Automatic driving planning trajectory evaluation method and device, electronic equipment and storage medium

Abstract

The invention discloses an automatic driving planning trajectory evaluation method and device, electronic equipment and a storage medium. The method comprises the steps of receiving trajectory data output by an automatic driving planning module; inputting the trajectory data into an evaluation model, wherein the evaluation model is used for judging whether a second evaluation index is met or not under the condition that the trajectory data meets a first evaluation index; and obtaining a planning trajectory evaluation result according to the evaluation model. Through the method provided by the invention, feasibility evaluation of the output trajectory of the automatic driving planning module is realized. In addition, by means of the method, the comfort level of passengers of the automatic driving vehicle can be improved.

Description

Technical Field

[0001] This application relates to the field of autonomous driving technology, and particularly to a method and device for evaluating a planned trajectory of autonomous driving, as well as an electronic device and a storage medium. Background Art

[0002] In the field of autonomous driving, the planning module of the host vehicle is mainly responsible for decomposing complex driving tasks into a series of executable actions, and ensuring the safe and efficient driving of the vehicle through decision-making and optimization. The main functions of the planning module include decomposing complex driving tasks into a series of small tasks for step-by-step completion. According to the current road conditions and traffic conditions, determine the actions that the vehicle should perform, such as following a vehicle, overtaking, avoiding obstacles, etc. Select the optimal path in the road network map to ensure that the vehicle can reach the destination efficiently and safely, and so on.

[0003] For the trajectory output by the planning module of the host vehicle, especially the trajectory output by the planning AI model, the feasibility of the trajectory needs to be ensured. Summary of the Invention

[0004] Embodiments of this application provide a method and device for evaluating a planned trajectory of autonomous driving, as well as an electronic device and a storage medium, to evaluate the feasibility of the trajectory output by the planning module.

[0005] Embodiments of this application adopt the following technical solutions:

[0006] In a first aspect, embodiments of this application provide a method for evaluating a planned trajectory of autonomous driving, where the evaluation method includes:

[0007] Receiving trajectory data output by an autonomous driving planning module;

[0008] Inputting the trajectory data into an evaluation model, where the evaluation model is used to determine whether the second evaluation index is satisfied when the trajectory data satisfies the first evaluation index;

[0009] Obtaining an evaluation result of the planned trajectory according to the evaluation model.

[0010] In some embodiments, the inputting the trajectory data into an evaluation model, where the evaluation model is used to determine whether the second evaluation index is satisfied when the trajectory data satisfies the first evaluation index includes:

[0011] Determining whether the trajectory data satisfies the first evaluation index;

[0012] If it is determined that the trajectory data satisfies the first evaluation index, then determining whether the trajectory data satisfies the second evaluation index;

[0013] When all of the trajectory data meet the first evaluation index and the second evaluation index, the output result of the evaluation model is obtained.

[0014] In some embodiments, determining whether the trajectory data meets the first evaluation index includes:

[0015] Determining whether the parameters received by the control module for autonomous driving in the trajectory data meet the maximum threshold and the minimum threshold;

[0016] Determining whether the trajectory data meets the restrictions of traffic regulations;

[0017] Determining whether the trajectory data meets the interaction restrictions with other traffic participants;

[0018] Determining whether the trajectory data meets the settings for the comfort of the passengers in the host vehicle;

[0019] If the trajectory data does not meet any one of the indicators, it is considered that the trajectory data does not meet the first evaluation index;

[0020] If the trajectory data meets all of the above indicators, it is considered that the trajectory data meets the first evaluation index.

[0021] In some embodiments, if it is determined that the trajectory data meets the first evaluation index, then determining whether the trajectory data meets the second evaluation index includes:

[0022] If it is determined that the trajectory data meets the first evaluation index, then determining whether there is a risk of the leading vehicle making a panic stop when the host vehicle is following the leading vehicle in the trajectory data;

[0023] If there is a risk of a panic stop, then determining the maximum deceleration of the host vehicle;

[0024] Calculate the maximum deceleration of the host vehicle to ensure that the host vehicle does not make a panic stop when the leading vehicle makes a panic stop during the following process of the host vehicle, then the trajectory data meets the second evaluation index.

[0025] In some embodiments, the method further includes:

[0026] Based on whether the trajectory data meets the condition that the distance traveled by the host vehicle Se - the distance traveled by the leading vehicle So > the distance between the rear of the leading vehicle and the front of the host vehicle S, as the feasibility evaluation result of the second evaluation index;

[0027] Wherein, the distance traveled by the host vehicle Se refers to the distance traveled by the host vehicle after braking at a deceleration of Ae while traveling at Ve1, and the deceleration Ae refers to the acceleration of the host vehicle when braking;

[0028] The forward distance So of the leading vehicle refers to the distance traveled by the leading vehicle when it travels at Vo1 and continuously brakes at a deceleration of Ao for To1 time and then starts to travel at a constant speed. The deceleration Ao refers to the acceleration of the leading vehicle when braking.

[0029] The distance S between the rear of the leading vehicle and the front of the host vehicle refers to the distances traveled by the host vehicle and the leading vehicle at Ve1 and Vo1 respectively when the host vehicle is following the leading vehicle.

[0030] In some embodiments, the evaluation model includes:

[0031] Collect any one or more of the upper and lower bound data of the input control module parameters, traffic regulation data, interaction limit data with other traffic participants, and comfort data of the passengers in the host vehicle as the training data set;

[0032] Train the evaluation model for judging whether the trajectory data output by the autonomous driving planning module meets the first evaluation index and the second evaluation index according to the training data set; and

[0033] Optimize the evaluation model according to the evaluation result of the planned trajectory.

[0034] In some embodiments, the method further includes:

[0035] Deploy the evaluation model online on an autonomous driving vehicle to intercept in the case where the trajectory data output by the autonomous driving planning module is infeasible and re-plan the trajectory.

[0036] In a second aspect, an embodiment of the present application further provides an apparatus for evaluating an autonomous driving planned trajectory, where the apparatus includes:

[0037] A first module for receiving the trajectory data output by the autonomous driving planning module;

[0038] A second module for inputting the trajectory data into an evaluation model, where the evaluation model is used to judge whether the second evaluation index is met when the trajectory data meets the first evaluation index;

[0039] A third module for obtaining an evaluation result of the planned trajectory according to the evaluation model.

[0040] In a third aspect, an embodiment of the present application further provides an electronic device, including: a processor; and a memory arranged to store computer-executable instructions, where the executable instructions, when executed, cause the processor to execute the above method.

[0041] Fourthly, an embodiment of the present application further provides a computer-readable storage medium, which stores one or more programs. When the one or more programs are executed by an electronic device including multiple application programs, the electronic device is enabled to execute the above method.

[0042] At least one of the above technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects: receiving the trajectory data output by the autonomous driving planning module, and then inputting the trajectory data into the evaluation model. It is judged by the evaluation model whether the second evaluation index is satisfied when the trajectory data satisfies the first evaluation index. Finally, according to the evaluation model, the evaluation result of the planned trajectory is obtained. The first evaluation index is used as a mandatory index that must be satisfied, and the second evaluation index is used as a soft index, which is an index that when the vehicle in front brakes suddenly during the process of the vehicle following the self-vehicle, the self-vehicle will not brake suddenly. The evaluation result of the planned trajectory is finally obtained through the judgment results of relevant indexes. Through the above method, not only can the feasibility of the output trajectory of the planning module be evaluated, but also the evaluation result can be fed back to the self-vehicle planning module according to the feasibility evaluation result, changing the output of the planning module, thereby adjusting the comfort of the driver and passengers when the self-vehicle is following. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The drawings described herein are used to provide a further understanding of the present application, and constitute a part of the present application. The schematic embodiments and descriptions thereof of the present application are used to explain the present application, and do not constitute an improper limitation of the present application. In the drawings:

[0044] Figure 1 It is a schematic structural diagram of an evaluation model for an autonomous driving planned trajectory evaluation method in an embodiment of the present application;

[0045] Figure 2 It is a schematic flow diagram of an autonomous driving planned trajectory evaluation method in an embodiment of the present application;

[0046] Figure 3 It is a schematic structural diagram of an autonomous driving planned trajectory evaluation device in an embodiment of the present application;

[0047] Figure 4 It is a schematic diagram of the speed and time relationship curve between the vehicle in front and the self-vehicle in an autonomous driving planned trajectory evaluation method in an embodiment of the present application;

[0048] Figure 5 It is a schematic structural diagram of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0049] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments of this application and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this application without creative efforts belong to the scope of protection of this application.

[0050] The following will, in conjunction with the drawings, elaborate on the technical solutions provided by each embodiment of this application.

[0051] As Figure 1 shown, the evaluation model, that is, the model for evaluating the feasibility of the trajectory, mainly serves as the evaluation of the feasibility of hard indicators. If all the evaluation indicators are met, it is feasible. If any one of the evaluation indicators is not met, it is not feasible. The evaluation model includes the following parts:

[0052] Statistical driving data, including but not limited to collecting data on manual driving in various scenarios. Collect manual driving data and statistically analyze driving data such as acceleration values at the 95% boundaries in the middle. This can ensure the margin of the hard indicators.

[0053] The upper and lower bounds of the acceptable parameters for the traffic regulations of the control module. That is, the result output by the planning module is used as the input of the control module. Therefore, it is necessary to consider the upper and lower bounds of the acceptable parameters in the control module.

[0054] The restrictions of traffic laws, including but not limited to current traffic laws.

[0055] TTC safety distance. When evaluating the feasibility of the collision trajectory, if the TTC safety distance is not met, the status of "not feasible" is output. TTC (Time To Collision) refers to the time required for a vehicle to travel to collide with an obstacle in front while maintaining the current speed and direction. The TTC safety distance is the safety distance within which the vehicle itself does not collide with the vehicle in front during the TTC collision time.

[0056] In addition, passenger comfort is also considered. Passenger comfort is obtained by statistically analyzing the acceleration, acceleration derivative, curvature, and curvature change rate based on the collected manual driving data.

[0057] An embodiment of this application provides a method for evaluating an autonomous driving planned trajectory. As Figure 2 shown, a schematic flowchart of the method for evaluating an autonomous driving planned trajectory in an embodiment of this application is provided. The method at least includes the following steps S210 to step S230:

[0058] It should be noted that the following method for evaluating the planned trajectory of autonomous driving obtains the evaluation result of trajectory feasibility based on the input of the actual scenario, which involves the verification stage of the evaluation model. It can be understood that the input data in the verification stage of the evaluation model is the same as that in the training stage.

[0059] Step S210: Receive the trajectory data output by the autonomous driving planning module.

[0060] The autonomous driving planning module plans and outputs trajectory data based on the perception result of the target at the current moment and the state of the vehicle itself. It can be understood that the current moment is only for example, and in actual use, planning needs to be based on the perception results at each moment / each position. It should be noted that the input of the autonomous driving planning module is not limited, and can be the fusion / independent perception results of various sensors carried by the vehicle itself, the perception results of roadside devices, and the perception results pushed by the cloud, etc.

[0061] Step S220: Input the trajectory data into the evaluation model, and the evaluation model is used to determine whether the second evaluation index is satisfied when the trajectory data meets the first evaluation index.

[0062] According to the trajectory data output by the autonomous driving planning module, the evaluation model is used to respectively evaluate the feasibility of the first evaluation index and the second evaluation index for the trajectory data. The trajectory data includes, but is not limited to, methods based on the RB (RuleBased) rule model, the AI model, and the fusion of the RB rule model and the AI model.

[0063] It should be noted that the "first evaluation index" is a hard evaluation index, and only when the "first evaluation index" is feasible will the feasibility evaluation of the soft evaluation index of the "second evaluation index" be carried out. If the "first evaluation index" cannot be met, it can be considered that the trajectory data does not meet the trajectory feasibility evaluation.

[0064] The first evaluation index can usually be established based on relevant indicators and empirical rules, and in the first evaluation index, hard indicators related to the comfort of the vehicle's passengers can be established according to the data of manual driving; the first evaluation index also establishes hard interaction limit indicators between the vehicle itself and other traffic participants based on the TTC time.

[0065] The second evaluation index is used to evaluate whether the output trajectory data of the planning module meets the feasibility that if the leading vehicle makes a point brake (such as 1 second) in the vehicle following scenario of the vehicle itself, what acceleration should be adopted when the vehicle itself brakes to avoid collision with the leading vehicle on the premise that the trajectory data meets the first evaluation index.

[0066] It can be understood that "tap braking" is to control the vehicle (the vehicle in front) by quickly and intermittently stepping on the brake pedal. The operation method of "tap braking" can usually maintain the stability of the vehicle while decelerating and prevent the wheels from locking up.

[0067] Step S230: Obtain the planned trajectory evaluation result according to the evaluation model.

[0068] Judge whether the trajectory data meets the requirement of feasibility according to the data result of the evaluation model as the planned trajectory evaluation result. If the feasibility of the evaluation result of the planned trajectory is satisfied, a trajectory that meets the safety of human driving can be output. Otherwise, the automatic driving planning module needs to output a degradation plan or a re-planning plan, and then output a trajectory that meets the safety of human driving.

[0069] Through the above method, the trajectory data is input into the evaluation model, and the evaluation model is used to judge whether the second evaluation index is satisfied when the trajectory data meets the first evaluation index. Through the evaluation model, it is necessary to judge not only whether there is a state that violates the hard index (the first evaluation index) in the automatic driving state, but also whether the soft index (the second evaluation index) of the vehicle's own tap braking will be caused when the vehicle in front performs tap braking.

[0070] Through the above method, when obtaining the planned trajectory evaluation result according to the evaluation model, if the feasibility of the evaluation result of the planned trajectory is satisfied, a trajectory that meets the safety of human driving can be output. Otherwise, the automatic driving planning module needs to output a degradation plan or a re-planning plan, and then output a trajectory that meets the safety of human driving. Or, according to the feasibility evaluation result, feedback to the evaluation model to optimize the evaluation model, and obtain the trajectory data that is more in line with the safety of human driving output by the planning module.

[0071] Through the above method, receive the trajectory data output by the automatic driving planning module, including but not limited to, methods based on the RB (Rule Based) rule model, the AI model, and the fusion of the RB rule model and the AI model. Thus, the feasibility evaluation of the trajectory data of different models can be realized.

[0072] Different from the related technology, which lacks the problem of evaluating the feasibility of the trajectory data output by the planning module, through the above method, the evaluation model is deployed on the vehicle during the on-vehicle deployment stage, ensuring that when the planned output trajectory has poor performance, it can intercept and re-plan the trajectory in time to ensure the safety of the automatic driving vehicle and improve the comfort of the passengers and drivers. Maximize comfort while ensuring safety.

[0073] In one embodiment of the present application, inputting the trajectory data into an evaluation model for determining whether a second evaluation criterion is satisfied when the trajectory data meets a first evaluation criterion includes: determining whether the trajectory data meets the first evaluation criterion; if it is determined that the trajectory data meets the first evaluation criterion, determining whether the trajectory data meets the second evaluation criterion; and obtaining an output result of the evaluation model when the trajectory data meets both the first evaluation criterion and the second evaluation criterion.

[0074] First, it is necessary to determine whether the trajectory data meets the first evaluation criterion, that is, whether there are violations of hard criteria during the autonomous driving state. If it is determined that the trajectory data meets the first evaluation criterion, then it is determined whether the trajectory data meets the second evaluation criterion. Secondly, only when the hard criteria are all met will it be determined whether the trajectory data meets the second evaluation criterion. At the same time, when the trajectory data meets both the first evaluation criterion and the second evaluation criterion, finally, an output result of the evaluation model is obtained.

[0075] It should be noted that the output result of the evaluation model will be fed back to the planning module or directly output. If it is fed back to the planning module, the evaluation model can be optimized. If it is directly output, it can be determined whether the output trajectory data of the planning module is feasible.

[0076] In one embodiment of the present application, determining whether the trajectory data meets the first evaluation criterion includes: determining whether the maximum threshold and minimum threshold of the parameters received by the control module for autonomous driving are met in the trajectory data; determining whether the trajectory data meets the restrictions of traffic regulations; determining whether the trajectory data meets the interaction restrictions with other traffic participants; determining whether the trajectory data meets the settings for the comfort of the passengers in the vehicle; if the trajectory data does not meet any one of the criteria, it is considered that the trajectory data does not meet the first evaluation criterion; if the trajectory data meets all of the above criteria, it is considered that the trajectory data meets the first evaluation criterion.

[0077] The model establishment stage includes the evaluation of the feasibility of hard criteria and soft criteria:

[0078] For the feasibility of hard criteria, only the states of feasible and infeasible are output. Specifically, it consists of the following aspects:

[0079] (a) The upper and lower bounds of parameters acceptable to the control module (the downstream module of the planning module), including information such as the maximum acceleration and steering acceleration that the chassis information can provide. When evaluating the feasibility of the trajectory, as long as one of them exceeds the limit, the status of "infeasible" is output. The parameter upper and lower bounds mainly refer to the maximum and minimum values of acceleration, angular velocity, etc. Acceleration corresponds to the brakes / throttle on the vehicle itself, and angular velocity corresponds to the steering of the vehicle's steering wheel.

[0080] (b) The restrictions of traffic regulations, including but not limited to information such as overtaking on the right, lane change across solid lines, driving out of the road boundary, running a red light, etc. When evaluating the feasibility of the trajectory, as long as one of them violates the traffic regulations, the status of "infeasible" is output.

[0081] (c) The interaction restrictions with other traffic participants, including collisions. When evaluating the feasibility of the trajectory, if the TTC safety distance is not met, the status of "infeasible" is output. The RSS model is more conservative than TTC. Using TTC can better conform to the braking scheme of the vehicle following the leading vehicle with the maximum deceleration (-a) when the leading vehicle "jams on the brakes".

[0082] (d) The settings of the comfort of the vehicle's passengers. Based on the collected manual driving data, statistical analysis is carried out on acceleration, acceleration derivative, curvature, and curvature change rate. The upper and lower bounds of each item are taken from the data values of the middle 95% of the data distribution. When evaluating the feasibility of the trajectory, as long as one of them exceeds the limit, the status of "infeasible" is output.

[0083] That is, the expression of the hard index feasibility evaluation model is:

[0084]

[0085] Among them,

[0086] T is the trajectory to be evaluated.

[0087] Ea(T) is the acceptability evaluation function of acceleration and steering acceleration.

[0088] Et(T) is the evaluation function of compliance with traffic regulations.

[0089] Ec(T) is the interactive safety evaluation function with other traffic participants, that is, including but not limited to collisions. When evaluating the feasibility of the trajectory, if the TTC safety distance is not met, the status of "infeasible" is output.

[0090] Es(T) is the evaluation function of passenger comfort.

[0091] The above evaluation functions can adopt the evaluation functions common among those skilled in the art, and are not specifically limited in the embodiments of the present application.

[0092] The source of the acceleration signal is mainly the acceleration signal provided by the IMU or through vehicle speed differentiation.

[0093] The steering acceleration acceptability of an autonomous vehicle refers to the impact of the acceleration change on passenger comfort and vehicle stability when the vehicle performs a steering operation in the autonomous driving mode. This impact can be measured in the following aspects:

[0094] Acceleration change rate: In the autonomous driving mode, when the vehicle performs a steering operation, the acceleration change rate has an important impact on passenger comfort. An excessive acceleration change rate will cause passengers to feel uncomfortable and may even cause problems such as motion sickness.

[0095] Vehicle stability: During the steering process, the stability of the vehicle is crucial. The autonomous driving system needs to continuously monitor the vehicle's state through sensors to ensure that the vehicle does not lose control during steering. This includes precise control of factors such as tire grip and vehicle center of gravity change to maintain the vehicle's stability during steering.

[0096] Passenger comfort: In addition to vehicle stability, passenger comfort is also an important indicator for measuring steering acceleration acceptability. The autonomous driving system needs to minimize sudden acceleration and deceleration as much as possible to avoid discomfort for passengers. This can be achieved by optimizing algorithms and adjusting steering strategies.

[0097] Environmental perception and decision-making: The autonomous driving system needs to rely on high-precision environmental perception technology to obtain information about the surrounding environment and make reasonable steering decisions based on this information. This helps to reduce unexpected acceleration or deceleration caused by environmental changes, thereby improving passenger comfort and safety.

[0098] In an embodiment of the present application, if it is determined that the trajectory data meets the first evaluation index, then it is determined whether the trajectory data meets the second evaluation index, including: if it is determined that the trajectory data meets the first evaluation index, then it is determined whether there is a risk of the leading vehicle braking suddenly when the host vehicle is following the leading vehicle in the trajectory data; if there is a risk of sudden braking, then the maximum deceleration of the host vehicle is determined; calculate the maximum deceleration of the host vehicle to ensure that the host vehicle does not brake suddenly when the leading vehicle brakes suddenly during the following process of the host vehicle, then the trajectory data meets the second evaluation index.

[0099] The feasibility of the soft index refers to whether there is a risk of sudden braking when the output of the hard feasibility is "feasible". That is to say, it is judged whether there is a state that violates the hard index in the autonomous driving state, and whether the host vehicle will brake suddenly when the leading vehicle brakes suddenly, which is a soft index. In short, the soft index is an index to ensure that the host vehicle will not brake suddenly (maximum deceleration) when the leading vehicle brakes suddenly.

[0100] If there is a risk of point braking, the maximum deceleration of the host vehicle is judged. Further, the maximum deceleration of the host vehicle is calculated and it is judged that the maximum deceleration is within the threshold range, so as to ensure that the host vehicle does not perform point braking when the leading vehicle performs point braking during the following process of the host vehicle. The design idea for the above scenario is: in the case of the host vehicle following in a driving scenario, if an abnormal accident occurs in the front, the host vehicle has to perform emergency braking, which is considered a reasonable situation and is not within the scope of soft indicators. However, in many cases, when the leading vehicle has short-term accidental braking due to inattentive driving or being too close to the vehicle in the adjacent lane, etc., it is necessary to analyze this situation, that is, to ensure that the host vehicle does not perform emergency braking (or perform point braking like the leading vehicle) in this situation, that is, the maximum deceleration of the host vehicle is within the threshold range.

[0101] It should be noted that the above maximum deceleration of the host vehicle can be set in the soft indicators. For example, if the maximum deceleration of the host vehicle is set to -1, the following distance of the leading vehicle will be relatively far (conservative); or if the maximum deceleration of the host vehicle is set to a maximum of -3, the following distance of the leading vehicle will be relatively close (aggressive). Generally, when the maximum deceleration is small, the following distance of the leading vehicle is relatively close and the host vehicle will perform emergency braking; when the maximum deceleration is large, the following distance of the leading vehicle is relatively far and the host vehicle will perform point braking. For example, when the leading vehicle performs point braking for 1 s, it is judged whether the maximum deceleration of the host vehicle exceeds the threshold. If it exceeds, it means that the host vehicle has performed point braking and the following distance of the host vehicle should be increased.

[0102] In an embodiment of the present application, the method further includes: according to whether the trajectory data satisfies that the distance Se traveled by the host vehicle during the process from the leading vehicle braking until the speed of the host vehicle braking is the same as that of the leading vehicle - the distance So traveled by the leading vehicle > the distance S between the rear of the leading vehicle and the front of the host vehicle, as the feasibility evaluation result of the second evaluation index; wherein, the distance Se traveled by the host vehicle refers to the distance traveled by the host vehicle after braking at a deceleration of Ae while traveling at Ve1, and the deceleration Ae refers to the acceleration of the host vehicle when braking; the distance So traveled by the leading vehicle refers to the distance traveled by the leading vehicle after braking at a deceleration of Ao for a duration of To1 and then starting to travel at a constant speed, and the deceleration Ao refers to the acceleration of the leading vehicle when braking; the distance S between the rear of the leading vehicle and the front of the host vehicle refers to the distance between the two vehicles when the host vehicle and the leading vehicle are traveling following each other at Ve1 and Vo1 respectively before the leading vehicle brakes.

[0103] When the leading vehicle is traveling following, assuming that the leading vehicle suddenly brakes for a short time and both before and after braking, the leading vehicle maintains a constant speed. On the premise of not colliding, the host vehicle takes braking with a threshold deceleration and the two vehicles do not collide.

[0104] Please refer to such as Figure 4As shown in Table 1, when following a vehicle, the host vehicle and the leading vehicle are traveling at speeds Ve1 and Vo1 respectively. The distance between the rear of the leading vehicle and the front of the host vehicle is S. The leading vehicle suddenly brakes with an acceleration of Ao for a duration of To1, and its speed drops to Vo2, then it continues to travel at a constant speed of Vo2.

[0105] Vo2 = Ve2

[0106] Vo2 = Vo1 + Ao * To1

[0107] Ve2 = Ve1 + Ae * Te2

[0108] Please refer to Figure 4 , then the host vehicle travels at a constant speed of Ve1 during the reaction time Te1, and then brakes with an acceleration of Ae until its speed is the same as that of the leading vehicle. The distance traveled by the leading vehicle is So, and the distance traveled by the host vehicle is Se. Eventually, the two vehicles will not collide, and this situation is feasible; otherwise, it is infeasible.

[0109] So = (Vo1 + Vo2) / 2 * To1 + Vo2 * (Te1 + Te2 - To1)

[0110] Se = Ve1 * Te1 + (Ve1 + Ve2) / 2 * Te2

[0111] Among them, Vo1 and Ve1 can be obtained through the autonomous driving system, and the preset values include: Ao, To1, Te1, and the braking acceleration threshold Ae used to judge whether it is an emergency brake.

[0112] According to whether Se - So > S is satisfied, the feasibility of the soft index is obtained. If Se - So > S, it is considered that the soft index is feasible; if not, it is considered that the soft index is infeasible. That is to say, when Se - So > S, it is ensured that the host vehicle does not perform an emergency brake when the leading vehicle performs a point brake during the following process. Conversely, during the following process of the host vehicle, the host vehicle will perform an emergency brake when the leading vehicle performs a point brake. The soft index is whether the leading vehicle's point brake will cause the host vehicle to perform an emergency brake. If the soft index is feasible, it is considered that there will be no situation where the leading vehicle's point brake causes the host vehicle to perform an emergency brake. If the soft index is infeasible, it is considered that there will be a situation where the leading vehicle's point brake causes the host vehicle to perform an emergency brake.

[0113] It should be noted that due to the emergency brake of the leading vehicle and the host vehicle being below the emergency braking threshold, the magnitude of the acceleration is compared, |Ao| > |Ae|.

[0114] Table 1

[0115]

[0116] In one embodiment of the present application, the evaluation model includes: collecting any one or more of the upper and lower bound data of input control module parameters, traffic regulation data, interaction limit data with other traffic participants, and self-vehicle passenger comfort data as a training data set; training the evaluation model for determining whether the trajectory data output by the autonomous driving planning module meets the first evaluation index and the second evaluation index according to the training data set; and optimizing the evaluation model according to the evaluation result of the planned trajectory.

[0117] First, the collected manual driving data is preprocessed and used as one of the training data sets. The valid information in the training data set includes, but is not limited to, the speed of the vehicle in front, the speed of the vehicle behind, the heading angle of the vehicle in front, the heading angle of the vehicle behind, the relative position between the front and the back, etc. Secondly, the upper and lower bound data of the input control module parameters, the interaction limit data with other traffic participants, and the self-vehicle passenger comfort data are used as one of the training data sets. Finally, feature extraction is performed on the training data set to obtain relevant features, and training is carried out on whether the trajectory to be evaluated is "feasible" or "infeasible".

[0118] Then, the training data set is used to train the evaluation model for determining whether the trajectory data output by the autonomous driving planning module meets the first evaluation index and the second evaluation index.

[0119] In order to obtain a better evaluation model, it is also necessary to optimize the evaluation model according to the evaluation result of the planned trajectory, such as using backpropagation.

[0120] In one embodiment of the present application, the method further includes: deploying the evaluation model online on an autonomous driving vehicle to intercept when the trajectory data output by the autonomous driving planning module is infeasible, and re-planning the trajectory.

[0121] A feedback relationship can be established between the trajectory feasibility of the planning module and the planning module. During the on-vehicle deployment stage, the evaluation model is deployed on the vehicle to ensure that it can be intercepted in time when the planned trajectory has poor performance, and the trajectory is re-planned to ensure the safety of the autonomous driving vehicle.

[0122] The evaluation model is used to perform feasibility evaluations on the trajectory data for the first evaluation index and the second evaluation index respectively. The trajectory data includes, but is not limited to, methods based on the RB (Rule Based) rule model, the AI model, and the fusion of the RB rule model and the AI model.

[0123] The embodiment of the present application also provides an autonomous driving planned trajectory evaluation device 300, as Figure 3As shown in the figure, a schematic structural diagram of the automatic driving planned trajectory evaluation device in the embodiments of the present application is provided. The automatic driving planned trajectory evaluation device 300 at least includes: a first module 310, a second module 320, and a third module 330, where:

[0124] In an embodiment of the present application, the first module 310 is specifically configured to: receive the trajectory data output by the automatic driving planning module.

[0125] The automatic driving planning module plans and outputs trajectory data according to the perception result of the target at the current moment and in combination with the state of the vehicle itself. It can be understood that the current moment is only an example, and in actual use, planning needs to be based on the perception results at each moment / each position. It should be noted that the input of the automatic driving planning module is not limited, and can be the fusion / independent perception results of various sensors carried by the vehicle itself, the perception results of roadside devices, and the perception results pushed by the cloud, etc.

[0126] In an embodiment of the present application, the second module 320 is specifically configured to: input the trajectory data into an evaluation model, and the evaluation model is used to judge whether the second evaluation index is satisfied when the trajectory data satisfies the first evaluation index.

[0127] According to the trajectory data output by the automatic driving planning module, the evaluation model is used to respectively evaluate the feasibility of the first evaluation index and the second evaluation index for the trajectory data. The trajectory data includes, but is not limited to, methods based on the RB (RuleBased) rule model, the AI model, and the fusion of the RB rule model and the AI model.

[0128] It should be noted that the "first evaluation index" is a hard evaluation index, and only on the premise that the "first evaluation index" is feasible will the feasibility evaluation of the soft evaluation index of the "second evaluation index" be carried out. If the "first evaluation index" cannot be satisfied, it can be considered that the trajectory data does not meet the trajectory feasibility evaluation.

[0129] The first evaluation index can usually be established according to relevant indexes and empirical rules, and in the first evaluation index, hard indexes related to the comfort of the vehicle's passengers can be established based on the data of manual driving; the first evaluation index also establishes hard indexes for the interaction restrictions between the vehicle itself and other traffic participants based on the TTC time.

[0130] The second evaluation index is used to evaluate whether the output trajectory data of the planning module meets the feasibility that when the vehicle in front brakes suddenly in the vehicle following scenario of the vehicle itself, what acceleration should be adopted when the vehicle itself brakes to avoid collision with the vehicle in front on the premise that the trajectory data meets the first evaluation index.

[0131] In one embodiment of the present application, the third module 330 is specifically configured to: obtain a planned trajectory evaluation result according to the evaluation model.

[0132] Judge whether the trajectory data meets the requirements of feasibility according to the data result of the evaluation model, as the planned trajectory evaluation result. If the feasibility of the evaluation result of the planned trajectory is satisfied, a trajectory meeting the safety of human driving can be output. Otherwise, the automatic driving planning module needs to output a downgraded plan or a replanning plan, and then output a trajectory meeting the safety of human driving.

[0133] In one embodiment of the present application, the second module 320 is further configured to:

[0134] Judge whether the trajectory data meets the first evaluation index;

[0135] If it is judged that the trajectory data meets the first evaluation index, then judge whether the trajectory data meets the second evaluation index;

[0136] When the trajectory data meets both the first evaluation index and the second evaluation index, obtain the output result of the evaluation model.

[0137] In one embodiment of the present application, the second module 320 is further configured to:

[0138] Judge whether the trajectory data meets the maximum threshold and the minimum threshold of the parameters received by the control module for automatic driving;

[0139] Judge whether the trajectory data meets the restrictions of traffic regulations;

[0140] Judge whether the trajectory data meets the interaction restrictions with other traffic participants;

[0141] Judge whether the trajectory data meets the settings of the comfort of the passengers in the vehicle;

[0142] If the trajectory data does not meet any one of the indicators, it is considered that the trajectory data does not meet the first evaluation index;

[0143] If the trajectory data meets all the above indicators, it is considered that the trajectory data meets the first evaluation index.

[0144] In one embodiment of the present application, the second module 320 is further configured to:

[0145] If it is judged that the trajectory data meets the first evaluation index, then judge whether there is a risk of the vehicle in front suddenly braking when the vehicle is following;

[0146] If there is a risk of sudden braking, judge the maximum deceleration of the vehicle;

[0147] If the maximum deceleration of the host vehicle is judged to ensure that the host vehicle does not brake when the leading vehicle brakes gently during the following process of the host vehicle, the trajectory data meets the second evaluation index.

[0148] In an embodiment of the present application, the second module 320 is further configured to:

[0149] Use whether the trajectory data meets the condition that the distance traveled by the host vehicle Se - the distance traveled by the leading vehicle So > the distance between the rear of the leading vehicle and the front of the host vehicle S as the feasibility evaluation result of the second evaluation index;

[0150] Wherein, the distance traveled by the host vehicle Se refers to the distance traveled by the host vehicle after braking at a deceleration of Ae while traveling at Ve1;

[0151] The distance traveled by the leading vehicle So refers to the distance traveled by the leading vehicle after continuously braking for To1 time at a deceleration of Ao while traveling at Vo1 and then starting to travel at a constant speed;

[0152] The distance between the rear of the leading vehicle and the front of the host vehicle S refers to the distance traveled by the host vehicle and the leading vehicle at Ve1 and Vo1 respectively during the following process of the host vehicle.

[0153] In an embodiment of the present application, the evaluation model includes:

[0154] Collect any one or more of the upper and lower bound data of the input control module parameters, traffic regulation data, interaction limit data with other traffic participants, and host vehicle passenger comfort data as the training data set;

[0155] Train the evaluation model for judging whether the trajectory data output by the autonomous driving planning module meets the first evaluation index and the second evaluation index according to the training data set; and

[0156] Optimize the evaluation model according to the planned trajectory evaluation result.

[0157] In an embodiment of the present application, it further includes a fourth module, which is further configured to:

[0158] Deploy the evaluation model online on an autonomous driving vehicle to intercept when the trajectory data output by the autonomous driving planning module is infeasible and re-plan the trajectory.

[0159] It can be understood that the above-mentioned autonomous driving planned trajectory evaluation device can implement each step of the autonomous driving planned trajectory evaluation method provided in the foregoing embodiment. The relevant explanations about the autonomous driving planned trajectory evaluation method are applicable to the autonomous driving planned trajectory evaluation device and will not be elaborated here.

[0160] Figure 5This is a schematic structural diagram of an electronic device according to an embodiment of the present application. Please refer to Figure 5 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. Among them, the memory may include a memory, such as a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory, etc. Of course, the electronic device may also include other hardware required for other services.

[0161] The processor, network interface, and memory can be interconnected through an internal bus, and the internal bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of representation, Figure 5 only a two-way arrow is used in

[0162] but it does not mean that there is only one bus or one type of bus.

[0163] The memory is used to store programs. Specifically, the program may include program code, and the program code includes computer operation instructions. The memory may include a memory and a non-volatile memory, and provide instructions and data to the processor.

[0164] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming an automatic driving planning trajectory evaluation device at the logical level. The processor executes the program stored in the memory and is specifically used to perform the following operations:

[0165] Receive the trajectory data output by the automatic driving planning module;

[0166] Input the trajectory data into the evaluation model, and the evaluation model is used to judge whether the second evaluation index is satisfied when the trajectory data meets the first evaluation index;

[0167] The above is as described in the present application Figure 2The method executed by the automatic driving planning trajectory evaluation device disclosed in the illustrated embodiment can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit in the hardware of the processor or the instructions in the form of software. The above-mentioned processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may 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. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present application can be directly embodied as being executed and completed by the hardware decoding processor, or executed and completed by the combination of the hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory, and the processor reads the information in the memory and combines its hardware to complete the steps of the above method.

[0168] The electronic device can also execute Figure 2 the method executed by the automatic driving planning trajectory evaluation device in Figure 2 the illustrated embodiment and implement the functions of the automatic driving planning trajectory evaluation device in

[0169] The embodiments of the present application also propose a computer-readable storage medium that stores one or more programs. The one or more programs include instructions that, when executed by an electronic device including multiple application programs, can enable the electronic device to execute Figure 2 the method executed by the automatic driving planning trajectory evaluation device in the illustrated embodiment, and specifically used to execute:

[0170] Receive the trajectory data output by the automatic driving planning module;

[0171] Input the trajectory data into an evaluation model, and the evaluation model is used to determine whether the second evaluation index is satisfied when the trajectory data satisfies the first evaluation index;

[0172] According to the evaluation model, a planned trajectory evaluation result is obtained.

[0173] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0174] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for realizing the functions specified in Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0175] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that realize the functions specified in Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0176] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for realizing the functions specified in Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0177] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and a memory.

[0178] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0179] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0180] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0181] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0182] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for evaluating an autonomous driving planning trajectory, wherein: The evaluation methods include: Receive trajectory data output by the autonomous driving planning module; Inputting the trajectory data into an evaluation model, wherein the evaluation model is used to determine whether the trajectory data satisfies a second evaluation indicator when the trajectory data satisfies a first evaluation indicator; According to the evaluation model, a planning trajectory evaluation result is obtained.

2. The method of claim 1, wherein: The step of inputting the trajectory data into an evaluation model, wherein the evaluation model is used to determine whether the trajectory data satisfies a second evaluation indicator when the trajectory data satisfies a first evaluation indicator, includes: Determining whether the trajectory data meets a first evaluation indicator; If it is determined that the trajectory data meets the first evaluation index, then determining whether the trajectory data meets the second evaluation index; When the trajectory data all satisfy the first evaluation index and the second evaluation index, an output result of the evaluation model is obtained.

3. The method of claim 2, wherein: The determining whether the trajectory data satisfies a first evaluation index includes: Determining whether the trajectory data satisfies a maximum threshold and a minimum threshold of parameters received by a control module for autonomous driving; Determining whether the trajectory data meets the restrictions of traffic regulations; Determining whether the trajectory data satisfies the interaction restriction with other traffic participants; Determining whether the trajectory data satisfies the setting of the comfort level of the passengers in the vehicle; If the trajectory data does not meet any one of the indicators, it is considered that the trajectory data does not meet the first evaluation indicator; If the trajectory data all meet the above indicators, it is considered that the trajectory data meets the first evaluation indicator.

4. The method of claim 2, wherein: If it is determined that the trajectory data satisfies the first evaluation index, then determining whether the trajectory data satisfies the second evaluation index includes: If it is determined that the trajectory data meets the first evaluation index, determining whether there is a risk of a front vehicle braking when the self-vehicle is following the vehicle in the trajectory data; If there is a risk of braking, determine the maximum deceleration of the vehicle; The maximum deceleration of the vehicle is calculated to ensure that the vehicle does not brake suddenly when the vehicle in front brakes suddenly during the following process, and the trajectory data meets the second evaluation index.

5. The method of claim 4, wherein: The method further comprises: Whether the trajectory data satisfies the condition that the distance Se of the self-vehicle traveled from the time when the front vehicle brakes to the time when the self-vehicle brakes at the same speed as the front vehicle - the distance So of the front vehicle traveled > the distance S between the rear end of the front vehicle and the front end of the self-vehicle, is used as the feasibility evaluation result of the second evaluation indicator; The distance Se traveled by the vehicle refers to the distance traveled by the vehicle after the vehicle travels according to Ve1 and brakes according to Ae deceleration, and Ae deceleration refers to the acceleration of the vehicle when braking; The distance So advanced by the preceding vehicle refers to the distance advanced by the preceding vehicle after the preceding vehicle is traveling at Vo1 and braking at Ao deceleration for To1 time and then starts to travel at a constant speed. The Ao deceleration refers to the acceleration of the preceding vehicle when braking. The distance S between the rear end of the leading vehicle and the front end of the own vehicle refers to the distance between the own vehicle and the leading vehicle when the own vehicle and the leading vehicle are traveling at Ve1 and Vo1 respectively before the leading vehicle brakes.

6. The method of claim 1, wherein: The evaluation model comprises: Collect any one or more of the following data as training data sets: upper and lower bounds of input control module parameters, traffic regulations, interaction restriction data with other traffic participants, and comfort data of passengers in the vehicle; According to the training data set, training the evaluation model for determining whether the trajectory data output by the autonomous driving planning module satisfies the first evaluation index and the second evaluation index; and The evaluation model is optimized according to the evaluation result of the planning trajectory.

7. The method of claim 1, wherein: The method further comprises: The evaluation model is deployed online on the autonomous driving vehicle to intercept and re-plan the trajectory when the trajectory data output by the autonomous driving planning module is not feasible.

8. An autonomous driving planning trajectory evaluation device, wherein: The device comprises: The first module is used to receive the trajectory data output by the automatic driving planning module; A second module is used to input the trajectory data into an evaluation model, and the evaluation model is used to determine whether the trajectory data satisfies a second evaluation index when the trajectory data satisfies the first evaluation index; The third module is used to obtain a planning trajectory evaluation result according to the evaluation model.

9. An electronic device, comprising: processor; as well as A memory arranged to store computer executable instructions, which when executed cause the processor to perform the method of any one of claims 1 to 7.

10. A computer-readable storage medium storing one or more programs, which, when executed by an electronic device including a plurality of application programs, causes the electronic device to execute any one of the methods of claims 1 to 7.

Images