A processing method for planning trajectory query based on planning branch time

Abstract

This invention relates to a method for querying planned trajectories based on planning branch times. The method includes: a planning module receiving a set of multi-target predicted trajectories sent by a prediction module at the start time t of the current work cycle; constructing one or more multi-target motion scenarios for future time periods based on the set of multi-target predicted trajectories; planning the vehicle trajectory for the future time periods based on all multi-target motion scenarios to generate a corresponding set of planned trajectories; performing planning trajectory branch time confirmation processing based on the set of planned trajectories to generate a corresponding planning branch time sequence; and performing planning trajectory branch time confirmation processing at each planning branch time t of the current work cycle. y The system responds to the planning trajectory query command sent by the control module and sends back the real-time planned trajectory that matches the real-time multi-target trajectory set in the planning trajectory query command. This invention solves the problem that conventional planning modules cannot output planning trajectories multiple times within a single work cycle.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a processing method for performing planning trajectory query based on planning branch time. Background Art

[0002] The planning control system of the vehicle's automatic driving is usually composed of a prediction module, a planning module and a control module; the prediction module predicts multiple possible trajectories of the target based on the target (obstacle) posture information collected by the perception sensor, and sends multiple predicted trajectories to the planning module; the planning module plans multiple corresponding planning trajectories based on the received multiple predicted trajectories, and selects the most likely trajectory from them and sends it to the control module; the control module uses the received planned trajectory as a reference trajectory and tracks it, and in the process of trajectory tracking, the actual motion trajectory is made as close to the reference trajectory as possible through driving control, that is, controlling the real-time motion state of the vehicle (such as steering wheel control, throttle / brake control, etc.). From the above description, it is not difficult to see that the conventional planning module will only output a planned trajectory to the control module once in a working cycle, and this planned trajectory is the most likely one among all planned trajectories.

[0003] In practical applications, we often encounter situations where the target trajectory suddenly changes. The so-called target trajectory sudden change means that the actual motion trajectory of the target does not match the target predicted trajectory corresponding to the current reference trajectory. In this case, the current reference trajectory should be invalidated and the planned trajectory matching the actual motion trajectory of the target should be used as the new reference trajectory. However, the processing method of the single-cycle single-planning output of the conventional planning module cannot update the planned trajectory in time according to the actual motion trajectory of the target when the target trajectory suddenly changes. This undoubtedly brings great safety risks to the driving control of the vehicle. Summary of the invention

[0004] The purpose of the present invention is to provide a processing method, electronic device and computer-readable storage medium for querying a planned trajectory based on a planned branch time in view of the defects of the prior art. Through the present invention, a planned trajectory reverse query mechanism is established between the planning module and the control module, which is initiated by the control module and responded to by the planning module, so as to solve the problem that the conventional planning module cannot output the planned trajectory multiple times within one working cycle; in addition, the planned branch time rather than the predicted branch time is used as the response time point of the planned trajectory reverse query mechanism between the planning module and the control module, so as to improve the accuracy of the planned trajectory reverse query.

[0005] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present invention provides a processing method for querying a planned trajectory based on a planned branch time, the method comprising:

[0006] The planning module receives the corresponding future time period [t,t end ] is a set of multi-target predicted trajectories; t≤t end ;

[0007] Construct the future time period [t,t end ] one or more multi-target motion scenes;

[0008] According to all the multi-target motion scenes, the future time period [t,t end ] to plan the ego-vehicle trajectory and generate the corresponding planning trajectory set;

[0009] According to the planned trajectory set, the planned trajectory branch time confirmation processing is performed to generate a corresponding planned branch time sequence; the planned branch time sequence includes one or more planned branch time t y ;t <t y <t end , y is the branch time index, 1≤y;

[0010] In each of the planning branch times t in the current working cycle y The control module responds to the planned trajectory query instruction sent by the control module, and sends back to the control module a real-time planned trajectory that matches the real-time multi-target trajectory set of the planned trajectory query instruction.

[0011] Preferably, the multi-target prediction trajectory set includes multiple single-target prediction trajectory sets A i ; i is the single target index, 1≤i≤n, n is the total number of targets;

[0012] The single target prediction trajectory set A i including one or more first predicted trajectories a i,j The first predicted trajectory a i,j The trajectory period is [t,t end ]; j is the predicted trajectory index, 1≤j≤m i , m i is the total number of predicted trajectories of single target i;

[0013] The planned trajectory set includes multiple planned trajectories b x ; The planned trajectory b x The trajectory period is [t,t end ]; x is the planned trajectory index, 1≤x≤m, m is the total number of planned trajectories, m=m i＝1 ×m i＝2 …×m i＝n .

[0014] Preferably, the future time period [t,t end ] one or more multi-target motion scenes, specifically including:

[0015] From each of the single target prediction trajectory sets A of the multi-target prediction trajectory set i In the above example, select any one of the first predicted trajectories a i,j Combine to generate a corresponding set of prediction trajectory groups; the total number of prediction trajectories of each prediction trajectory group is the target total number n, and the trajectory sets of any two prediction trajectory groups are not repeated; the total number of prediction trajectory groups is m i＝1 ×m i＝2 …×m i＝n ;

[0016] Each predicted trajectory group is regarded as a future period [t,t end ] of the multi-target motion scene, thereby obtaining one or more of the multi-target motion scenes.

[0017] Preferably, the future time period [t,t end ] to plan the self-vehicle trajectory and generate the corresponding planning trajectory set, including:

[0018] All the first predicted trajectories a in each of the multi-target motion scenes i,j is the future period [t,t end ], and predict the future time period [t,t based on the known starting position and target position of the ego vehicle. end ] to plan the vehicle motion trajectory and generate the corresponding planned trajectory b x ; and all the planned trajectories b are obtained x Form the corresponding planning trajectory set.

[0019] Preferably, performing planning trajectory branch time confirmation processing according to the planning trajectory set to generate a corresponding planning branch time sequence specifically includes:

[0020] For any two of the planned trajectories b in the planned trajectory set x The branch time is identified to obtain the corresponding first branch time;

[0021] When all the first branch times are not 0, time deduplication processing is performed on the obtained multiple first branch times; and each of the first branch times remaining after the deduplication processing is used as the planned branch time t y , and obtain all the planning branch times t y The corresponding planning branch time series are arranged in chronological order.

[0022] Furthermore, for any two of the planned trajectories b in the planned trajectory set x The branch time is identified to obtain the corresponding first branch time, specifically including:

[0023] Step 61: select any two planned trajectories b from the set of planned trajectories. x As the corresponding first and second planned trajectories; the trajectory period of the first and second planned trajectories is the future period [t,t end ];

[0024] Step 62, set the start time t a is the future period [t,t end ] start time t, set end time t b is the future period [t,t end ] end time t end ;

[0025] Step 63, determining whether the first and second planned trajectories are at the starting time t a If the track points match, go to step 64; if not, set the first branch time to 0 and go to step 68;

[0026] Step 64, determining whether the first and second planned trajectories are at the end time t b If not, go to step 65; if matched, set the first branch time to the end time t b and go to step 68;

[0027] Step 65, from the starting time t a To the end time t b The central time of the future period is taken as the central time t m ;

[0028] Step 66: Determine whether the first and second planned trajectories are within the central time t m Whether the trajectory points match; if they match, the starting time t a Let the central time t m , and maintain the end time t b unchanged; if not matched, keep the starting time t a unchanged, and the end time t b Let the central time t m ;

[0029] Step 67, determine from the starting time t a To the end time t bWhether the length of the future time period is greater than the preset track point time interval △t; if it is greater than the track point time interval △t, go to step 65; if it is less than or equal to the track point time interval △t, set the first branch time to the latest center time t m and go to step 68;

[0030] Step 68: output the first branch time as a recognition result.

[0031] Preferably, each of the planned branch times t in the current working cycle y Responding to the planned trajectory query instruction sent by the control module, and sending back to the control module a real-time planned trajectory that matches the real-time multi-target trajectory set of the planned trajectory query instruction, specifically includes:

[0032] In each of the planning branch times t in the current working cycle y Receive the planned trajectory query instruction sent by the control module; the planned trajectory query instruction includes the real-time multi-target trajectory set; the real-time multi-target trajectory set includes multiple single-target real-time trajectories c i ; Each of the single target real-time trajectory c i The trajectory period is [t s ,t y ], t <t s <t y ;

[0033] Confirming the multi-target motion scene that matches the real-time multi-target trajectory set and recording it as a corresponding matching motion scene;

[0034] The planned trajectory b corresponding to the matching motion scene x As matching planning trajectory;

[0035] The planning branch time t is intercepted from the matching planning trajectory y The subsequent trajectory segments are used as the corresponding real-time planning trajectories;

[0036] The real-time planning trajectory is sent back to the control module.

[0037] Further, the confirming the multi-target motion scene that matches the real-time multi-target trajectory set and recording it as a corresponding matching motion scene specifically includes:

[0038] In the multi-target predicted trajectory set, the trajectory period [t s ,t y ] and the real-time trajectory of each single target c i The first predicted trajectory a i,jRecorded as the corresponding first trajectory; and all the first trajectories obtained form a corresponding first trajectory group; and the predicted trajectory group matching the first trajectory group is recorded as the corresponding first predicted trajectory group; and the multi-target motion scene corresponding to the first predicted trajectory group is used as the matching motion scene.

[0039] A second aspect of an embodiment of the present invention provides an electronic device, including: a memory, a processor, and a transceiver;

[0040] The processor is used to be coupled to the memory, read and execute instructions in the memory, so as to implement the method steps described in the first aspect above;

[0041] The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

[0042] A third aspect of an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions. When the computer instructions are executed by a computer, the computer executes the instructions of the method described in the first aspect above.

[0043] The embodiment of the present invention provides a processing method, electronic device and computer-readable storage medium for querying a planned trajectory based on a planned branch time, so that the planning module plans the driving trajectory of the vehicle in various possible motion scenarios in the future period based on the multi-target predicted trajectory set sent by the prediction module at the start time of the working cycle to obtain the corresponding planned trajectory set; and compares each planned trajectory in the planned trajectory set to obtain multiple planned trajectory branch time points in the future period; and uses each trajectory branch time point in the same cycle as the planned trajectory query instruction response time point of the control module, and sends the planned trajectory that best matches the actual motion trajectory of the current target back to the control module as the latest reference trajectory. Through the present invention, a planned trajectory reverse query mechanism initiated by the control module and responded by the planning module is established between the planning module and the control module, which solves the problem that the conventional planning module cannot output the planned trajectory multiple times within a working cycle; in addition, the response time point of the planned trajectory reverse query mechanism is based on the planned branch time instead of the predicted branch time between the planning module and the control module, which improves the accuracy of the planned trajectory reverse query. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic diagram of a processing method for querying a planned trajectory based on planned branch time provided in the first embodiment of the present invention;

[0045] Figure 2 A schematic diagram of the structure of an electronic device provided in Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] Embodiment 1 of the present invention provides a processing method for querying a planned trajectory based on a planned branch time, such as Figure 1 As shown in the schematic diagram of a processing method for querying a planned trajectory based on a planned branch time provided in the first embodiment of the present invention, the method mainly includes the following steps:

[0048] Embodiment 1 of the present invention provides a processing method for querying a planned trajectory based on a planned branch time, such as Figure 1 As shown in the schematic diagram of a processing method for querying a planned trajectory based on a planned branch time provided in the first embodiment of the present invention, the method mainly includes the following steps:

[0049] Step 1: The planning module receives the corresponding future time period [t,t end ]’s multi-target prediction trajectory set;

[0050] Where t≤t end ; The multi-target prediction trajectory set includes multiple single-target prediction trajectory sets A i ; i is the single target index, 1≤i≤n, n is the total number of targets; single target prediction trajectory set A i including one or more first predicted trajectories a i,j ; The first predicted trajectory a i,j The trajectory period is [t,t end ]; j is the predicted trajectory index, 1≤j≤m i , m i is the total number of predicted trajectories of single target i.

[0051] Here, the multi-target prediction trajectory set includes multiple single-target prediction trajectory sets A i ; Each single target prediction trajectory set A i Corresponding to an obstacle target i in the traffic environment where the vehicle is located; each single target prediction trajectory set A i is the corresponding target i in the future period [t,t end ] is a set of multiple possible motion trajectories, each first predicted trajectory a in the set i,j Corresponding to a specific possible motion trajectory; it should be noted that the first predicted trajectory a of the same target i,jIn fact, the trajectories in the initial stage are the same. The reason why the subsequent trajectories change is that they may be caused by various movements that may occur at some point in the future.

[0052] For example, suppose there are two targets around the vehicle, namely vehicle 1 and vehicle 2, and vehicle 1 is in the future period [t,t end ]There are two possible motion trajectories: continuous straight driving and driving straight for a while and then turning left. In the future period [t,t end ] There is only one possible motion trajectory, which is continuous straight movement; then, the multi-target prediction trajectory set should include two single-target prediction trajectory sets, A1 and A2; among them, the single-target prediction trajectory set A1 should include two first prediction trajectories, a 1,1 、a 1,2 , the corresponding total number of predicted trajectories m1 = 2, a 1,1 、a 1,2 The trajectory of vehicle 1 before turning left is the same; the single target predicted trajectory set A2 should include a first predicted trajectory a 2,1 , the corresponding total number of predicted trajectories is m2=1.

[0053] Step 2: Construct the future time period [t,t end ] one or more multi-target motion scenes;

[0054] Specifically, step 21 includes: i In the above example, select a first prediction trajectory a i,j Combining and generating a set of corresponding prediction trajectory groups;

[0055] The total number of predicted trajectories of each predicted trajectory group is the target total number n, and the trajectory sets of any two predicted trajectory groups are not repeated; the total number of predicted trajectory groups is m i＝1 ×m i＝2 …×m i＝n ;

[0056] For example, it is known that the multi-target prediction trajectory set includes two single-target prediction trajectory sets A1 and A2, wherein the single-target prediction trajectory set A1 includes two first prediction trajectories a 1,1 、a 1,2 , the corresponding total number of predicted trajectories m1 = 2, and the single target predicted trajectory set A2 includes 1 first predicted trajectory a 2,1 , the corresponding total number of predicted trajectories m2 = 1; then, any first predicted trajectory selected from A1 and A2 of the multi-target predicted trajectory set can be combined to obtain m1×m2=2 predicted trajectory groups: Prediction trajectory group 1 (a 1,1 , a 2,1 ), predicted trajectory group 2 (a1,2 , a 2,1 );

[0057] Step 22: Take each predicted trajectory group as a future period [t,t end ] to obtain one or more multi-target motion scenes.

[0058] Here, each predicted trajectory group actually consists of a single trajectory of all targets, corresponding to the future period [t,t end ]A possible multi-target motion scenario.

[0059] For example, there are two prediction trajectory groups: prediction trajectory group 1 (a 1,1 , a 2,1 ), predicted trajectory group 2 (a 1,2 , a 2,1 ); then, it can be confirmed that in the future period [t,t end ] There are two possible multi-target motion scenarios:

[0060] According to the predicted trajectory group 1 (a 1,1 , a 2,1 ) corresponds to the multi-target motion scene 1, in which vehicles 1 and 2 will start from their current positions and continue to move forward; 1,2 , a 2,1 ) corresponds to a multi-target motion scene 2, in which vehicle 1 will start from its current position and go straight for a while before turning left, while vehicle 2 will start from its current position and continue to go straight forward;

[0061] It is also known that a 1,1 and a 1,2 The trajectories of vehicle 1 before turning left are consistent, so the scene combination trajectories of multi-target motion scenes 1 and 2 before vehicle 1 turns left are actually the same. At this time, the time point when vehicle 1 turns left can be regarded as the predicted branch time of all current scenes (multi-target motion scenes 1 and 2). It should be noted that the number of multi-target motion scenes in the above example is 2, so there is only one predicted branch time, but in an actual complex traffic environment, the number of multi-target motion scenes is often greater than 2. Because the predicted branch time is the branch time of any two multi-target motion scenes, there will be multiple corresponding predicted branch time points.

[0062] Step 3: According to all multi-target motion scenes, the future time period [t,t end ] to plan the ego-vehicle trajectory and generate the corresponding planning trajectory set;

[0063] Among them, the planning trajectory set includes multiple planning trajectories b x ; Planning trajectory b xThe trajectory period is [t,t end ]; x is the planned trajectory index, 1≤x≤m, m is the total number of planned trajectories, m=m i＝1 ×m i＝2 …×m i＝n ;

[0064] Specifically, it includes: taking all the first predicted trajectories a in each multi-target motion scene i,j is the future period [t,t end ], and predict the future time period [t,t based on the known starting position and target position of the ego vehicle. end ] to plan the vehicle motion trajectory and generate the corresponding planning trajectory b x ; and all the planned trajectories b are obtained x Form a corresponding set of planning trajectories.

[0065] Here, for the planning module, the starting position and the target position of the vehicle are known by default, and both can be obtained from the map data based on the vehicle's positioning system; when performing trajectory planning, the trajectory planning algorithm used by the planning module of the embodiment of the present invention is similar to the trajectory planning algorithm used by the traditional planning module, both of which are based on the known starting position and target position of the vehicle, and the first predicted trajectory a of each single target in the current multi-target motion scene. i,j The reference trajectory of the ego vehicle is used to avoid the shortest path and motion state, and the corresponding planning trajectory b is obtained by combining the driving state of the ego vehicle. x , the specific implementation can refer to the traditional planning algorithm technology implementation, which will not be further described here. Here each planning trajectory b x Corresponding to a multi-target motion scene.

[0066] For example, in the future period [t,t end There are two possible multi-target motion scenarios: 1,1 , a 2,1 ) corresponds to the multi-target motion scene 1, and the predicted trajectory group 2 (a 1,2 , a 2,1 ) corresponding to the multi-target motion scene 2; then, the planning trajectory b1 corresponding to the multi-target motion scene 1 and the planning trajectory b2 corresponding to the multi-target motion scene 2 can be obtained;

[0067] Because the scene combination trajectories of multi-target motion scenes 1 and 2 before vehicle 1 makes a left turn are actually the same, the trajectories of planned trajectories b1 and b2 in the initial stage are actually the same. The reason for the planned trajectory branch is that vehicle 1 may make a left turn at some point in the future, and the corresponding time point when the planned trajectory branch occurs is also called the planned branch time; from the conventional planning algorithm, it can be seen that the trajectory point time of the planned trajectory is not necessarily aligned with the trajectory point time of the predicted trajectory, especially the time point when the planned trajectory branches is often later than the time point when the predicted trajectory branches, so the planned branch time of multi-target motion scenes 1 and 2 here is later than the predicted branch time of multi-target motion scenes 1 and 2. In addition, although there is only one planned branch time in the above example, in an actual complex traffic environment, there are often multiple predicted branch time points, that is, there are also multiple corresponding planned branch times, and the number of planned branch times should be consistent with the number of predicted branch times.

[0068] Step 4, performing planning trajectory branch time confirmation processing according to the planning trajectory set to generate a corresponding planning branch time sequence;

[0069] Among them, the planning branch time sequence includes one or more planning branch time t y ;t <t y <t end , y is the branch time index, 1≤y;

[0070] Specifically, step 41 includes: for any two planned trajectories b in the planned trajectory set, x The branch time is identified to obtain the corresponding first branch time;

[0071] Specifically, step 411, select two planned trajectories b from the planned trajectory set x As the corresponding first and second planning trajectories;

[0072] Among them, the trajectory period of the first and second planned trajectories is the future period [t,t end ];

[0073] For example, if the set of planned trajectories includes planned trajectories b1 and b2, then the first and second planned trajectories are b1 and b2 respectively, and the trajectory periods of b1 and b2 are the future period [t,t end ];

[0074] Step 412, set the start time t a is the future period [t,t end ] start time t, set end time t b is the future period [t,t end ] end time t end ;

[0075] For example, it is known that the first and second planned trajectories are b1 and b2 respectively, and the trajectory period is the future period [t,t end ], then, t a =t,t b =t end ;

[0076] Step 413, determine the first and second planned trajectories at the starting time t a If the track points match, go to step 414; if not, set the first branch time to 0 and go to step 418;

[0077] Here, the embodiment of the present invention has a default restriction on each planned trajectory of the planned trajectory set, that is, the head trajectory segment of each planned trajectory must be partially overlapped; if the trajectory points of the first and second planned trajectories at the starting time t match, it means that the first and second planned trajectories are overlapping at the starting time t, and the above restriction is met, and the next step is to go to step 414 for further processing; if the trajectory points of the first and second planned trajectories at the starting time t do not match, it means that the first and second planned trajectories are completely separated at the starting time t, and the above restriction is not met, so the first branch time is set as the trajectory comparison error code (default is 0) and directly go to step 418 to output it;

[0078] Step 414, determine whether the first and second planned trajectories are at the end time t b If not, go to step 415; if not, set the first branch time to the end time t b and go to step 418;

[0079] Here, if the first and second planning trajectories end at time t end If the trajectory points match, it means that the first and second planning trajectories are actually two completely overlapping planning trajectories. In this case, the first branch time is set to the maximum time, that is, the end time t end Then go to step 418 to output it; if the first and second planned trajectories are at the end time t end If the trajectory points do not match, it means that the first and second planned trajectories are in the period [t,t end ] If there is a trajectory branch in the data file, the next step is to go to step 415 for further processing;

[0080] Step 415, starting from the starting time t a To end time t b The central time of the future period is taken as the central time t m ;

[0081] Step 416, determine the first and second planned trajectories at the center time t mWhether the trajectory points match; if they match, the starting time t a Set as the center time t m , and keep the end time t b unchanged; if not matched, keep the start time t a unchanged, and the end time t b Set as the center time t m ;

[0082] Step 417, determine from the start time t a To end time t b Whether the length of the future time period is greater than the preset track point time interval △t; if it is greater than the track point time interval △t, go to step 415; if it is less than or equal to the track point time interval △t, set the first branch time to the latest center time t m and go to step 418;

[0083] Here, in the above steps 415-416, the embodiment of the present invention adopts a binary search algorithm to improve the efficiency of confirming the trajectory branch time. Specifically, each time the center time t of the first and second planned trajectories is used. m The trajectory point matching result is used as the binary period selection condition, and the current period [t a ,t b ] duration from [t,t end ] time length is divided into two parts successively until it is less than or equal to the time interval △t of the track points; each time the track points are matched and compared, if the current center time t m The trajectory point matching result is a match, which means that the branch point is at the current center time t m After that, the corresponding selection of the current center time t m The following half period is the new period [t a ,t b ], and based on the new period [t a ,t b The central time t of m Continue to judge the trajectory point matching; if the current center time t m If the trajectory point matching result is not matched, it means that the branch point is at the current center time t m Before, then the corresponding selection is the current center time t m The previous half period is taken as the new period [t a ,t b ], and based on the new period [t a ,t b The central time t of m Continue to perform trajectory point matching judgment;

[0084] Step 418, outputting the first branch time as a recognition result;

[0085] Here, if there are only two planned trajectories in the planned trajectory set, only one round of steps 411-418 needs to be executed to obtain a first branch time; if the number of planned trajectories in the planned trajectory set exceeds 2, then it is necessary to execute Steps 411-418 are performed, where K is the number of planned trajectories; for example, if the number of planned trajectories in the planned trajectory set is K=4, then Repeat steps 411-418 to obtain 6 first branch times;

[0086] Step 42: when all first branch times are not 0, perform time deduplication processing on the obtained multiple first branch times; and use the remaining first branch times after deduplication processing as the planned branch time t y , and all the planning branch times t are obtained y The corresponding planning branch time series are arranged in chronological order.

[0087] Here, as mentioned above, the embodiment of the present invention has a default limitation on each planned trajectory set, that is, the head trajectory segments of each planned trajectory must overlap partially; if all first branch times are not 0, it means that the current planned trajectory set meets the limitation of the embodiment of the present invention, and then the redundant repeated times in all the first branch times are eliminated by time deduplication processing. After the time deduplication is completed, the remaining non-repeating multiple first branch times are used as the corresponding planned branch times t y Thus, a planning branch time sequence is obtained; if all first branch times cannot be non-zero, it means that there are at least two completely separated planning trajectories in the current planning trajectory set, which does not comply with the definition of the embodiment of the present invention. In this case, the embodiment of the present invention will terminate the execution of subsequent steps and issue a planning error alarm.

[0088] Here, the planning module of the embodiment of the present invention plans the driving trajectory of the vehicle in various possible motion scenarios in the future time period based on the multi-target predicted trajectory set sent by the prediction module at the start time of the working cycle through the above steps 1-4, thereby obtaining a corresponding planned trajectory set; and compares each planned trajectory in the planned trajectory set to obtain a plurality of planned trajectory branch time points in the future time period to form a planned branch time sequence. After obtaining the planned trajectory set and the planned branch time sequence, the planning module of the embodiment of the present invention will not actively send the planned trajectory to the downstream control module like a conventional planning module, but will use the planned trajectory set as the trajectory set to be queried to wait for the planned trajectory query instruction of the response control module. It should be noted that in order to improve the synergy between the downstream control module and the planning module, the planning module of the embodiment of the present invention will also provide the control module with the latest planned branch time sequence through data sharing or instruction sending.

[0089] If the planned branch time t of the planned branch time series y More than one or only one but not t end , it means that there are multiple multi-target motion scenes in the future period; to solve the uncertainty of multiple scenes, the downstream control module needs to make a planning trajectory query based on the multi-target real-time trajectory that has occurred at each possible trajectory branch time point; when querying, the control module should obtain the real-time trajectories of all targets in the environment from the perception system of the vehicle to form a real-time multi-target trajectory set and send it to the planning module, and use the planning trajectory sent back by the planning module as the reference trajectory that best matches the current real scene to control the driving state of the vehicle; ideally, the time when the control module initiates a query each time should be related to the predicted branch time of multiple multi-target motion scenes; but in the actual implementation process, we found that if the query is initiated at the predicted branch time, the target predicted trajectory after the branch time in the real-time multi-target trajectory set is very short and the data robustness is poor, which can easily lead to misjudgment of the planning module when performing motion scene recognition, thereby causing errors in the output planning trajectory; therefore, the embodiment of the present invention specifically selects a planning branch time t later than the predicted branch time y As the query initiation time, because the planning branch time t y Later than the predicted branch time, the target predicted trajectory after the branch time in the real-time multi-target trajectory set will increase and the data robustness will be enhanced. In this way, the accuracy of the planning module in motion scene recognition and planning trajectory output will be improved. The control module is at the planning branch time t y After sending the planning trajectory query instruction carrying the multi-target trajectory set matching to the planning module, wait for the planning module to return the corresponding planning trajectory. Correspondingly, the planning module is in the planning branch time t of the planning branch time sequence y More than one or only one but not t end At each planning branch time t y The processing flow of the planning trajectory query instruction sent by the upper response control module is completed by the following step 5.

[0090] Step 5: At each planning branch time t in the current working cycle y Respond to the planned trajectory query instruction sent by the control module, and send back to the control module the real-time planned trajectory that matches the real-time multi-target trajectory set of the planned trajectory query instruction;

[0091] Specifically, step 51 includes: y Receive the planning trajectory query instruction sent by the control module;

[0092] The planned trajectory query instruction includes a real-time multi-target trajectory set; the real-time multi-target trajectory set includes multiple single-target real-time trajectories ci ; Real-time trajectory of each single target c i The trajectory period is [t s ,t y ], t <t s <t y ;

[0093] Here, if the delay δ of the perception system is considered during the implementation process, the real-time trajectory c of each single target i The trajectory period should be [t s ,t ’ y ], t ’ y =(t y -δ);

[0094] Step 52, confirming the multi-target motion scene that matches the real-time multi-target trajectory set and recording it as the corresponding matching motion scene;

[0095] Specifically, it includes: in the multi-target prediction trajectory set, the trajectory period [t s ,t y ] and the real-time trajectory of each single target c i The first predicted trajectory a i,j Recorded as the corresponding first trajectory; and all the obtained first trajectories form a corresponding first trajectory group; and the predicted trajectory group matching the first trajectory group is recorded as the corresponding first predicted trajectory group; and the multi-target motion scene corresponding to the first predicted trajectory group is taken as the matching motion scene;

[0096] For example, it is known that the multi-target prediction trajectory set includes two single-target prediction trajectory sets A1 and A2, and the single-target prediction trajectory set A1 includes two first prediction trajectories a 1,1 、a 1,2 , the single target prediction trajectory set A2 includes a first prediction trajectory a 2,1 , the corresponding two prediction trajectory groups are: prediction trajectory group 1 (a 1,1 , a 2,1 ), prediction trajectory group 2 (a 1,2 , a 2,1 ), the predicted trajectory group 1 corresponds to the multi-target motion scene 1, and the predicted trajectory group 2 corresponds to the multi-target motion scene 2; in addition, it is also known that the real-time multi-target trajectory set includes two single-target real-time trajectories c1 and c2;

[0097] Assume a 1,1 In the trajectory period [t s ,t y ] is segmented into segment 11, a 1,2 In the trajectory period [t s ,t y] is segmented into segment 12, a 2,1 In the trajectory period [t s ,t y ] is segment 21; and the segment that matches c1 is segment 12, and the segment that matches c2 is segment 21;

[0098] Then, the two first trajectories 1 and 2 obtained should be a 1,2 、a 2,1 , the first trajectory group should be (a 1,2 ,a 2,1 ), the first prediction trajectory group should be prediction trajectory group 2 (a 1,2 , a 2,1 ), the matching motion scene should be the multi-target motion scene 2 corresponding to the prediction trajectory group 2;

[0099] Step 53: The planned trajectory b corresponding to the matching motion scene x As matching planning trajectory;

[0100] For example, it is known that multi-target motion scenes 1 and 2 correspond to planned trajectories b1 and b2 respectively. If the matching motion scene is multi-target motion scene 2, the matching planned trajectory should be planned trajectory b2;

[0101] Step 54: extract the planning branch time t from the matching planning trajectory y The subsequent trajectory segments are used as the corresponding real-time planning trajectories;

[0102] Here, because the planning branch time t y The previous time has become historical time, and the corresponding trajectory segment has no effect on the future time, so the planning branch time t is intercepted from the matching planning trajectory y The subsequent trajectory segments are used as the corresponding real-time planning trajectories;

[0103] It should be noted that if the current branch time t in the planned branch time sequence y There are other branch times t after that y To improve the certainty of the output trajectory, the next branch time t can also be set on the real-time planning trajectory. y The subsequent trajectory is deleted in segments;

[0104] Step 55: Send the real-time planned trajectory back to the control module.

[0105] Here, after receiving the real-time planned trajectory, the control module will use it as the latest reference trajectory and control the vehicle's driving state accordingly.

[0106] It should be noted that the above step 5 is for the planning branch time t of the planning branch time series. y More than one or only one but not tend On the contrary, if the planning branch time series of the planning branch time t y There is only one and it is t end , it means that there is only one multi-target motion scene in the future period. In this case, the control module does not need to perform multiple queries, and there is no specific constraint on the query time. The control module can be in the current working cycle period and the period [t,t end ] at any time point within the intersection period of ] to send a planning trajectory query instruction carrying a real-time multi-target trajectory set to the planning module. Correspondingly, the planning module sends a planning branch time t y There is only one and it is t end When the current working cycle period and the period [t,t end ] at any time point within the intersection period of the control module, and send back to the control module a real-time planned trajectory that matches the real-time multi-target trajectory set of the planned trajectory query instruction, specifically including:

[0107] Step A1: In the current working cycle period and the period [t,t end Any time t within the intersection period of ] z Receive the planning trajectory query instruction sent by the control module;

[0108] Where t≤t z ≤t+T p , T p is the working cycle length of the planning module; the planning trajectory query instruction includes a real-time multi-target trajectory set; the real-time multi-target trajectory set includes multiple single-target real-time trajectories c i ; Real-time trajectory of each single target c i The trajectory period is [t s ,t z ], t <t s <t z ;

[0109] Step A2, determining whether the real-time multi-target trajectory set matches the current unique multi-target motion scene for confirmation;

[0110] Specifically: If the trajectory period [t s ,t z ] and the real-time trajectory of each single target c i The first predicted trajectory a i,j If it is not empty, the real-time multi-target trajectory set matches the current unique multi-target motion scene, otherwise it does not match;

[0111] Step A3: If the real-time multi-target trajectory set matches the current unique multi-target motion scene, the unique planning trajectory corresponding to the current unique multi-target motion scene is used as the matching planning trajectory; and the time t is intercepted from the matching planning trajectory. z The subsequent trajectory segments are used as the corresponding real-time planning trajectories; and the real-time planning trajectories are sent back to the control module.

[0112] Figure 2 This is a schematic diagram of the structure of an electronic device provided in the second embodiment of the present invention. The electronic device may be the aforementioned terminal device or server, or may be a terminal device or server connected to the aforementioned terminal device or server to implement the method of the embodiment of the present invention. Figure 2 As shown, the electronic device may include: a processor 301 (such as a CPU), a memory 302, and a transceiver 303; the transceiver 303 is coupled to the processor 301, and the processor 301 controls the transceiver 303. Various instructions may be stored in the memory 302 to complete various processing functions and implement the processing steps described in the aforementioned method embodiment. Preferably, the electronic device involved in the embodiment of the present invention also includes: a power supply 304, a system bus 305 and a communication port 306. The system bus 305 is used to realize the communication connection between components. The above-mentioned communication port 306 is used for connecting and communicating between the electronic device and other peripherals.

[0113] exist Figure 2 The system bus 305 mentioned in the figure can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The system bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 2 The term "communication interface" is represented by only one thick line, but it does not mean that there is only one bus or one type of bus. The communication interface is used to realize the communication between the database access device and other devices (such as clients, read-write libraries, and read-only libraries). The memory may include random access memory (RAM) and may also include non-volatile memory (Non-Volatile Memory), such as at least one disk storage.

[0114] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), a graphics processing unit (GPU), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0115] It should be noted that an embodiment of the present invention further provides a computer-readable storage medium, in which instructions are stored. When the computer-readable storage medium is run on a computer, the computer executes the method and processing process provided in the above embodiments.

[0116] An embodiment of the present invention further provides a chip for executing instructions, wherein the chip is used to execute the processing steps described in the aforementioned method embodiment.

[0117] The embodiment of the present invention provides a processing method, electronic device and computer-readable storage medium for querying a planned trajectory based on a planned branch time, so that the planning module plans the driving trajectory of the vehicle in various possible motion scenarios in the future period based on the multi-target predicted trajectory set sent by the prediction module at the start time of the working cycle to obtain the corresponding planned trajectory set; and compares each planned trajectory in the planned trajectory set to obtain multiple planned trajectory branch time points in the future period; and uses each trajectory branch time point in the same cycle as the planned trajectory query instruction response time point of the control module, and sends the planned trajectory that best matches the actual motion trajectory of the current target back to the control module as the latest reference trajectory. Through the present invention, a planned trajectory reverse query mechanism initiated by the control module and responded by the planning module is established between the planning module and the control module, which solves the problem that the conventional planning module cannot output the planned trajectory multiple times within a working cycle; in addition, the response time point of the planned trajectory reverse query mechanism is based on the planned branch time instead of the predicted branch time between the planning module and the control module, which improves the accuracy of the planned trajectory reverse query.

[0118] The professionals should further realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0119] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0120] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, 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 processing method for planning trajectory query based on planning branch time, characterized in that: The method comprises: The planning module receives the corresponding future time period [t,t end ] is a set of multi-target predicted trajectories; t≤t end ; Construct the future time period [t,t end ] one or more multi-target motion scenes; According to all the multi-target motion scenes, the future time period [t,t end ] to plan the ego-vehicle trajectory and generate the corresponding planning trajectory set; According to the planned trajectory set, the planned trajectory branch time confirmation processing is performed to generate a corresponding planned branch time sequence; the planned branch time sequence includes one or more planned branch time t y ;t <t y <t end , y is the branch time index, 1≤y; In each of the planning branch times t in the current working cycle y Responding to the planned trajectory query instruction sent by the control module, and sending back to the control module a real-time planned trajectory that matches the real-time multi-target trajectory set of the planned trajectory query instruction; The multi-target prediction trajectory set includes multiple single-target prediction trajectory sets A i ; i is the single target index, 1≤i≤n, n is the total number of targets; The single target prediction trajectory set A i including one or more first predicted trajectories a i,j The first predicted trajectory a i,j The trajectory period is [t,t end ]; j is the predicted trajectory index, 1≤j≤m i , m i is the total number of predicted trajectories of single target i; The planned trajectory set includes multiple planned trajectories b x ; The planned trajectory b x The trajectory period is [t,t end ]; x is the planned trajectory index, 1≤x≤m, m is the total number of planned trajectories, m=m i=1 ×m i=2 …×m i=n ; The step of performing planning trajectory branch time confirmation processing according to the planning trajectory set to generate a corresponding planning branch time sequence specifically includes: For any two of the planned trajectories b in the planned trajectory set x The branch time is identified to obtain the corresponding first branch time; When all the first branch times are not 0, time deduplication processing is performed on the obtained multiple first branch times; and each of the first branch times remaining after the deduplication processing is used as the planned branch time t y , and obtain all the planning branch times t y Arrange the planning branches in chronological order to form a corresponding time sequence; The planning branch time t of each of the current working cycle y Responding to the planned trajectory query instruction sent by the control module, and sending back to the control module a real-time planned trajectory that matches the real-time multi-target trajectory set of the planned trajectory query instruction, specifically includes: In each of the planning branch times t in the current working cycle y Receive the planned trajectory query instruction sent by the control module; the planned trajectory query instruction includes the real-time multi-target trajectory set; the real-time multi-target trajectory set includes multiple single-target real-time trajectories c i ; Each of the single target real-time trajectory c i The trajectory period is [t s ,t y ], t <t s <t y ; Confirming the multi-target motion scene that matches the real-time multi-target trajectory set and recording it as a corresponding matching motion scene; The planned trajectory b corresponding to the matching motion scene x As matching planning trajectory; The planning branch time t is intercepted from the matching planning trajectory y The subsequent trajectory segments are used as the corresponding real-time planning trajectories; The real-time planning trajectory is sent back to the control module.

2. The processing method for planning trajectory query based on planning branch time according to claim 1 is characterized in that: The future time period [t,t is constructed according to the multi-target predicted trajectory set end ] one or more multi-target motion scenes, specifically including: From each of the single target prediction trajectory sets A of the multi-target prediction trajectory set i In the above example, select any one of the first predicted trajectories a i,j Combine to generate a corresponding set of prediction trajectory groups; the total number of prediction trajectories of each prediction trajectory group is the target total number n, and the trajectory sets of any two prediction trajectory groups are not repeated; the total number of prediction trajectory groups is m i=1 ×m i=2 …×m i=n ; Each predicted trajectory group is regarded as a future period [t,t end ] of the multi-target motion scene, thereby obtaining one or more of the multi-target motion scenes.

3. The processing method for planning trajectory query based on planning branch time according to claim 2 is characterized in that: The future time period [t,t end ] to plan the self-vehicle trajectory and generate the corresponding planning trajectory set, including: All the first predicted trajectories a in each of the multi-target motion scenes i,j is the future period [t,t end ], and predict the future time period [t,t based on the known starting position and target position of the ego vehicle. end ] to plan the vehicle motion trajectory and generate the corresponding planned trajectory b x ; and all the planned trajectories b are obtained x Form the corresponding planning trajectory set.

4. The processing method for planning trajectory query based on planning branch time according to claim 1 is characterized in that: For any two of the planned trajectories b in the planned trajectory set x The branch time is identified to obtain the corresponding first branch time, specifically including: Step 61: select any two planned trajectories b from the set of planned trajectories. x As the corresponding first and second planned trajectories; the trajectory period of the first and second planned trajectories is the future period [t,t end ]; Step 62, set the start time t a is the future period [t,t end ] start time t, set end time t b is the future period [t,t end ] end time t end ; Step 63, determining whether the first and second planned trajectories are at the starting time t a If the track points match, go to step 64; if not, set the first branch time to 0 and go to step 68; Step 64, determining whether the first and second planned trajectories are at the end time t b If not, go to step 65; if matched, set the first branch time to the end time t b and go to step 68; Step 65, from the starting time t a To the end time t b The central time of the future period is taken as the central time t m ; Step 66: Determine whether the first and second planned trajectories are within the central time t m Whether the trajectory points match; if they match, the starting time t a Let the central time t m , and maintain the end time t b unchanged; if not matched, keep the starting time t a unchanged, and the end time t b Let the central time t m ; Step 67, determine from the starting time t a To the end time t b Whether the length of the future time period is greater than the preset track point time interval △t; if it is greater than the track point time interval △t, go to step 65; if it is less than or equal to the track point time interval △t, set the first branch time to the latest center time t m and go to step 68; Step 68: output the first branch time as a recognition result.

5. The processing method for planning trajectory query based on planning branch time according to claim 2 is characterized in that: The confirming the multi-target motion scene that matches the real-time multi-target trajectory set and recording it as a corresponding matching motion scene specifically includes: In the multi-target predicted trajectory set, the trajectory period [t s ,t y ] and the real-time trajectory of each single target c i The first predicted trajectory a i,j Recorded as the corresponding first trajectory; and all the first trajectories obtained form a corresponding first trajectory group; and the predicted trajectory group matching the first trajectory group is recorded as the corresponding first predicted trajectory group; and the multi-target motion scene corresponding to the first predicted trajectory group is used as the matching motion scene.

6. An electronic device, characterized in that: include: memory, processors, and transceivers; The processor is used to couple with the memory, read and execute instructions in the memory, so as to implement the method according to any one of claims 1 to 5; The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a computer, the computer is enabled to execute the method according to any one of claims 1 to 5.

Images