, a data simulation module 404, a cost calculation module 406 and a parameter adjustment module 408, wherein:
- the data acquisition module 402 is used to acquire the trajectory data of the target vehicle in the preset time period, the road condition data within the preset distance range from the target vehicle, the road information corresponding to the target vehicle in the preset time period, and the preset time period of the target vehicle. Route data for the time period.
- the data simulation module 404 is used to input the trajectory data, road condition data, road information and route data into the data simulation unit, the data simulation unit includes the current trajectory planning parameters, and the data simulation unit according to the trajectory data, road condition data, road information, The route data and the current trajectory planning parameters generate the simulated trajectory of the target vehicle.
- the cost calculation module 406 is configured to perform cost calculation on the simulated trajectory to obtain the cost value corresponding to the simulated trajectory.
- the parameter adjustment module 408 is used to adjust the current trajectory planning parameters according to the cost value, obtain the adjusted trajectory planning parameters, use the adjusted trajectory planning parameters as the current trajectory planning parameters, and return to the data simulation unit according to the trajectory data, road conditions The step of generating the simulated trajectory of the target vehicle from the data, road information, route data and current trajectory planning parameters, until the preset conditions are met, then the parameter adjustment is stopped, and the target trajectory planning parameters are obtained.
- the data simulation module 404 is further configured to use the data simulation unit to extract the trajectory data of each frame from the trajectory data, extract the road condition data of each frame from the road condition data, and extract each frame from the road information and extract the route data of each frame from the route data; generate the route data according to the trajectory data of each frame, the road condition data of each frame, the road information of each frame, the route data of each frame and the current trajectory planning parameters.
- the simulated trajectory fragment of each frame after obtaining the simulated trajectory fragment of the previous frame, according to the trajectory data of the next frame, the road condition data of the next frame, the road information of the next frame, the route data of the next frame and the current trajectory
- the planning parameters generate the simulated trajectory segment corresponding to the next frame until the simulated trajectory segment of the last frame is obtained, and the simulated trajectory of the target vehicle is obtained according to the simulated trajectory segments of multiple frames.
- the data simulation unit includes a simulator and a trajectory planner
- the data simulation module 404 is further configured to send trajectory data and road condition data to the simulator in the data simulation unit, and send road information and route data to
- the trajectory planner in the data simulation unit the trajectory planner includes the current trajectory planning parameters; the trajectory data of each frame is extracted from the trajectory data by the simulator and the road condition data of each frame is extracted from the road condition data.
- the trajectory data of the frame and the road condition data of each frame are sent to the trajectory planner; the road information of each frame is extracted from the road information by the trajectory planner, and the route data of each frame is extracted from the route data.
- the trajectory data, the road condition data of each frame, the road information of each frame, the route data of each frame and the current trajectory planning parameters generate the trajectory planning result corresponding to the corresponding frame, and send the trajectory planning result corresponding to the corresponding frame to the simulator;
- the simulator generates simulated trajectory segments of corresponding frames according to the trajectory planning results.
- the cost calculation module 406 is further configured to calculate the similarity between the trajectory data and the simulated trajectory; and calculate the cost value corresponding to the simulated trajectory according to the similarity.
- the cost calculation module 406 is further configured to determine the driving trajectory data corresponding to the preset driving parameters in the simulated trajectory; determine the driving road condition data corresponding to the preset driving parameters in the road condition data; according to the driving trajectory data and The cost value corresponding to the simulated trajectory is calculated from the driving road condition data.
- the parameter adjustment module 408 is further configured to input the current trajectory planning parameters and the cost value into the non-gradient optimizer, and the non-gradient optimizer adjusts the current trajectory planning parameters within the parameter adjustment range according to the cost value , obtain the adjusted trajectory planning parameters, and send the adjusted trajectory planning parameters to the data simulation unit; use the adjusted trajectory planning parameters as the current trajectory planning parameters through the data simulation unit, and return to the data simulation unit according to the trajectory data, road conditions The step of generating the simulated trajectory of the target vehicle from the data, road information, route data and current trajectory planning parameters, until the preset conditions are met, then the parameter adjustment is stopped, and the target trajectory planning parameters are obtained.
- Each module in the above-mentioned trajectory data processing device can be implemented in whole or in part by software, hardware and combinations thereof.
- the above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.
- a computer device is provided, and the computer device may be a server, and its internal structure diagram may be as shown in FIG. 5 .
- the computer device includes a processor, memory, a communication interface, and a database connected by a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities.
- the memory of the computer device includes a non-volatile storage medium, an internal memory.
- the non-volatile storage medium stores an operating system, computer readable instructions and a database.
- the internal memory provides an environment for the execution of the operating system and computer-readable instructions in the non-volatile storage medium.
- the database of the computer device is used for storing data of a trajectory data processing method.
- the communication interface of the computer device is used to connect and communicate with an external terminal.
- the computer readable instructions when executed by a processor, implement a trajectory data processing method.
- FIG. 5 is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer equipment to which the solution of the present application is applied. Include more or fewer components than shown in the figures, or combine certain components, or have a different arrangement of components.
- a computer device comprising a memory and one or more processors, the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the one or more processors, makes the one or more processors execute the above methods to implement steps in the example.
- One or more non-volatile computer-readable storage media storing computer-readable instructions, when the computer-readable instructions are executed by one or more processors, cause the one or more processors to execute the above method embodiments. step.
- Nonvolatile memory may include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory.
- Volatile memory may include random access memory (RAM) or external cache memory.
- RAM is available in various forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Road (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Traffic Control Systems (AREA)
Abstract
一种轨迹数据处理方法,包括:获取预设时间段内目标车辆的轨迹数据、道路信息、航路数据和与目标车辆在预设距离范围内的路况数据;将轨迹数据、路况数据、道路信息和航路数据输入数据模拟单元,数据模拟单元中包括当前轨迹规划参数,根据轨迹数据、路况数据、道路信息、航路数据和当前轨迹规划参数生成目标车辆的模拟轨迹;对模拟轨迹进行代价计算,得到模拟轨迹对应的代价值;根据代价值对当前轨迹规划参数进行调节,将调节后的轨迹规划参数作为当前轨迹规划参数,返回至根据轨迹数据、路况数据、道路信息、航路数据和当前轨迹规划参数生成目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
Description
本申请涉及一种轨迹数据处理方法、装置、计算机设备和存储介质。
人工智能技术的发展,促进了自动驾驶技术的发展。在自动驾驶过程中,自动驾驶车辆会面临越来越开放、越来越复杂的交通场景。自动驾驶车辆在交通场景中的驾驶行为对自动驾驶车辆的安全起着决定性作用,而轨迹规划作为驾驶行为的核心组成部分,为了应对复杂的交通场景,会在轨迹规划器所采用的轨迹规划算法中设置大量的参数,由于参数之间的平衡关系十分复杂,导致通过调节参数来适应多样化的交通场景变得异常困难。为了使轨迹规划算法能够适应多样化的交通场景,传统方式是通过学习轨迹规划算法中的轨迹规划参数的方法,如逆强化学习的方法,来针对多样化的交通场景进行参数调节,从而在自动驾驶过程中,采用参数调节后的轨迹规划算法进行轨迹规划。
然而传统方式只适应于特定的轨迹规划算法,当利用传统方式对特定规划算法以外的轨迹规划算法进行参数调节时,会导致轨迹规划算法的参数不准确。
发明内容
根据本申请公开的各种实施例,提供一种能够提高轨迹规划算法的参数准确性的轨迹数据处理方法、装置、计算机设备和存储介质。
一种轨迹数据处理方法,包括:
获取预设时间段内目标车辆的轨迹数据、与所述目标车辆在预设距离范围内的路况数据、所述目标车辆在所述预设时间段内对应的道路信息和所述目标车辆在所述预设时间段内的航路数据;
将所述轨迹数据、所述路况数据、所述道路信息和所述航路数据输入至数据模拟单元中,所述数据模拟单元中包括当前轨迹规划参数,通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹;
对所述模拟轨迹进行代价计算,得到所述模拟轨迹对应的代价值;及
根据所述代价值对所述当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将所 述调节后的轨迹规划参数作为所述当前轨迹规划参数,返回至所述通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
一种轨迹数据处理装置,包括:
数据获取模块,用于获取预设时间段内目标车辆的轨迹数据、与所述目标车辆在预设距离范围内的路况数据、所述目标车辆在所述预设时间段内对应的道路信息和所述目标车辆在所述预设时间段内的航路数据;
数据模拟模块,用于将所述轨迹数据、所述路况数据、所述道路信息和所述航路数据输入至数据模拟单元中,所述数据模拟单元中包括当前轨迹规划参数;通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹;
代价计算模块,用于对所述模拟轨迹进行代价计算,得到所述模拟轨迹对应的代价值;及
参数调节模块,用于根据所述代价值对所述当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将所述调节后的轨迹规划参数作为所述当前轨迹规划参数,返回至所述通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
一种计算机设备,包括存储器和一个或多个处理器,所述存储器中储存有计算机可读指令,所述计算机可读指令被所述处理器执行时,使得所述一个或多个处理器执行以下步骤:
获取预设时间段内目标车辆的轨迹数据、与所述目标车辆在预设距离范围内的路况数据、所述目标车辆在所述预设时间段内对应的道路信息和所述目标车辆在所述预设时间段内的航路数据;
将所述轨迹数据、所述路况数据、所述道路信息和所述航路数据输入至数据模拟单元中,所述数据模拟单元中包括当前轨迹规划参数,通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹;
对所述模拟轨迹进行代价计算,得到所述模拟轨迹对应的代价值;及
根据所述代价值对所述当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将所述调节后的轨迹规划参数作为所述当前轨迹规划参数,返回至所述通过所述数据模拟单元根 据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
一个或多个存储有计算机可读指令的非易失性计算机可读存储介质,计算机可读指令被一个或多个处理器执行时,使得一个或多个处理器执行以下步骤:
获取预设时间段内目标车辆的轨迹数据、与所述目标车辆在预设距离范围内的路况数据、所述目标车辆在所述预设时间段内对应的道路信息和所述目标车辆在所述预设时间段内的航路数据;
将所述轨迹数据、所述路况数据、所述道路信息和所述航路数据输入至数据模拟单元中,所述数据模拟单元中包括当前轨迹规划参数,通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹;
对所述模拟轨迹进行代价计算,得到所述模拟轨迹对应的代价值;及
根据所述代价值对所述当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将所述调节后的轨迹规划参数作为所述当前轨迹规划参数,返回至所述通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
本申请的一个或多个实施例的细节在下面的附图和描述中提出。本申请的其它特征和优点将从说明书、附图以及权利要求书变得明显。
为了更清楚地说明本申请实施例中的技术方案,下面将对实施例中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本申请的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其它的附图。
图1为一个或多个实施例中轨迹数据处理方法的应用环境图。
图2为一个或多个实施例中轨迹数据处理方法的流程示意图。
图3为一个或多个实施例中对当前轨迹规划参数进行调节步骤的流程示意图。
图4为一个或多个实施例中轨迹数据处理装置的框图。
图5为一个或多个实施例中计算机设备的框图。
为了使本申请的技术方案及优点更加清楚明白,以下结合附图及实施例,对本申请进行进一步详细说明。应当理解,此处描述的具体实施例仅仅用以解释本申请,并不用于限定本申请。
本申请提供的轨迹数据处理方法,可以应用于如图1所示的应用环境中。终端102与服务器104通过网络进行通信。在需要调节轨迹规划参数时,终端102可以发送参数调节请求至服务器104,服务器104在获取到参数调节请求后,对参数调节请求进行解析,得到预设时间段内目标车辆的轨迹数据、与目标车辆在预设距离范围内的路况数据、目标车辆所在的道路信息和目标车辆在预设时间段内的航路数据。还可以是服务器104下发参数调节指令至终端102,终端102根据参数调节指令获取到预设时间段内目标车辆的轨迹数据、与目标车辆在预设距离范围内的路况数据、目标车辆所在的道路信息和目标车辆在预设时间段内的航路数据,并将获取到的数据发送至服务器104。服务器104将轨迹数据、路况数据、道路信息和航路数据输入至数据模拟单元中,数据模拟单元中包括当前轨迹规划参数,通过数据模拟单元根据轨迹数据、路况数据、道路信息、航路数据和当前轨迹规划参数生成目标车辆的模拟轨迹,从而服务器104对模拟轨迹进行代价计算,得到模拟轨迹对应的代价值,进而根据代价值对当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将调节后的轨迹规划参数作为当前轨迹规划参数,返回至通过数据模拟单元根据轨迹数据、路况数据、道路信息、航路数据和当前轨迹规划参数生成目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标规划参数。其中,终端102可以但不限于是各种个人计算机、笔记本电脑、智能手机、平板电脑和便携式可穿戴设备。服务器104可以用独立的服务器或者是多个服务器组成的服务器集群来实现。
在其中一个实施例中,如图2所示,提供了一种轨迹数据处理方法,以该方法应用于图1中的服务器为例进行说明,包括以下步骤:
步骤202,获取预设时间段内目标车辆的轨迹数据、与目标车辆在预设距离范围内的路况数据、目标车辆所在的道路信息和目标车辆在预设时间段内的航路数据。
自动驾驶车辆会面临越来越开放、越来越复杂的交通场景。自动驾驶车辆在交通场景中的驾驶行为对自动驾驶车辆的安全起着决定性作用,而轨迹规划作为驾驶行为的核心组成部分,为了应对复杂的交通场景,会在轨迹规划器所采用的轨迹规划算法中设置大量的参数。在轨迹规划过程中,可以通过位置信息、周围环境信息等,规划出一段时间内的行驶路线,并执行具体动作以改变自动驾驶车辆的行驶状态,例如停车、直行、换道、转弯等。本申请提供的轨迹数据处理方法,可以适用于任意一种轨迹规划算法来确定任意一种复杂、多样化 的交通场景的最优轨迹规划参数,例如,轨迹规划算法可以包括快速扩展随机树(rapidly-exploring random tree,简称RRT)方法、可视图法(visibility map)、概率路线图算法(probability roadmap method,简称PRM)等,交通场景可以包括换道场景、转弯场景、车道保持场景、停车场景等。
预设时间段内目标车辆的轨迹数据是指目标车辆在实际行驶过程中的轨迹数据,与目标车辆在预设距离范围内的路况数据是指与车辆在预设距离范围内的环境车辆在实际行驶过程中的轨迹数据。轨迹数据和路况数据均是已经录制的历史数据。预设时间段是用于调节轨迹规划参数所选择的时间段。目标车辆是指用于自动驾驶的任意一辆车,其轨迹数据用于调节轨迹规划参数。与目标车辆在预设距离范围内的路况数据是指目标车辆周围的环境车辆的轨迹数据。航路数据是指目标车辆需要到达的目的地数据。
为了使自动驾驶车辆在交通场景中进行准确地轨迹规划,可以预先对交通场景所采用的轨迹规划算法的参数进行调节。具体的,服务器可以获取终端发送的参数调节请求,对参数调节请求进行解析,得到预设时间段内目标车辆的轨迹数据、与目标车辆在预设距离范围内的路况数据、目标车辆所在的道路信息和目标车辆在预设时间段内的航路数据。服务器还可以下发参数调节指令至终端,终端根据参数调节指令获取到预设时间段内目标车辆的轨迹数据、与目标车辆在预设距离范围内的路况数据、目标车辆所在的道路信息和目标车辆在预设时间段内的航路数据,并将获取到的数据发送至服务器。目标车辆的轨迹数据可以包括目标车辆在每一时刻的位置、速度、加速度、加加速度等。与目标车辆在预设距离范围内的路况数据包括与目标车辆在预设距离范围内的环境车辆在每一时刻的位置、速度、加速度、加加速度等。目标车辆所在的道路信息可以包括目标车辆所在的交通车道、道路边界点、车道线信息等,道路信息可以是从高精地图中获取的。航路数据可以包括目标车辆的目的地标识、目的地的位置等。例如,预设时间段可以是40s。预设距离范围可以是以目标车辆所在位置为中心,以10m为距离半径的范围。
预设时间段内目标车辆的轨迹数据、与目标车辆在预设距离范围内的路况数据、目标车辆所在的道路信息和目标车辆在预设时间段内的航路数据可以是终端从车载计算机设备中提取的。车载计算机设备是安装在自动驾驶车辆中的,车载计算机设备可以将目标车辆的轨迹数据以及与目标车辆在已损坏距离范围内的路况数据、目标车辆所在的道路信息和目标车辆在预设时间段内的航路数据进行存储。
进一步的,服务器还可以获取终端发送的参数调节请求,对参数调节请求进行解析,得到请求参数,根据请求参数在车载计算机设备中提取预设时间段内目标车辆的轨迹数据、与目标车辆在预设距离范围内的路况数据、目标车辆所在的道路信息和目标车辆在预设时间段 内的航路数据。
步骤204,将轨迹数据、路况数据、道路信息和航路数据输入至数据模拟单元中,数据模拟单元中包括当前轨迹规划参数通过数据模拟单元根据轨迹数据、路况数据、道路信息、航路数据和当前轨迹规划参数生成目标车辆的模拟轨迹。
数据模拟单元用于规划轨迹以及根据规划轨迹来生成模拟轨迹。模拟轨迹是指在数据模拟过程中,目标车辆在数据模拟单元中运行的轨迹。
服务器在获取到轨迹数据、路况数据、道路信息和航路数据后,调用数据模拟单元,将轨迹数据、路况数据、道路信息和航路数据输入至数据模拟单元中,数据模拟单元中包括当前轨迹规划参数,从而可以通过数据模拟单元根据轨迹数据、路况数据、道路信息、航路数据以及当前轨迹规划参数来模拟目标车辆在预设时间段内运行的轨迹。
具体的,服务器可以通过数据模拟单元将轨迹数据、路况数据、道路信息和航路数据进行分帧处理,得到多帧数据。每帧数据中包括该帧对应的轨迹数据、路况数据、道路信息和航路数据。从而通过数据模拟单元根据当前轨迹规划参数生成每一帧数据对应的模拟轨迹片段。当前轨迹规划参数可以是预先指定需要调节的参数。模拟轨迹片段是指在数据模拟过程中,控制目标车辆在每一帧运行的轨迹。进一步的,通过数据模拟单元根据每一帧数据之间的时间先后顺序,依次生成每一帧数据对应的模拟轨迹片段。进而通过数据模拟单元根据多帧数据对应的模拟轨迹片段生成目标车辆的模拟轨迹。模拟轨迹可以包括每一时刻的位置、速度、加速度、加加速度等。
步骤206,对模拟轨迹进行代价计算,得到模拟轨迹对应的代价值。
为了更好的调节轨迹规划参数,服务器可以对模拟轨迹进行代价计算,代价计算的方式有多种,可以通过计算模拟轨迹与目标车辆的轨迹数据之间的相似度来进行代价计算,也可以通过比较轨迹模拟过程中,目标车辆与预设距离范围内环境车辆中各行驶参数的参数值来进行代价计算。对代价计算的方式不作限定,可以根据实际需要采取相应的代价计算方式。
在其中一个实施例中,对模拟轨迹进行代价计算,得到模拟轨迹对应的代价值包括:计算轨迹数据与模拟轨迹之间的相似度;根据相似度计算得到模拟轨迹对应的代价值。
当服务器通过计算模拟轨迹与目标车辆的轨迹数据之间的相似度来进行代价计算时,服务器可以采用模仿代价方法计算轨迹数据与模拟轨迹之间的距离,将距离作为相似度。距离越小,表明轨迹数据与模拟轨迹之间越相似。从而服务器采用模仿代价方法根据计算得到的相似度来计算模拟轨迹对应的代价值,模拟轨迹对应的代价值可以是多个代价值的加权和。具体的,服务器根据模仿代价方法中的代价函数确定需要计算的多个代价值,进而根据计算得到的相似度、代价函数以及每个代价值对应的代价权重计算得到模拟轨迹对应的代价值。 代价函数中多个代价值对应的代价权重可以作为轨迹规划参数。例如,当代价函数是效率代价、安全性代价、舒适性代价的加权和,若需要自动调节效率代价、安全性代价、舒适性代价这三项代价对应的权重,可以将这三项代价对应的权重作为轨迹规划参数。在本实施例中,通过计算轨迹数据模拟轨迹与目标车辆的轨迹数据之间的相似度来进行代价计算,能够在后续自动驾驶过程中得到更符合人类驾驶习惯的行驶轨迹,从而提高轨迹规划的灵活性。
在其中一个实施例中,对模拟轨迹进行代价计算,得到模拟轨迹对应的代价值包括:在模拟轨迹中确定与预设行驶参数对应的行驶轨迹数据;在路况数据中确定预设行驶参数对应的行驶路况数据;根据行驶轨迹数据和行驶路况数据计算模拟轨迹对应的代价值。
当服务器通过比较轨迹模拟过程中,目标车辆与预设距离范围内环境车辆中各行驶参数的参数值来进行代价计算时,可以在模拟轨迹中确定与预设行驶参数对应的行驶轨迹数据,以及在路况数据中确定预设行驶参数对应的行驶路况数据。预设行驶参数可以包括距离、速度、加速度、加加速度等。行驶轨迹数据可以包括目标车辆在每一时刻对应的各行驶参数的参数值。路况数据中包括与目标车辆在预设距离范围内的环境车辆的轨迹数据。行驶轨迹数据中包括环境车辆在每一时刻对应的各行驶参数的参数值。服务器从而采用自监督代价方法根据目标车辆与环境车辆在每一时刻的各行驶参数的参数值计算模拟轨迹对应的代价值。模拟轨迹对应的代价值可以是多个代价值的加权和。具体的,服务器根据自监督代价方法中的代价函数确定需要计算的多个代价值,进而根据目标车辆与环境车辆在每一时刻的各行驶参数的参数值、代价函数以及每个代价值对应的代价权重计算得到模拟轨迹对应的代价值。代价函数中多个代价值对应的代价权重可以作为轨迹规划参数。例如,当代价函数是距离代价、速度代价的加权和时,若需要自动调节距离代价、速度代价这两项代价对应的权重,可以将这两项代价对应的权重作为轨迹规划参数。在本实施例中,通过比较目标车辆与预设距离范围内环境车辆中各行驶参数的参数值,能够避免目标车辆与环境车辆发生碰撞,避免出现同速飙车的现象,能够提高后续自动驾驶过程中轨迹规划的准确性,保证自动驾驶的安全性。
步骤208,根据代价值对当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将调节后的轨迹规划参数作为当前轨迹规划参数,返回至通过数据模拟单元根据轨迹数据、路况数据、道路信息、航路数据和当前轨迹规划参数生成目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
数据规划单元中包括多个轨迹规划参数,轨迹规划参数可以根据数据模拟单元所采用的轨迹规划算法来决定,而轨迹规划算法中需要利用代价函数进行代价计算,因此,轨迹规划参数可以根据代价函数来确定的。不同的代价函数所对应的轨迹规划参数可以是不同的。服务器可以根据代价函数中的轨迹规划参数对数据规划单元进行设置,以便后续数据规划单元 根据轨迹规划参数进行轨迹规划。
在进行轨迹轨迹参数调节之前,服务器可以在数据规划单元的大量轨迹规划参数中预先指定轨迹规划参数和预先设置轨迹规划参数的参数调节范围。将预先指定的轨迹规划参数作为当前轨迹规划参数。服务器在进行参数调节过程中,在计算得到模拟轨迹对应的代价值后,可以根据代价值以及轨迹规划范围对数据模拟单元中的当前轨迹规划参数进行调节,将调节后的轨迹规划参数作为当前轨迹规划参数,返回至通过数据模拟单元根据轨迹数据、路况数据、道路信息、航路数据和当前轨迹规划参数生成目标车辆的模拟轨迹的步骤,进行重复参数调节,直至满足预设条件。预设条件可以是当前轨迹规划参数已经达到参数调节范围内的最优值。最优值是预先存储在服务器中的,以便服务器进行参数调节。服务器可以采用贝叶斯方法根据代价值生成调节后的轨迹规划参数,从而将调节后的轨迹规划参数作为当前轨迹规划参数,输入至数据模拟单元中。服务器在当前轨迹规划参数满足预设条件时,停止参数调节,将此时的轨迹规划参数作为目标轨迹规划参数。例如,在当前轨迹规划参数为效率代价和安全性代价这两项代价的权重时,服务器可以根据代价值在参数调节范围内对两项代价的权重进行迭代调节,当效率代价的权重和安全性代价的权重均到达最优值时,停止参数调节,将此时效率代价的权重和安全性代价的权重作为目标轨迹规划参数。
在其中一个实施例中,服务器可以将目标轨迹规划参数存储在数据模拟单元中。在自动驾驶过程中,数据模拟单元可以利用目标轨迹规划参数对自动驾驶车辆进行轨迹规划,由于目标轨迹规划参数为预先调节的、准确的轨迹规划参数,可以提高轨迹规划的准确性。
在本实施例中,将获取到的预设时间段内目标车辆的轨迹数据、与目标车辆在预设距离范围内的路况数据、目标车辆在预设时间段内对应的道路信息和目标车辆在预设时间段内的航路数据输入至数据模拟单元中,通过数据模拟单元根据轨迹数据、路况数据、道路信息、航路数据和当前轨迹规划参数生成目标车辆的模拟轨迹,从而对模拟轨迹进行代价计算,得到模拟轨迹对应的代价值,进而根据代价值对当前轨迹规划参数进行调节,将调节后的轨迹规划参数作为当前轨迹规划参数,返回至通过数据模拟单元根据轨迹数据、路况数据、道路信息、航路数据和当前轨迹规划参数生成目标车辆的模拟轨迹的步骤,进行重复参数调节,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。在轨迹规划参数的调节过程中,并不关注轨迹规划器的轨迹规划方式,能够独立于轨迹规划器进行参数调节,因此能够适用于多样化交通场景下的任意轨迹规划方法,有效提高了轨迹规划算法的参数准确性。同时,只需要当前轨迹规划参数、设置预设条件,即可实现参数的自动调节,提高了轨迹规划算法参数的调节效率。
在其中一个实施例中,通过数据模拟单元根据轨迹数据、路况数据、道路信息、航路数 据和当前轨迹规划参数生成目标车辆的模拟轨迹的步骤包括:通过数据模拟单元在轨迹数据中提取每一帧的轨迹数据、在路况数据中提取每一帧的路况数据、在道路信息中提取每一帧的道路信息以及在航路数据中提取每一帧的航路数据;根据每一帧的轨迹数据、每一帧的路况数据、每一帧的道路信息、每一帧的航路数据和当前轨迹规划参数生成每一帧的模拟轨迹片段;在得到上一帧的模拟轨迹片段后,根据下一帧的轨迹数据、下一帧的路况数据、下一帧的道路信息、下一帧的航路数据和当前轨迹规划参数生成下一帧对应的模拟轨迹片段,直至得到最后一帧的模拟轨迹片段,根据多帧的模拟轨迹片段得到目标车辆的模拟轨迹。
轨迹数据可以包括目标车辆在每一时刻的位置、速度、加速度、加加速度等。路况数据包括与目标车辆在预设距离范围内的环境车辆在每一时刻的位置、速度、加速度、加加速度等。道路信息可以包括目标车辆所在的交通车道、道路边界点、车道线信息等,道路信息可以是从高精地图中获取的。航路数据可以包括目标车辆的目的地标识、目的地的位置等。例如,预设时间段可以是40s。预设距离范围可以是以目标车辆所在位置为中心,以10m为距离半径的范围。
由于服务器发送至数据模拟单元中的轨迹数据、路况数据、道路信息和航路数据中均为预设时间段内的数据,数据模拟单元可以将轨迹数据、路况数据、道路信息和航路数据进行分帧处理,得到多帧数据。每帧数据中包括该帧对应的轨迹数据、路况数据、道路信息和航路数据。从而数据模拟单元根据每一帧数据之间的时间先后顺序,依次提取每一帧的数据,根据当前轨迹规划参数对每一帧的数据进行轨迹规划,得到轨迹规划结果。数据模拟单元根据轨迹规划结果控制目标车辆运行一段时间,直至达到下一帧对应的时间戳,得到该帧对应的模拟轨迹片段。模拟轨迹片段是指在数据模拟过程中,控制目标车辆在每一帧运行的轨迹。数据模拟单元在生成上一帧的模拟轨迹片段后,获取下一帧的数据,根据下一帧的数据生成下一帧对应的模拟轨迹片段,直至得到最后一帧的模拟轨迹片段,进而数据模拟单元根据多帧的模拟轨迹片段得到目标车辆的模拟轨迹。下一帧的数据包括下一帧的轨迹数据、下一帧的路况数据、下一帧的道路信息和下一帧的航路数据。模拟轨迹可以包括每一时刻的位置、速度、加速度、加加速度等。
在其中一个实施例中,针对任意一种交通场景,数据模拟单元的模拟帧数可以是固定的,因此生成模拟片段的循环次数也是固定的。
在本实施例中,根据每一帧的轨迹数据、每一帧的路况数据、每一帧的道路信息、每一帧的航路数据和当前轨迹规划参数生成每一帧的模拟轨迹片段,在得到上一帧的模拟轨迹片段后,根据下一帧的轨迹数据、下一帧的路况数据、下一帧的道路信息、下一帧的航路数据和当前轨迹规划参数生成下一帧对应的模拟轨迹片段,直至得到最后一帧的模拟轨迹片段, 根据多帧的模拟轨迹片段得到目标车辆的模拟轨迹。有利于后续根据模拟轨迹计算代价值,以实现对轨迹规划参数进行调节。
在其中一个实施例中,数据模拟单元包括模拟器和轨迹规划器,上述方法还包括:将轨迹数据和路况数据发送至数据模拟单元中的模拟器,以及将道路信息和航路数据发送至数据模拟单元中的轨迹规划器,轨迹规划器包括当前轨迹规划参数;通过模拟器在轨迹数据中提取每一帧的轨迹数据和在路况数据中提取每一帧的路况数据,将提取的每一帧的轨迹数据和每一帧的路况数据发送至轨迹规划器;通过轨迹规划器在道路信息中提取每一帧的道路信息和在航路数据中提取每一帧的航路数据,根据每一帧的轨迹数据、每一帧的路况数据、每一帧的道路信息、每一帧的航路数据和当前轨迹规划参数生成相应帧对应的轨迹规划结果,将相应帧对应的轨迹规划结果发送至模拟器;通过模拟器根据轨迹规划结果生成相应帧的模拟轨迹片段。
数据模拟单元包括模拟器和轨迹规划器,模拟器用于根据轨迹规划器发送的轨迹规划结果生成模拟轨迹。轨迹规划器用于根据模拟器发送的每一帧的轨迹数据、每一帧的路况数据、每一帧的道路信息、每一帧的航路数据和当前轨迹规划参数生成相应帧对应的轨迹规划结果。
通过模拟器按照时间先后顺序在轨迹数据中依次提取每一帧的轨迹数据和在路况数据中提取每一帧的路况数据,并将提取到的数据发送至轨迹规划器。轨迹规划器根据接收到的数据在道路信息中提取相应帧的道路信息和在航路数据中提取相应帧的航路数据,可以将提取的相同帧的轨迹数据、路况数据、道路信息和航路数据作为一帧数据。轨迹规划器可以根据该帧数据以及和当前轨迹规划参数生成相应帧对应的轨迹规划结果,将相应帧对应的轨迹规划结果发送至模拟器。模拟器从而控制目标车辆按照轨迹规划结果行驶,直至达到下一帧对应的时间戳,模拟器在轨迹数据中提取下一帧的的轨迹数据和在路况数据中提取每一帧的路况数据,将提取的下一帧的轨迹数据和下一帧的路况数据发送至轨迹规划器。轨迹规划器根据下一帧的数据以及当前轨迹规划参数生成下一帧对应的轨迹规划结果,将下一帧对应的轨迹规划结果发送至模拟器,模拟器从而控制目标车辆按照轨迹规划结果行驶,直至达到下一帧对应的时间戳。模拟器和轨迹规划器重复上述生成模拟轨迹片段的步骤,直至得到最后一帧的模拟轨迹片段。最终模拟器根据多帧的模拟轨迹片段输出目标车辆的模拟轨迹。
在本实施例中,通过模拟器和轨迹规划器之间的循环,生成多帧的模拟轨迹片段,最终生成目标车辆的模拟轨迹,在当前轨迹规划参数的调节过程中,并不关注轨迹规划器的具体轨迹规划方式,只需要轨迹规划器输出的轨迹规划结果,即可生成目标车辆的模拟轨迹。实现独立于轨迹规划器进行参数调节,能够适用于多样化交通场景下的任意轨迹规划方法,有效提高了轨迹规划算法的参数准确性。
在其中一个实施例中,如图3所示,对当前轨迹规划参数进行调节的步骤包括:
步骤302,将当前轨迹规划参数和代价值输入至非梯度优化器中,通过非梯度优化器根据代价值在参数调节范围内对当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将调节后的轨迹规划参数发送至数据模拟单元。
步骤304,通过数据模拟单元将调节后的轨迹规划参数作为当前轨迹规划参数,返回至通过数据模拟单元根据轨迹数据、路况数据、道路信息、航路数据和当前轨迹规划参数生成目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
当前轨迹规划参数可以是服务器预先设置的需要进行调节的参数。服务器在计算得到模拟轨迹的代价值后,获取数据模拟单元的当前轨迹规划参数。数据模拟单元中包括模拟器和轨迹规划器,当前轨迹规划参数为轨迹规划器的参数。服务器可以调用非梯度优化器,将当前轨迹规划参数和代价值输入至非梯度优化器中,通过非梯度优化器采用贝叶斯方法在参数调节范围内对当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,非梯度优化器将调节后的轨迹规划参数输入至数据模拟单元中,数据模拟单元将调节后的轨迹规划参数作为当前轨迹规划参数,返回至通过数据模拟单元根据轨迹数据、路况数据、道路信息、航路数据和当前轨迹规划参数生成目标车辆的模拟轨迹的步骤,重复进行参数调节的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。预设条件可以是当前轨迹规划参数已经达到参数调节范围内的最优值。
在本实施例中,通过非梯度优化器在参数调节范围内对当前轨迹规划参数进行调节,能够快速计算得到一个新的轨迹规划参数,从而再次根据新的轨迹规划参数生成模拟轨迹,对模拟轨迹进行代价计算,得到模拟轨迹对应的代价值,通过非梯度优化器在参数调节范围内对当前轨迹规划参数进行调节,通过迭代循环,得到目标轨迹规划参数。
在其中一个实施例中,如图4所示,提供了一种轨迹数据处理装置,包括:数据获取模块402、数据模拟模块404、代价计算模块406和参数调节模块408,其中:
数据获取模402,用于获取预设时间段内目标车辆的轨迹数据、与目标车辆在预设距离范围内的路况数据、目标车辆在预设时间段内对应的道路信息和目标车辆在预设时间段内的航路数据。
数据模拟模块404,用于将轨迹数据、路况数据、道路信息和航路数据输入至数据模拟单元中,数据模拟单元中包括当前轨迹规划参数,通过数据模拟单元根据轨迹数据、路况数据、道路信息、航路数据和当前轨迹规划参数生成目标车辆的模拟轨迹。
代价计算模块406,用于对模拟轨迹进行代价计算,得到模拟轨迹对应的代价值。
参数调节模块408,用于根据代价值对当前轨迹规划参数进行调节,得到调节后的轨迹 规划参数,将调节后的轨迹规划参数作为当前轨迹规划参数,返回至通过数据模拟单元根据轨迹数据、路况数据、道路信息、航路数据和当前轨迹规划参数生成目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
在其中一个实施例中,数据模拟模块404还用于通过数据模拟单元在轨迹数据中提取每一帧的轨迹数据、在路况数据中提取每一帧的路况数据、在道路信息中提取每一帧的道路信息以及在航路数据中提取每一帧的航路数据;根据每一帧的轨迹数据、每一帧的路况数据、每一帧的道路信息、每一帧的航路数据和当前轨迹规划参数生成每一帧的模拟轨迹片段;在得到上一帧的模拟轨迹片段后,根据下一帧的轨迹数据、下一帧的路况数据、下一帧的道路信息、下一帧的航路数据和当前轨迹规划参数生成下一帧对应的模拟轨迹片段,直至得到最后一帧的模拟轨迹片段,根据多帧的模拟轨迹片段得到目标车辆的模拟轨迹。
在其中一个实施例中,数据模拟单元包括模拟器和轨迹规划器,数据模拟模块404还用于将轨迹数据和路况数据发送至数据模拟单元中的模拟器,以及将道路信息和航路数据发送至数据模拟单元中的轨迹规划器,轨迹规划器包括当前轨迹规划参数;通过模拟器在轨迹数据中提取每一帧的轨迹数据和在路况数据中提取每一帧的路况数据,将提取的每一帧的轨迹数据和每一帧的路况数据发送至轨迹规划器;通过轨迹规划器在道路信息中提取每一帧的道路信息和在航路数据中提取每一帧的航路数据,根据每一帧的轨迹数据、每一帧的路况数据、每一帧的道路信息、每一帧的航路数据和当前轨迹规划参数生成相应帧对应的轨迹规划结果,将相应帧对应的轨迹规划结果发送至模拟器;通过模拟器根据轨迹规划结果生成相应帧的模拟轨迹片段。
在其中一个实施例中,代价计算模块406还用于计算轨迹数据与模拟轨迹之间的相似度;根据相似度计算得到模拟轨迹对应的代价值。
在其中一个实施例中,代价计算模块406还用于在模拟轨迹中确定与预设行驶参数对应的行驶轨迹数据;在路况数据中确定预设行驶参数对应的行驶路况数据;根据行驶轨迹数据和行驶路况数据计算模拟轨迹对应的代价值。
在其中一个实施例中,参数调节模块408还用于将当前轨迹规划参数和代价值输入至非梯度优化器中,通过非梯度优化器根据代价值在参数调节范围内对当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将调节后的轨迹规划参数发送至数据模拟单元;通过数据模拟单元将调节后的轨迹规划参数作为当前轨迹规划参数,返回至通过数据模拟单元根据轨迹数据、路况数据、道路信息、航路数据和当前轨迹规划参数生成目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
关于轨迹数据处理装置的具体限定可以参见上文中对于轨迹数据处理方法的限定,在此 不再赘述。上述轨迹数据处理装置中的各个模块可全部或部分通过软件、硬件及其组合来实现。上述各模块可以硬件形式内嵌于或独立于计算机设备中的处理器中,也可以以软件形式存储于计算机设备中的存储器中,以便于处理器调用执行以上各个模块对应的操作。
在其中一个实施例中,提供了一种计算机设备,该计算机设备可以是服务器,其内部结构图可以如图5所示。该计算机设备包括通过***总线连接的处理器、存储器、通信接口和数据库。其中,该计算机设备的处理器用于提供计算和控制能力。该计算机设备的存储器包括非易失性存储介质、内存储器。该非易失性存储介质存储有操作***、计算机可读指令和数据库。该内存储器为非易失性存储介质中的操作***和计算机可读指令的运行提供环境。该计算机设备的数据库用于存储一种轨迹数据处理方法的数据。该计算机设备的通信接口用于与外部的终端连接通信。该计算机可读指令被处理器执行时以实现一种轨迹数据处理方法。
本领域技术人员可以理解,图5中示出的结构,仅仅是与本申请方案相关的部分结构的框图,并不构成对本申请方案所应用于其上的计算机设备的限定,具体的计算机设备可以包括比图中所示更多或更少的部件,或者组合某些部件,或者具有不同的部件布置。
一种计算机设备,包括存储器及一个或多个处理器,存储器中储存有计算机可读指令,计算机可读指令被一个或多个处理器执行时,使得一个或多个处理器执行上述各个方法实施例中的步骤。
一个或多个存储有计算机可读指令的非易失性计算机可读存储介质,计算机可读指令被一个或多个处理器执行时,使得一个或多个处理器执行上述各个方法实施例中的步骤。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程,是可以通过计算机可读指令来指令相关的硬件来完成,上述计算机可读指令可存储于一非易失性计算机可读取存储介质中,该计算机可读指令在执行时,可包括如上述各方法的实施例的流程。其中,本申请所提供的各实施例中所使用的对存储器、存储、数据库或其它介质的任何引用,均可包括非易失性和/或易失性存储器。非易失性存储器可包括只读存储器(ROM)、可编程ROM(PROM)、电可编程ROM(EPROM)、电可擦除可编程ROM(EEPROM)或闪存。易失性存储器可包括随机存取存储器(RAM)或者外部高速缓冲存储器。作为说明而非局限,RAM以多种形式可得,诸如静态RAM(SRAM)、动态RAM(DRAM)、同步DRAM(SDRAM)、双数据率SDRAM(DDRSDRAM)、增强型SDRAM(ESDRAM)、同步链路(Synchlink)DRAM(SLDRAM)、存储器总线(Rambus)直接RAM(RDRAM)、直接存储器总线动态RAM(DRDRAM)、以及存储器总线动态RAM(RDRAM)等。
以上实施例的各技术特征可以进行任意的组合,为使描述简洁,未对上述实施例中的各个技术特征所有可能的组合都进行描述,然而,只要这些技术特征的组合不存在矛盾,都应 当认为是本说明书记载的范围。
以上所述实施例仅表达了本申请的几种实施方式,其描述较为具体和详细,但并不能因此而理解为对发明专利范围的限制。应当指出的是,对于本领域的普通技术人员来说,在不脱离本申请构思的前提下,还可以做出若干变形和改进,这些都属于本申请的保护范围。因此,本申请专利的保护范围应以所附权利要求为准。
Claims (20)
- 一种轨迹数据处理方法,包括:获取预设时间段内目标车辆的轨迹数据、与所述目标车辆在预设距离范围内的路况数据、所述目标车辆在所述预设时间段内对应的道路信息和所述目标车辆在所述预设时间段内的航路数据;将所述轨迹数据、所述路况数据、所述道路信息和所述航路数据输入至数据模拟单元中,所述数据模拟单元中包括当前轨迹规划参数,通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹;对所述模拟轨迹进行代价计算,得到所述模拟轨迹对应的代价值;及根据所述代价值对所述当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将所述调节后的轨迹规划参数作为所述当前轨迹规划参数,返回至所述通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
- 根据权利要求1所述的方法,其特征在于,所述通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹包括:通过所述数据模拟单元在所述轨迹数据中提取每一帧的轨迹数据、在所述路况数据中提取每一帧的路况数据、在所述道路信息中提取每一帧的道路信息以及在所述航路数据中提取每一帧的航路数据;根据每一帧的轨迹数据、每一帧的路况数据、每一帧的道路信息、每一帧的航路数据和所述当前轨迹规划参数生成每一帧的模拟轨迹片段;及在得到上一帧的模拟轨迹片段后,根据下一帧的轨迹数据、下一帧的路况数据、下一帧的道路信息、下一帧的航路数据和所述当前轨迹规划参数生成下一帧对应的模拟轨迹片段,直至得到最后一帧的模拟轨迹片段,根据多帧的模拟轨迹片段得到所述目标车辆的模拟轨迹。
- 根据权利要求2所述的方法,其特征在于,所述数据模拟单元包括模拟器和轨迹规划器,所述方法还包括:将所述轨迹数据和所述路况数据发送至所述数据模拟单元中的模拟器,以及将所述道 路信息和所述航路数据发送至所述数据模拟单元中的轨迹规划器,所述轨迹规划器包括所述当前轨迹规划参数;通过所述模拟器在所述轨迹数据中提取每一帧的轨迹数据和在所述路况数据中提取每一帧的路况数据,将提取的每一帧的轨迹数据和每一帧的路况数据发送至所述轨迹规划器;通过所述轨迹规划器在所述道路信息中提取每一帧的道路信息和在所述航路数据中提取每一帧的航路数据,根据每一帧的轨迹数据、每一帧的路况数据、每一帧的道路信息、每一帧的航路数据和所述当前轨迹规划参数生成相应帧对应的轨迹规划结果,将所述相应帧对应的轨迹规划结果发送至所述模拟器;及通过所述模拟器根据所述轨迹规划结果生成相应帧的模拟轨迹片段。
- 根据权利要求1所述的方法,其特征在于,所述对所述模拟轨迹进行代价计算,得到所述模拟轨迹对应的代价值包括:计算所述轨迹数据与所述模拟轨迹之间的相似度;及根据所述相似度计算得到所述模拟轨迹对应的代价值。
- 根据权利要求1所述的方法,其特征在于,所述对所述模拟轨迹进行代价计算,得到所述模拟轨迹对应的代价值包括:在所述模拟轨迹中确定与预设行驶参数对应的行驶轨迹数据;在所述路况数据中确定所述预设行驶参数对应的行驶路况数据;及根据所述行驶轨迹数据和所述行驶路况数据计算所述模拟轨迹对应的代价值。
- 根据权利要求1所述的方法,其特征在于,所述根据所述代价值对所述当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将所述调节后的轨迹规划参数作为所述当前轨迹规划参数,返回至所述通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数包括:将所述当前轨迹规划参数和所述代价值输入至非梯度优化器中,通过所述非梯度优化器根据所述代价值在参数调节范围内对所述当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将所述调节后的轨迹规划参数发送至所述数据模拟单元;及通过所述数据模拟单元将所述调节后的轨迹规划参数作为所述当前轨迹规划参数,返回至所述通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹的步骤,直至满足预设条 件,停止参数调节,得到目标轨迹规划参数。
- 一种轨迹数据处理装置,包括:数据获取模块,用于获取预设时间段内目标车辆的轨迹数据、与所述目标车辆在预设距离范围内的路况数据、所述目标车辆在所述预设时间段内对应的道路信息和所述目标车辆在所述预设时间段内的航路数据;数据模拟模块,用于将所述轨迹数据、所述路况数据、所述道路信息和所述航路数据输入至数据模拟单元中,所述数据模拟单元中包括当前轨迹规划参数,通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹;代价计算模块,用于对所述模拟轨迹进行代价计算,得到所述模拟轨迹对应的代价值;及参数调节模块,用于根据所述代价值对所述当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将所述调节后的轨迹规划参数作为所述当前轨迹规划参数,返回至所述通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
- 根据权利要求7所述的装置,其特征在于,数据模拟模块还用于通过所述数据模拟单元在所述轨迹数据中提取每一帧的轨迹数据、在所述路况数据中提取每一帧的路况数据、在所述道路信息中提取每一帧的道路信息以及在所述航路数据中提取每一帧的航路数据;根据每一帧的轨迹数据、每一帧的路况数据、每一帧的道路信息、每一帧的航路数据和所述当前轨迹规划参数生成每一帧的模拟轨迹片段;及在得到上一帧的模拟轨迹片段后,根据下一帧的轨迹数据、下一帧的路况数据、下一帧的道路信息、下一帧的航路数据和所述当前轨迹规划参数生成下一帧对应的模拟轨迹片段,直至得到最后一帧的模拟轨迹片段,根据多帧的模拟轨迹片段得到所述目标车辆的模拟轨迹。
- 根据权利要求7所述的装置,其特征在于,所述代价计算模块还用于计算所述轨迹数据与所述模拟轨迹之间的相似度;及根据所述相似度计算得到所述模拟轨迹对应的代价值。
- 根据权利要求7所述的装置,其特征在于,所述参数调节模块,还用于将所述当前轨迹规划参数和所述代价值输入至非梯度优化器中,通过所述非梯度优化器根据所述代价值在参数调节范围内对所述当前轨迹规划参数进行调节,得到调节后的轨迹规划参数, 将所述调节后的轨迹规划参数发送至所述数据模拟单元;及通过所述数据模拟单元将所述调节后的轨迹规划参数作为所述当前轨迹规划参数,返回至所述通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
- 一种计算机设备,包括存储器及一个或多个处理器,所述存储器中存储有计算机可读指令,所述计算机可读指令被所述一个或多个处理器执行时,使得所述一个或多个处理器执行以下步骤:获取预设时间段内目标车辆的轨迹数据、与所述目标车辆在预设距离范围内的路况数据、所述目标车辆在所述预设时间段内对应的道路信息和所述目标车辆在所述预设时间段内的航路数据;将所述轨迹数据、所述路况数据、所述道路信息和所述航路数据输入至数据模拟单元中,所述数据模拟单元中包括当前轨迹规划参数,通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹;对所述模拟轨迹进行代价计算,得到所述模拟轨迹对应的代价值;及根据所述代价值对所述当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将所述调节后的轨迹规划参数作为所述当前轨迹规划参数,返回至所述通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
- 根据权利要求11所述的计算机设备,其特征在于,所述处理器执行所述计算机可读指令时还执行以下步骤:通过所述数据模拟单元在所述轨迹数据中提取每一帧的轨迹数据、在所述路况数据中提取每一帧的路况数据、在所述道路信息中提取每一帧的道路信息以及在所述航路数据中提取每一帧的航路数据;根据每一帧的轨迹数据、每一帧的路况数据、每一帧的道路信息、每一帧的航路数据和所述当前轨迹规划参数生成每一帧的模拟轨迹片段;及在得到上一帧的模拟轨迹片段后,根据下一帧的轨迹数据、下一帧的路况数据、下一帧的道路信息、下一帧的航路数据和所述当前轨迹规划参数生成下一帧对应的模拟轨迹片段,直至得到最后一帧的模拟轨迹片段,根据多帧的模拟轨迹片段得到所述目标车辆的模拟轨迹。
- 根据权利要求12所述的计算机设备,其特征在于,所述数据模拟单元包括模拟器和轨迹规划器,所述处理器执行所述计算机可读指令时还执行以下步骤:将所述轨迹数据和所述路况数据发送至所述数据模拟单元中的模拟器,以及将所述道路信息和所述航路数据发送至所述数据模拟单元中的轨迹规划器,所述轨迹规划器包括所述当前轨迹规划参数;通过所述模拟器在所述轨迹数据中提取每一帧的轨迹数据和在所述路况数据中提取每一帧的路况数据,将提取的每一帧的轨迹数据和每一帧的路况数据发送至所述轨迹规划器;通过所述轨迹规划器在所述道路信息中提取每一帧的道路信息和在所述航路数据中提取每一帧的航路数据,根据每一帧的轨迹数据、每一帧的路况数据、每一帧的道路信息、每一帧的航路数据和所述当前轨迹规划参数生成相应帧对应的轨迹规划结果,将所述相应帧对应的轨迹规划结果发送至所述模拟器;及通过所述模拟器根据所述轨迹规划结果生成相应帧的模拟轨迹片段。
- 根据权利要求11所述的计算机设备,其特征在于,所述处理器执行所述计算机可读指令时还执行以下步骤:计算所述轨迹数据与所述模拟轨迹之间的相似度;及根据所述相似度计算得到所述模拟轨迹对应的代价值。
- 根据权利要求11所述的计算机设备,其特征在于,所述处理器执行所述计算机可读指令时还执行以下步骤:将所述当前轨迹规划参数和所述代价值输入至非梯度优化器中,通过所述非梯度优化器根据所述代价值在参数调节范围内对所述当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将所述调节后的轨迹规划参数发送至所述数据模拟单元;及通过所述数据模拟单元将所述调节后的轨迹规划参数作为所述当前轨迹规划参数,返回至所述通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
- 一个或多个存储有计算机可读指令的非易失性计算机可读存储介质,所述计算机可读指令被一个或多个处理器执行时,使得所述一个或多个处理器执行以下步骤:获取预设时间段内目标车辆的轨迹数据、与所述目标车辆在预设距离范围内的路况数据、所述目标车辆在所述预设时间段内对应的道路信息和所述目标车辆在所述预设时间段内的航路数据;将所述轨迹数据、所述路况数据、所述道路信息和所述航路数据输入至数据模拟单元中,所述数据模拟单元中包括当前轨迹规划参数,通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标 车辆的模拟轨迹;对所述模拟轨迹进行代价计算,得到所述模拟轨迹对应的代价值;及根据所述代价值对所述当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将所述调节后的轨迹规划参数作为所述当前轨迹规划参数,返回至所述通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
- 根据权利要求16所述的存储介质,其特征在于,所述计算机可读指令被所述处理器执行时还执行以下步骤:通过所述数据模拟单元在所述轨迹数据中提取每一帧的轨迹数据、在所述路况数据中提取每一帧的路况数据、在所述道路信息中提取每一帧的道路信息以及在所述航路数据中提取每一帧的航路数据;根据每一帧的轨迹数据、每一帧的路况数据、每一帧的道路信息、每一帧的航路数据和所述当前轨迹规划参数生成每一帧的模拟轨迹片段;及在得到上一帧的模拟轨迹片段后,根据下一帧的轨迹数据、下一帧的路况数据、下一帧的道路信息、下一帧的航路数据和所述当前轨迹规划参数生成下一帧对应的模拟轨迹片段,直至得到最后一帧的模拟轨迹片段,根据多帧的模拟轨迹片段得到所述目标车辆的模拟轨迹。
- 根据权利要求17所述的存储介质,其特征在于,所述数据模拟单元包括模拟器和轨迹规划器,所述计算机可读指令被所述处理器执行时还执行以下步骤:将所述轨迹数据和所述路况数据发送至所述数据模拟单元中的模拟器,以及将所述道路信息和所述航路数据发送至所述数据模拟单元中的轨迹规划器,所述轨迹规划器包括所述当前轨迹规划参数;通过所述模拟器在所述轨迹数据中提取每一帧的轨迹数据和在所述路况数据中提取每一帧的路况数据,将提取的每一帧的轨迹数据和每一帧的路况数据发送至所述轨迹规划器;通过所述轨迹规划器在所述道路信息中提取每一帧的道路信息和在所述航路数据中提取每一帧的航路数据,根据每一帧的轨迹数据、每一帧的路况数据、每一帧的道路信息、每一帧的航路数据和所述当前轨迹规划参数生成相应帧对应的轨迹规划结果,将所述相应帧对应的轨迹规划结果发送至所述模拟器;及通过所述模拟器根据所述轨迹规划结果生成相应帧的模拟轨迹片段。
- 根据权利要求16所述的存储介质,其特征在于,所述计算机可读指令被所述处理器执行时还执行以下步骤:计算所述轨迹数据与所述模拟轨迹之间的相似度;及根据所述相似度计算得到所述模拟轨迹对应的代价值。
- 根据权利要求16所述的存储介质,其特征在于,所述计算机可读指令被所述处理器执行时还执行以下步骤:将所述当前轨迹规划参数和所述代价值输入至非梯度优化器中,通过所述非梯度优化器根据所述代价值在参数调节范围内对所述当前轨迹规划参数进行调节,得到调节后的轨迹规划参数,将所述调节后的轨迹规划参数发送至所述数据模拟单元;及通过所述数据模拟单元将所述调节后的轨迹规划参数作为所述当前轨迹规划参数,返回至所述通过所述数据模拟单元根据所述轨迹数据、所述路况数据、所述道路信息、所述航路数据和所述当前轨迹规划参数生成所述目标车辆的模拟轨迹的步骤,直至满足预设条件,停止参数调节,得到目标轨迹规划参数。
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/CN2020/138520 WO2022133775A1 (zh) | 2020-12-23 | 2020-12-23 | 轨迹数据处理方法、装置、计算机设备和存储介质 |
CN202080103187.1A CN116097193A (zh) | 2020-12-23 | 2020-12-23 | 轨迹数据处理方法、装置、计算机设备和存储介质 |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/CN2020/138520 WO2022133775A1 (zh) | 2020-12-23 | 2020-12-23 | 轨迹数据处理方法、装置、计算机设备和存储介质 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022133775A1 true WO2022133775A1 (zh) | 2022-06-30 |
Family
ID=82158542
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2020/138520 WO2022133775A1 (zh) | 2020-12-23 | 2020-12-23 | 轨迹数据处理方法、装置、计算机设备和存储介质 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN116097193A (zh) |
WO (1) | WO2022133775A1 (zh) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116069041A (zh) * | 2023-03-20 | 2023-05-05 | 北京集度科技有限公司 | 轨迹规划方法、装置、车辆及介质 |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106926844A (zh) * | 2017-03-27 | 2017-07-07 | 西南交通大学 | 一种基于实时环境信息的动态自动驾驶换道轨迹规划方法 |
US20190050000A1 (en) * | 2017-08-08 | 2019-02-14 | Skydio, Inc. | Image space motion planning of an autonomous vehicle |
CN109885891A (zh) * | 2019-01-24 | 2019-06-14 | 中国科学院合肥物质科学研究院 | 一种智能车gpu并行加速轨迹规划方法 |
CN112068586A (zh) * | 2020-08-04 | 2020-12-11 | 西安交通大学 | 一种空间时间联合优化的四旋翼无人飞行器轨迹规划方法 |
-
2020
- 2020-12-23 WO PCT/CN2020/138520 patent/WO2022133775A1/zh unknown
- 2020-12-23 CN CN202080103187.1A patent/CN116097193A/zh active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106926844A (zh) * | 2017-03-27 | 2017-07-07 | 西南交通大学 | 一种基于实时环境信息的动态自动驾驶换道轨迹规划方法 |
US20190050000A1 (en) * | 2017-08-08 | 2019-02-14 | Skydio, Inc. | Image space motion planning of an autonomous vehicle |
CN109885891A (zh) * | 2019-01-24 | 2019-06-14 | 中国科学院合肥物质科学研究院 | 一种智能车gpu并行加速轨迹规划方法 |
CN112068586A (zh) * | 2020-08-04 | 2020-12-11 | 西安交通大学 | 一种空间时间联合优化的四旋翼无人飞行器轨迹规划方法 |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116069041A (zh) * | 2023-03-20 | 2023-05-05 | 北京集度科技有限公司 | 轨迹规划方法、装置、车辆及介质 |
Also Published As
Publication number | Publication date |
---|---|
CN116097193A (zh) | 2023-05-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11427215B2 (en) | Systems and methods for generating a task offloading strategy for a vehicular edge-computing environment | |
CN113032285B (zh) | 一种高精地图测试方法、装置、电子设备及存储介质 | |
US20220363259A1 (en) | Method for generating lane changing decision-making model, method for lane changing decision-making of unmanned vehicle and electronic device | |
CN110197027B (zh) | 一种自动驾驶测试方法、装置、智能设备和服务器 | |
CN112286206B (zh) | 自动驾驶的模拟方法、***、设备、可读存储介质及平台 | |
CN110377025A (zh) | 用于自动驾驶车辆的传感器聚合框架 | |
US20230350399A1 (en) | Operational testing of autonomous vehicles | |
JP2022547611A (ja) | 道路シーンにおける多様な長期将来軌道のシミュレーション | |
KR102139172B1 (ko) | 가상환경에서의 자율주행차량 시뮬레이션 방법 | |
CN113515463A (zh) | 自动化测试方法、装置、计算机设备和存储介质 | |
EP3710993B1 (en) | Image segmentation using neural networks | |
CN111923919B (zh) | 车辆控制方法、装置、计算机设备和存储介质 | |
EP3847619A1 (en) | Unsupervised depth prediction neural networks | |
CN113561994A (zh) | 轨迹规划方法、装置、存储介质及电子设备 | |
WO2022133775A1 (zh) | 轨迹数据处理方法、装置、计算机设备和存储介质 | |
CN112382165A (zh) | 驾驶策略生成方法、装置、介质、设备及仿真*** | |
CN113192175A (zh) | 模型训练方法、装置、计算机设备和可读存储介质 | |
CN112815948A (zh) | 识别偏航模式的方法、装置、计算机设备和存储介质 | |
CN116403174A (zh) | 一种端到端自动驾驶方法、***、仿真***及存储介质 | |
CN112507857B (zh) | 一种车道线更新方法、装置、设备及存储介质 | |
CN115782921A (zh) | 用于自动驾驶的环境感知方法、装置及电子设备 | |
CN115454082A (zh) | 车辆避障方法及***、计算机可读存储介质和电子设备 | |
CN111381575A (zh) | 自动测试方法、装置、服务器、电子设备及存储介质 | |
US20220108172A1 (en) | Generating a simplified model for xil systems | |
CN115576317A (zh) | 一种基于神经网络的多预瞄点路径跟踪控制方法及*** |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 20966355 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 28/11/2023) |