Vehicle path planning method and device, vehicle and storage medium

Abstract

The invention discloses a vehicle path planning method and device, a vehicle and a storage medium, and relates to the technical field of vehicles. The method comprises the following steps: acquiring driving related parameters of a vehicle at the current moment, wherein the driving related parameters at least comprise vehicle positioning information, obstacle information, map information, navigation information and predicted trajectory information of traffic participants around the vehicle; the driving related parameters are input into a pre-trained large language model, an interaction text is obtained, and the interaction text is used for representing a driving strategy of the vehicle after the current moment; inputting the map information, the navigation information and the interactive text into a pre-trained path planning model to obtain a planned path output by the path planning model; and controlling the vehicle to run according to the planned path. Therefore, a more suitable driving strategy and a more suitable driving path can be flexibly deduced, and the vehicle is controlled to drive according to the more suitable driving path, so that the driving safety and the driving efficiency of the vehicle are improved.

Description

Technical Field

[0001] The present application relates to the field of vehicle technology, and in particular to a vehicle path planning method, device, vehicle and storage medium. Background Art

[0002] In the field of autonomous driving, the vehicle obtains information about the surrounding environment through the environmental perception module, and combines the vehicle's own position information and motion status information to make behavioral decisions, path planning and motion control, and generate corresponding control instructions; the control instructions are then sent to the vehicle's chassis actuators to achieve autonomous driving of the vehicle.

[0003] In related technologies, driving trajectories are generally generated using search algorithms. However, this generation method is relatively rigid, so that the generated driving trajectory is often not optimal, which greatly affects the performance of vehicle autonomous driving. Summary of the invention

[0004] The present application proposes a vehicle path planning method, device, vehicle and storage medium to better implement vehicle path planning.

[0005] In a first aspect, an embodiment of the present application provides a vehicle path planning method, the method comprising: obtaining driving-related parameters of the vehicle at a current moment, the driving-related parameters including at least the vehicle's positioning information, obstacle information, map information, navigation information, and predicted trajectory information of traffic participants around the vehicle; inputting the driving-related parameters into a pre-trained large language model to obtain an interactive text output by the large language model, the interactive text being used to characterize the vehicle's driving strategy after the current moment; inputting the map information, the navigation information, and the interactive text into a pre-trained path planning model to obtain a planned path output by the path planning model; and controlling the vehicle to travel along the planned path.

[0006] In the second aspect, an embodiment of the present application provides a vehicle path planning device, which includes: a parameter acquisition module, a strategy generation module, a path planning module and a control module. The parameter acquisition module is used to obtain the driving-related parameters of the vehicle at the current moment, and the driving-related parameters at least include the vehicle's positioning information, obstacle information, map information, navigation information and predicted trajectory information of traffic participants around the vehicle; the strategy generation module is used to input the driving-related parameters into a pre-trained large language model to obtain an interactive text output by the large language model, and the interactive text is used to characterize the driving strategy of the vehicle after the current moment; the path planning module is used to input the map information, the navigation information and the interactive text into a pre-trained path planning model to obtain a planned path output by the path planning model; the control module is used to control the vehicle to travel according to the planned path.

[0007] In a third aspect, an embodiment of the present application provides a vehicle, comprising: one or more processors; a memory; one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs are configured to execute the above-mentioned method.

[0008] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which a program code is stored. The program code can be called by a processor to execute the above method.

[0009] In the solution provided by the present application, the driving-related parameters of the vehicle at the current moment are obtained, and the driving-related parameters at least include the vehicle positioning information, obstacle information, map information, navigation information, and predicted trajectory information of traffic participants around the vehicle; the driving-related parameters are input into the pre-trained large language model to obtain the interactive text output by the large language model, and the interactive text is used to characterize the driving strategy of the vehicle after the current moment; the map information, navigation information, and interactive text are input into the pre-trained path planning model to obtain the planned path output by the path planning model; and the vehicle is controlled to drive according to the planned path. In this way, the pre-trained large language model can flexibly deduce and output a more appropriate driving strategy for the driving scene after understanding the driving scene of the vehicle at the current moment based on the driving-related parameters of the vehicle, and then the pre-trained path planning model generates a more appropriate driving path according to the more appropriate driving strategy, map information, and navigation information, and controls the vehicle to drive according to the more appropriate driving path, thereby improving the driving safety and driving efficiency of the vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0011] Figure 1 A flow chart of a vehicle path planning method provided in an embodiment of the present application is shown.

[0012] Figure 2 A flow chart of a vehicle path planning method provided in another embodiment of the present application is shown.

[0013] Figure 3 Shows Figure 2 A flowchart of the sub-steps of step S230 in one implementation manner.

[0014] Figure 4 A model architecture diagram of a path planning model provided in an embodiment of the present application is shown.

[0015] Figure 5 Shows Figure 2 A flowchart of the sub-steps of step S240 in one implementation manner.

[0016] Figure 6 A flow chart of a model training method provided in one embodiment of the present application is shown.

[0017] Figure 7 A model architecture diagram of an initial language model provided in an embodiment of the present application is shown.

[0018] Figure 8 It is a block diagram of a vehicle path planning device provided according to an embodiment of the present application.

[0019] Fig. 9 It is a block diagram of a vehicle for executing a vehicle path planning method according to an embodiment of the present application.

[0020] Fig.10 It is a storage unit of an embodiment of the present application for storing or carrying program codes for implementing the vehicle path planning method according to an embodiment of the present application. DETAILED DESCRIPTION

[0021] In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0022] It should be noted that in some processes described in the specification, claims and the above-mentioned drawings of the present application, multiple operations that appear in a specific order are included, and these operations may not be executed or executed in parallel in the order in which they appear in this article. The sequence numbers of operations such as S110, S120, etc. are only used to distinguish different operations, and the sequence numbers themselves do not represent any execution order. In addition, these processes may include more or less operations, and these operations may be executed in sequence or in parallel. And, the terms "first", "second", etc. in the specification, claims and the above-mentioned drawings of the present application are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or server that includes a series of steps or sub-modules is not necessarily limited to those steps or sub-modules explicitly listed, but may include other steps or sub-modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0023] In related technologies, driving trajectories are generally generated using search algorithms. The search algorithm mainly scatters points, then connects the scattered points into multiple lines, and optimizes each line. This method is relatively rigid and inflexible, and cannot cover scenarios where deceleration is required in schools, construction sections, etc. Moreover, in this method, time and space separation is difficult to interact, that is, it is impossible to generate trajectories / control sequences by performing lateral and longitudinal actions at the same time, and it is impossible to generate a driving trajectory of "first slowing down to let pedestrian A pass, then accelerating to overtake vehicle B on the right".

[0024] In view of the above problems, the inventor proposes a vehicle path planning method, device, vehicle and storage medium. The vehicle path planning method provided in the embodiment of the present application is described in detail below.

[0025] Please refer to Figure 1 , Figure 1 A flow chart of a vehicle path planning method provided in an embodiment of the present application. Figure 1 The vehicle path planning method provided in the embodiment of the present application is described in detail. The vehicle path planning method may include the following steps:

[0026] Step S110: Acquire the driving-related parameters of the vehicle at the current moment, wherein the driving-related parameters at least include the vehicle positioning information, obstacle information, map information, navigation information, and predicted trajectory information of traffic participants around the vehicle.

[0027] In this embodiment, the vehicle positioning information can be obtained through the Global Positioning System (GPS), and the obstacle information can include the coordinate position, speed, direction and obstacle type of the obstacle. The obstacle information can be determined based on the sensor data collected by the camera, laser radar and other sensors of the vehicle. Map information can include static information such as lane lines, road topology, traffic rules and road attributes. Traffic participants around the vehicle can include other vehicles and other organisms (for example, pedestrians or animals) around the vehicle; the predicted trajectory information of the traffic participants around the vehicle can be obtained by predicting and analyzing the current coordinate position, speed and direction of the traffic participants.

[0028] Step S120: input the driving related parameters into a pre-trained large language model to obtain an interactive text output by the large language model, wherein the interactive text is used to characterize the driving strategy of the vehicle after the current moment.

[0029] Optionally, the large language model can be obtained by training a language model based on Transformer. That is to say, through a large amount of training, the large language model can understand the driving scenario of the vehicle at the current moment through the input driving-related parameters, and deduce the driving strategy of the vehicle after the current moment based on the understood driving scenario, that is, output the aforementioned interactive text.

[0030] Exemplarily, based on the input driving-related parameters, the large language model understands that the vehicle's current driving scenario is "unprotected left turn, pedestrian A is walking towards each other on the left sidewalk, and car B on the right is driving from right to left relative to the vehicle"; based on this, the large language model can deduce a driving strategy such as "the vehicle first slows down to let pedestrian A pass, and then accelerates to overtake car B on the right" and outputs it as the above-mentioned interactive text.

[0031] Optionally, the large language model can deduce an optimal driving strategy, that is, output an optimal interaction text; of course, the large language model can also deduce multiple driving strategies, that is, output multiple interaction texts, which is not limited in this embodiment.

[0032] Step S130: inputting the map information, the navigation information and the interactive text into a pre-trained path planning model to obtain a planned path output by the path planning model.

[0033] The path planning model may be obtained by training an initial planning model based on an Actor-Critic Algorithm using a large amount of sample data.

[0034] For example, the interactive text is "the vehicle first slows down to let pedestrian A pass, then accelerates to overtake vehicle B on the right".

[0035] At this time, the planned path output by the path planning model can be "after pedestrian A passes through the sidewalk, a driving trajectory is generated behind pedestrian A, and a driving trajectory is generated in front of the right vehicle B."

[0036] Step S140: Control the vehicle to travel along the planned path.

[0037] In this embodiment, that is to say, in addition to outputting the planned path of the vehicle selected under the semantics of the interactive text, the path planning model also outputs the control sequence for controlling the vehicle to travel along the planned path. The control sequence includes at least the acceleration parameter and the steering wheel angle parameter of the vehicle. Based on this, the acceleration parameter and the steering wheel angle parameter can be transmitted to the actuator control module of the vehicle through the CAN bus, and then the actuator control module controls the vehicle to travel along the above-mentioned planned path according to the acceleration parameter and the steering wheel angle parameter.

[0038] Specifically, the above planning path and control sequence can be expressed by the following two formulas:

[0039] traj={(x,y)|x=f(t),y=g(t),t∈T}

[0040] ControlSeq={(a,SteerAngle)|a=h(t),SteerAngle=m(t),t∈T}

[0041] Among them, traj is used to represent the planned path, (x, y) is the coordinate position of the trajectory point in the planned path, t represents each prediction moment, T represents the prediction time domain (for example, 3 seconds), function f(t) is used to represent the horizontal coordinate value of the trajectory point that changes with the prediction moment, and function g(t) is used to represent the vertical coordinate value of the trajectory point that changes with the prediction moment. ControlSeq is used to represent the control sequence, a is the acceleration parameter, SteerAngle is the steering wheel angle parameter, function h(t) is used to represent the parameter value of the acceleration parameter of the vehicle that changes with the prediction moment, and function m(t) is used to represent the parameter value of the steering wheel angle parameter of the vehicle that changes with the prediction moment.

[0042] In this embodiment, the pre-trained large language model can understand the current driving scenario of the vehicle based on the driving-related parameters of the vehicle, and flexibly deduce and output a more appropriate driving strategy for the driving scenario. Then, the pre-trained path planning model generates a suitable driving path according to the appropriate driving strategy, map information and navigation information, and controls the vehicle to drive along the appropriate driving path, thereby improving the driving safety and efficiency of the vehicle.

[0043] Please refer to Figure 2 , Figure 2 A flow chart of a vehicle path planning method provided in another embodiment of the present application. Figure 2 The vehicle path planning method provided in the embodiment of the present application is described in detail. The vehicle path planning method may include the following steps:

[0044] Step S210: Acquire driving-related parameters at the current moment, wherein the driving-related parameters at least include vehicle positioning information, obstacle information, map information, navigation information, and predicted trajectory information of traffic participants around the vehicle.

[0045] In this embodiment, the specific implementation of step S210 can refer to the content of the above-mentioned embodiment, which will not be repeated here.

[0046] Step S220: input the driving-related parameters into a pre-trained large language model to obtain a plurality of interactive texts output by the large language model, wherein the interactive texts are used to characterize the driving strategy of the vehicle after the current moment.

[0047] In this embodiment, after the large language model understands the current driving scenario of the vehicle based on the driving-related parameters of the vehicle, it can deduce a variety of feasible driving strategies for the driving scenario, that is, it can output multiple interactive texts, and different interactive texts represent different driving strategies.

[0048] For example, based on the input driving-related parameters, the large language model understands that the current driving scenario of the vehicle is "unprotected left turn, pedestrian A is walking towards each other on the left sidewalk, and the right car B is driving from right to left relative to the vehicle". Based on this, the large language model can deduce multiple different interactive texts shown below:

[0049] Interactive text 1: The vehicle accelerates to overtake pedestrian A, and then accelerates to overtake vehicle B on the right;

[0050] Interactive text 2: On the premise that the right car B overtakes the pedestrian A first, the self-vehicle accelerates to overtake the pedestrian A and gives way to the right car B;

[0051] Interactive text 3: The vehicle slows down to let pedestrian A pass, then accelerates to overtake vehicle B on the right;

[0052] Interactive text 4: The vehicle slows down to give way to pedestrian A, and also slows down to give way to vehicle B on the right.

[0053] Step S230: input the map information, the navigation information and each of the interactive texts into the path planning model to obtain a planned path corresponding to each of the interactive texts.

[0054] It can be understood that when a large language model outputs multiple interactive texts, the path planning model can be used to perform path planning for each interactive text in turn; that is, the map information, navigation information and each of the interactive texts are input into the path planning model to obtain the planned path output by the path planning model for each interactive text, that is, multiple planned paths corresponding one by one to the multiple interactive texts are obtained.

[0055] Exemplarily, if the multiple interactive texts include the aforementioned interactive text 1, interactive text 2, interactive text 3 and interactive text 4, the path planning model can then output the planned path 1 corresponding to interactive text 1, the planned path 2 corresponding to interactive text 2, the planned path 3 corresponding to interactive text 3, and the planned path 4 corresponding to interactive text 4 in sequence.

[0056] In some embodiments, see Figure 3 , step S230 may include the contents of step S231 to step S233:

[0057] Step S231: input the interactive text, the driving state parameters of the vehicle, the map information and the navigation information into the path generation unit to obtain a plurality of candidate paths output by the path generation unit.

[0058] Specifically, see Figure 4 , the path planning model may include an input communication port, a prediction unit, a data preprocessing unit, a path generation unit, a path selection unit, a path tracking unit, and an output communication port. First, the driving state parameters of the vehicle (i.e., the vehicle state) and the obstacle state can be filtered out only from the perception information, the vehicle positioning information, and the chassis information, and transmitted based on the CAN (Controller Area Network) bus, and transmitted to the prediction unit through the input communication port. Among them, the vehicle state may include the state parameters such as the coordinate position, speed, acceleration, and orientation of the vehicle, and the obstacle state may include the state parameters such as the coordinate position, speed, acceleration, and orientation of the obstacle.

[0059] Furthermore, the prediction unit can predict the driving trajectory of the obstacle within a preset time after the current moment according to the state parameters such as the coordinate position, speed, acceleration and direction of the obstacle, and obtain the predicted trajectory of the obstacle. The preset time is a pre-set time, such as 3 seconds or 5 seconds, etc. Of course, the time value can also be adjusted according to actual needs, and this embodiment does not limit this.

[0060] Then, the interactive text, the vehicle state, and the predicted trajectory of the obstacle are input into the data preprocessing unit for preprocessing. The data preprocessing unit contains a transformer-based network model, which can filter out information with greater relevance and interaction with the vehicle from the predicted trajectories of many obstacles based on the attention mechanism, such as the predicted trajectory of obstacles that are closer to the vehicle. In other words, the data preprocessing unit can further encode the information input into the unit and finally output the encoded information with greater relevance and interaction with the vehicle.

[0061] Furthermore, the driving state parameters of the vehicle preprocessed by the data preprocessing unit are transmitted to the path generation unit, and the path generation unit can generate multiple feasible candidate paths according to the semantics of the interactive text, the driving state parameters of the vehicle, the map information and the navigation information.

[0062] Exemplarily, taking the interactive text "The vehicle first slows down to give way to pedestrian A, and then accelerates to overtake vehicle B on the right" as an example, candidate path 1 can be generated: "The vehicle slows down within the current lane line to give way to pedestrian A, and accelerates from the adjacent lane on the left of vehicle B to overtake vehicle B on the right"; candidate path 2 can also be generated: "The vehicle slows down within the current lane line to give way to pedestrian A, and accelerates from the adjacent lane on the right of vehicle B to overtake vehicle B on the right"; candidate path 3 can also be generated: "The vehicle slows down within the current lane line to give way to pedestrian A, and accelerates from the second lane on the left of vehicle B to overtake vehicle B on the right".

[0063] Step S232: inputting a plurality of candidate paths and the interactive text into the path selection unit, and obtaining the optimal path output by the path selection unit as the planned path.

[0064] Among them, the path selection unit is the Critic module mentioned above, which is an action value function responsible for evaluating the performance of the Actor module under a given state s(State) and guiding the next stage of the Actor module's actions. In layman's terms, the path selection unit can select the candidate path with the highest function value of the action value function from multiple candidate paths based on the current driving state of the vehicle as the optimal path, and use the optimal path as the planned path of the vehicle under the interactive text.

[0065] Step S233: input the planned path, the driving state parameters of the vehicle and the predicted trajectory information of the obstacle into the path tracking unit to obtain the planned path output by the path tracking unit and the control sequence corresponding to the planned path.

[0066] Furthermore, the driving state parameters of the vehicle, the predicted trajectory information of the obstacle, and the planned path output by the path selection unit are input into the path tracking unit, and the planned path output by the path tracking unit and the control sequence corresponding to the planned path can be obtained. Among them, the path tracking unit is the Actor module mentioned above, which outputs the action space of the intelligent agent (i.e., the vehicle). The expression formulas of the planned path and the control sequence can refer to the contents in the above embodiments, which will not be repeated here.

[0067] Step S240: using a preset scoring strategy to score each of the planned paths, and obtaining a path score for each of the planned paths.

[0068] It can be understood that since the path planning model outputs multiple different planned paths, each planned path can be scored using a preset scoring strategy to select the best planned path for the vehicle to travel based on the path score of each planned path.

[0069] In some embodiments, see Figure 5 , step S240 may include the contents of step S241 to step S242:

[0070] Step S241: Calculate the index score of each of the planned paths using a plurality of index calculation formulas to obtain a plurality of index scores for each of the planned paths, wherein the plurality of index calculation formulas correspond one-to-one to a plurality of preset performance evaluation indicators, and the plurality of preset performance evaluation indicators include at least a safety indicator, a comfort indicator, a communication efficiency indicator, and a compliance indicator.

[0071] Optionally, the indicator calculation formula corresponding to the security indicator can be expressed as follows:

[0072] I safety =Ave(PODAR)

[0073]

[0074] w(·)=w D (d) or w T (t)

[0075] w(0)=1

[0076]

[0077] Among them, Isafety represents the safety score under the safety index, t is the prediction time, T is the prediction time domain [s], n is the weekly vehicle number of the vehicles around the vehicle, N is the total number of weekly vehicles around the vehicle, To predict the risk value caused by the nth weekly vehicle at time t, is the potential collision damage with the nth vehicle at time t, ω T is the time reduction factor, ω D is the space reduction coefficient, d is the contour distance between the two workshops [m], which can be calculated using the circle method.

[0078] Optionally, the index calculation formula corresponding to the comfort index can be expressed as follows:

[0079]

[0080] Among them, I comfort Represents the comfort score under the comfort index, a includes instantaneous lateral acceleration and longitudinal acceleration, the unit is m / s 2 , ω is the instantaneous yaw rate, in rad / s, jerk is the longitudinal acceleration, in m / s 3 , T is the statistical time step interval, which is generally set to 20. Of course, the value can also be adjusted according to actual needs, and this embodiment does not limit this.

[0081] Optionally, the indicator calculation formula corresponding to the communication efficiency indicator can be expressed as follows:

[0082]

[0083] Among them, I traffic represents the traffic efficiency score under the traffic efficiency index, v is the driving speed of the vehicle, is the average speed of the vehicles around the vehicle. That is to say, when the vehicle is traveling along the planned path, the higher the traffic efficiency, the higher the traffic efficiency score of the planned path.

[0084] Optionally, the indicator calculation formula corresponding to the compliance indicator can be expressed as follows:

[0085] I lawful =score-x

[0086] Among them, l lawfulrepresents the legality score under the legality index, score represents the initial score of the vehicle when it does not violate any traffic rules, and the initial score can be set to 0 points; x represents the deduction points when the vehicle violates traffic rules. It can be understood that the deduction points for violating different traffic rules can be different or the same, and this embodiment does not limit this; it should be noted that the value of x is a positive number. That is, when the vehicle is driving along the planned path, the more traffic rules it violates, the higher the legality score I. lawul The lower.

[0087] Step S242: performing weighted summation on the multiple indicator scores of each planned path according to the weight parameters corresponding to each indicator score, and obtaining the comprehensive indicator score of each planned path as the path score of each planned path.

[0088] Specifically, the calculation formula of the comprehensive score of the indicator can be expressed as follows:

[0089] I total =α1·I safety +α2·I comfort +α3·I traffic +α4·I lawful

[0090] Among them, I total Characterizes the comprehensive score of the indicators of the planned path, α1 characterizes the weight parameter corresponding to the safety indicator, α2 characterizes the weight parameter corresponding to the comfort indicator, α3 characterizes the weight parameter corresponding to the traffic efficiency indicator, and α4 characterizes the weight parameter corresponding to the compliance indicator. It should be noted that the parameter values ​​of α1, α2, α3 and α4 are pre-set values, and the parameter values ​​of different weight parameters can be set to be the same or different, which is not limited in this embodiment. Optionally, considering that the most important thing for the vehicle during driving is driving safety, the second most important thing is the compliance of driving behavior, and the third most important thing is comfort and traffic efficiency during driving, therefore, α1>α4 can be set. Of course, the parameter values ​​of the weight parameter α3 corresponding to the traffic efficiency index and the weight parameter α2 corresponding to the comfort index can also be adjusted according to the current intelligent driving mode of the vehicle; optionally, if the current intelligent driving mode is a comfortable intelligent driving mode, the aforementioned weight parameters can be set to α1>α4>α2>α3; optionally, if the current intelligent driving mode is an efficient intelligent driving mode, the aforementioned weight parameters can be set to α1>α4>α3>α2.

[0091] Step S250: determining a planned path with the highest path score from the multiple planned paths as the target planned path.

[0092] It can be understood that the higher the path score of the planned path, the higher the safety, comfort and traffic efficiency of the vehicle when driving along the planned path, and the lower the probability of violating traffic rules. Therefore, after obtaining the path score of each planned path among multiple planned paths, the planned path with the highest path score is determined as the target planned path; the target planned path can also be understood as the optimal planned path among multiple planned paths.

[0093] Step S260: Control the vehicle to travel according to the target planned path.

[0094] Similarly, the target control sequence corresponding to the target planning path includes at least a target acceleration parameter and a target steering wheel parameter. Based on this, the vehicle can be controlled to travel along the target planning path according to the target acceleration parameter and the target steering wheel parameter.

[0095] In this embodiment, a pre-trained large language model is used to understand the current driving scene of the vehicle based on the driving-related parameters of the vehicle, and based on the understood driving scene, multiple feasible driving strategies are deduced more quickly and comprehensively (i.e., multiple interactive texts are output); all driving strategies are deduced faster and more comprehensively than human drivers, and the interpretability and interactive performance are greatly improved. Then, the pre-trained path planning model corresponds to multiple feasible driving paths according to multiple feasible driving strategies, map information, and navigation information, and then the driving path with the highest path score is selected as the optimal driving path, and the vehicle is controlled to drive according to the optimal driving path, thereby effectively improving the safety, traffic efficiency, compliance, and comfort of the vehicle during driving. In addition, the interactive text containing the driving strategy is generated using the large language model. Compared with directly outputting the trajectory / control sequence or control amount, the interactive text of the driving strategy considers the driving possibility of a spatiotemporal segment, and can obtain a feasible solution with a larger area and richer interactive methods, and then the corresponding trajectory / control sequence is generated according to the interactive text containing the driving strategy. In other words, the present application simulates the process of interaction of implicit semantic decisions made by humans when driving, so as to flexibly realize the deduction of driving strategies and realize the deduction of driving strategies separated in time and space.

[0096] Please refer to Figure 6 , Figure 6 A flow chart of a model training method provided in another embodiment of the present application. Figure 6 The model training method provided in the embodiment of the present application is described in detail. The model training method may include the following steps:

[0097] Step S301: Obtain a second sample set, where the second sample set includes multiple second sample data, each of the second sample data includes driving-related parameters of the vehicle at a historical moment, and each of the second sample data carries a preset label, and the preset label includes a preset interactive text, and the preset interactive text is used to characterize a preset driving strategy of the vehicle after the historical moment.

[0098] The second sample set may be established based on the vehicle driving data set, and each second sample data includes the driving-related parameters of the vehicle at a historical moment. The specific parameter information contained in the driving-related parameters is the same as the information type contained in the driving-related parameters in the aforementioned embodiment. Please refer to the content in the aforementioned embodiment and will not be repeated here. In addition, each second sample data carries a preset label, and the preset label includes a preset interactive text for characterizing the preset driving strategy of the vehicle after the historical moment.

[0099] Step S302: input each of the second sample data into the initial language model to obtain a second predicted interactive text output by the initial language model.

[0100] Step S303: based on the difference between the second predicted interaction text and the preset interaction text, iteratively update the model parameters of the initial language model until a second preset condition is met, thereby obtaining the pre-trained initial language model.

[0101] Furthermore, each second sample data is input into the initial language model to obtain the second predicted interaction text output by the initial language model for each second sample data; and the second loss value is determined based on the difference between the second predicted interaction text corresponding to each second sample data and the preset interaction text carried by each second sample data; and based on the second loss value, the model parameters of the initial language model are iteratively updated until the second preset condition is met to obtain the pre-trained initial language model. Optionally, the initial language model can be a Transformer model, and the network architecture of the initial language model is as follows: Figure 7 As shown, the Transformer model includes an encoder module and a decoder module. The encoder module includes Figure 7 The multi-head attention layer, residual & normalization layer, feed-forward network layer, and residual & normalization layer in the decoder module include Figure 7 The multi-head attention layer, residual & normalization layer, multi-head attention layer, residual & normalization layer, feed-forward network layer, and residual & normalization layer in .

[0102] Among them, the second preset condition can be: the second loss value is less than the preset value, the second loss value no longer changes, or the number of training times reaches the preset number, etc. It can be understood that after the initial language model is iteratively trained for multiple training cycles according to the second sample set, wherein each training cycle includes multiple iterative trainings, the parameters in the initial language model are continuously optimized, so that the above-mentioned second loss value becomes smaller and smaller, and finally becomes a fixed value, or is less than the above preset value. At this time, it means that the initial language model has converged; of course, it can also be determined that the initial language model has converged after the number of training times reaches the preset number. At this time, the initial language model can be used as the initial language model after pre-training. Among them, the preset value and the preset number of times are pre-set, and their values ​​can also be adjusted according to different application scenarios. This embodiment does not limit this.

[0103] That is to say, a large amount of second sample data is used to train the initial language model to have the ability to understand the vehicle's driving scenarios and the ability to plan driving strategies for the understood driving scenarios based on the vehicle's driving related parameters; in this way, the pre-trained initial language model will initially have the ability to understand the vehicle's driving scenarios and the ability to plan driving strategies for the understood driving scenarios.

[0104] Step S304: Acquire a first sample set, where the first sample set includes a plurality of first sample data, each of which includes driving-related parameters of the vehicle at a historical moment and a historical driving path after the historical moment.

[0105] Further, after completing the pre-training of the initial language model, a first sample set can be obtained, wherein each first sample data in the first sample set includes the driving-related parameters of the vehicle at the historical moment and the historical driving path after the historical moment. It should be noted that the driving behavior of the vehicle according to the historical driving path in each first sample data is reasonable and safe.

[0106] Step S305: input each of the first sample data into a pre-trained initial language model to obtain a first predicted interactive text output by the pre-trained initial language model.

[0107] Step S306: input the first predicted interaction text into an initial planning model to obtain a first planning path output by the initial planning model.

[0108] In this way, by using the pre-trained initial language model and based on each first sample data, the appropriate driving strategy under each first sample data can be output, that is, the appropriate first predicted interaction text can be output. The first predicted interaction text is the driving strategy of the vehicle after the historical moment predicted by the initial language model; and the first predicted interaction text is input into the initial planning model to obtain the first planning path output by the initial planning model for the first predicted interaction text.

[0109] Step S307: Based on the degree of difference between the first planned path and the historical driving path, the model parameters of the pre-trained initial language model and the initial planning model are iteratively updated until the first preset condition is met, and the updated initial language model is obtained as the large language model, and the updated initial planning model is obtained as the path planning model.

[0110] Specifically, based on the difference between the first planned path and the historical driving path, a first loss value is determined; and based on the first loss value, the model parameters of the pre-trained initial language model and the initial planning model are iteratively updated until the first preset condition is met, and the updated initial language model is obtained as the aforementioned large language model, and the updated initial planning model is obtained as the aforementioned path planning model. The model architecture of the initial language model can be found in Figure 4 The model architecture shown will not be described in detail here.

[0111] Among them, the first preset condition can be: the first loss value is less than the preset value, the first loss value no longer changes, or the number of training times reaches the preset number of times, etc. It can be understood that after the pre-trained initial language model and the initial planning model are iteratively trained for multiple training cycles according to the first sample set, wherein each training cycle includes multiple iterative trainings, the parameters in the pre-trained initial language model and the initial planning model are continuously optimized, so that the above-mentioned first loss value becomes smaller and smaller, and finally becomes a fixed value, or is less than the above preset value. At this time, it means that the pre-trained initial language model and the initial planning model have converged; of course, it can also be determined that the pre-trained initial language model and the initial planning model have converged after the training times reach the preset number of times. At this time, the converged pre-trained initial language model can be used as a large language model, and the converged initial planning model can be used as a path planning model. Among them, the preset value and the preset number of times are both pre-set, and their values ​​can also be adjusted according to different application scenarios. This embodiment does not limit this.

[0112] That is to say, by using the labeled second sample data and pre-training the initial language model through supervised learning, and then using the unlabeled first sample data to simultaneously train the pre-trained initial language model and the initial planning model through self-supervised learning, the labeling workload can be greatly reduced, and the training of the large language model and the path planning model can be completed.

[0113] It should be noted that the model training process of this embodiment is performed on a computer device. After completing the training of the large language model and the path planning model, the large language model and the path planning model can be deployed in the vehicle. The vehicle can generate the optimal planning path in the manner of the aforementioned embodiment during the intelligent driving process and drive according to the optimal planning path. No further details are given here. Optionally, the computer device can be an electronic terminal with data processing function, which includes but is not limited to smart phones, laptops and desktop computers; of course, the computer device can also be a server, which can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, network acceleration services (Content Delivery Network, CDN), and basic cloud computing services such as big data and artificial intelligence platforms.

[0114] In this embodiment, after pre-training the initial language model by supervised learning using the labeled second sample data, the pre-trained initial language model and the initial planning model are simultaneously trained by self-supervised learning using the unlabeled first sample data, thereby greatly reducing the workload of labeling the training sample data, and realizing the training of the large language model and the path planning model.

[0115] Please refer to Figure 8 , which shows a structural block diagram of a vehicle path planning device 400 provided in an embodiment of the present application. The device 400 may include: a parameter acquisition module 410, a strategy generation module 420, a path planning module 430 and a control module 440.

[0116] The parameter acquisition module 410 is used to obtain the driving-related parameters of the vehicle at the current moment, and the driving-related parameters at least include the vehicle positioning information, obstacle information, map information, navigation information, and predicted trajectory information of traffic participants around the vehicle.

[0117] The strategy generation module 420 is used to input the driving related parameters into a pre-trained large language model to obtain an interactive text output by the large language model, and the interactive text is used to characterize the driving strategy of the vehicle after the current moment.

[0118] The path planning module 430 is used to input the map information, the navigation information and the interactive text into a pre-trained path planning model to obtain a planned path output by the path planning model.

[0119] The control module 440 is used to control the vehicle to travel along the planned path.

[0120] In some embodiments, the path planning model includes at least a path generation unit, a path selection unit, and a path tracking unit. The path planning module 430 can be specifically used to: input the interactive text, the driving state parameters of the vehicle, the map information, and the navigation information into the path generation unit to obtain multiple candidate paths output by the path generation unit; input the multiple candidate paths and the interactive text into the path selection unit to obtain the optimal path output by the path selection unit as the planned path; input the planned path, the driving state parameters of the vehicle, and the predicted trajectory information of the obstacles into the path tracking unit to obtain the planned path output by the path tracking unit and the control sequence corresponding to the planned path.

[0121] In this manner, the control sequence at least includes the acceleration parameter and the steering wheel angle parameter of the vehicle. The control module 440 may be specifically configured to: based on the acceleration parameter and the steering wheel angle parameter, control the vehicle to travel along the planned path.

[0122] In some embodiments, the number of the interactive texts is multiple, the number of the planned paths is multiple, and the multiple planned paths correspond one-to-one to the multiple interactive texts. The vehicle path planning device 400 may also include: a scoring module and a target path determination module. Among them, the scoring module can be used to score each of the planned paths using a preset scoring strategy before controlling the vehicle to travel along the planned path to obtain a path score for each of the planned paths. The target path determination module can be used to determine the planned path with the highest path score from the multiple planned paths as the target planned path. The control module 440 can be used to control the vehicle to travel along the target planned path.

[0123] In this manner, the scoring module can be specifically used to: calculate the indicator score of each of the planned paths using a plurality of indicator calculation formulas to obtain a plurality of indicator scores for each of the planned paths, wherein the plurality of indicator calculation formulas correspond one-to-one to a plurality of preset performance evaluation indicators, and the plurality of preset performance evaluation indicators include at least a safety indicator, a comfort indicator, a communication efficiency indicator, and a compliance indicator; perform weighted summation of the plurality of indicator scores of each of the planned paths according to the weight parameters corresponding to each indicator score to obtain a comprehensive indicator score for each of the planned paths as the path score of each of the planned paths.

[0124] In some embodiments, the vehicle path planning device 400 may further include: a language model training module. The language model training module may be used to: obtain a first sample set, the first sample set includes a plurality of first sample data, each of the first sample data includes the driving-related parameters of the vehicle at a historical moment and the historical driving path after the historical moment; input each of the first sample data into the pre-trained initial language model to obtain the first predicted interactive text output by the pre-trained initial language model; input the first predicted interactive text into the initial planning model to obtain the first planned path output by the initial planning model; based on the degree of difference between the first planned path and the historical driving path, iteratively update the model parameters of the pre-trained initial language model and the initial planning model until the first preset condition is met, and obtain the updated initial language model as the large language model, and obtain the updated initial planning model as the path planning model.

[0125] In some embodiments, the vehicle path planning device 400 may further include: a path planning model training module. The path planning model training module may be used to: obtain a second sample set, the second sample set includes a plurality of second sample data, each of the second sample data includes driving-related parameters of the vehicle at a historical moment, each of the second sample data carries a preset label, the preset label includes a preset interactive text, and the preset interactive text is used to characterize the preset driving strategy of the vehicle after the historical moment; input each of the second sample data into the initial language model to obtain a second predicted interactive text output by the initial language model; based on the degree of difference between the second predicted interactive text and the preset interactive text, iteratively update the model parameters of the initial language model until the second preset condition is met to obtain the pre-trained initial language model.

[0126] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and modules can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here.

[0127] In several embodiments provided in the present application, the coupling between modules may be electrical, mechanical or other forms of coupling.

[0128] In addition, each functional module in each embodiment of the present application can be integrated into a processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The above integrated modules can be implemented in the form of hardware or software functional modules.

[0129] To summarize, the present application utilizes a pre-trained large language model to understand the current driving scenario of the vehicle based on the driving-related parameters of the vehicle, and based on the understood driving scenario, more quickly and comprehensively deduce a variety of feasible driving strategies (i.e., output multiple interactive texts); then the pre-trained path planning model corresponds to a plurality of feasible driving paths according to a variety of feasible driving strategies, map information, and navigation information, and then selects the driving path with the highest path score as the optimal driving path, and controls the vehicle to drive along the optimal driving path, thereby effectively improving the safety, traffic efficiency, compliance, and comfort of the vehicle during driving.

[0130] The following will be combined Fig. 9 A vehicle provided in the present application is described.

[0131] Reference Fig. 9 , Fig. 9 A structural block diagram of a vehicle 500 provided in an embodiment of the present application is shown, and the above method provided in the embodiment of the present application can be executed by the vehicle 500.

[0132] The vehicle 500 in the embodiment of the present application may include one or more of the following components: a processor 501, a memory 502, and one or more applications, wherein the one or more applications may be stored in the memory 502 and configured to be executed by one or more processors 501, and the one or more programs are configured to execute the method as described in the aforementioned method embodiment.

[0133] The processor 501 may include one or more processing cores. The processor 501 uses various interfaces and lines to connect various parts of the entire vehicle 500, and executes various functions and processes data of the vehicle 500 by running or executing instructions, programs, code sets or instruction sets stored in the memory 502, and calling data stored in the memory 502. Optionally, the processor 501 can be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), and programmable logic array (PLA). The processor 501 can integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. Among them, the CPU mainly processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing display content; and the modem is used to process wireless communications. It is understandable that the above-mentioned modem can also be integrated into the processor 501 and implemented separately through a communication chip.

[0134] The memory 502 may include a random access memory (RAM) or a read-only memory (ROM). The memory 502 may be used to store instructions, programs, codes, code sets or instruction sets. The memory 502 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for implementing at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the following various method embodiments, etc. The data storage area may also store data created by the vehicle 500 during use (such as the various corresponding relationships described above), etc.

[0135] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and modules can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here.

[0136] In several embodiments provided in the present application, the coupling or direct coupling or communication connection between the modules shown or discussed may be an indirect coupling or communication connection through some interfaces, devices or modules, which may be electrical, mechanical or other forms.

[0137] In addition, each functional module in each embodiment of the present application can be integrated into a processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The above integrated modules can be implemented in the form of hardware or software functional modules.

[0138] Please refer to Fig.10 , which shows a structural block diagram of a computer-readable storage medium provided in an embodiment of the present application. The computer-readable medium 600 stores program codes, which can be called by a processor to execute the method described in the above method embodiment.

[0139] The computer readable storage medium 600 may be an electronic memory such as a flash memory, an EEPROM (electrically erasable programmable read-only memory), an EPROM, a hard disk, or a ROM. Optionally, the computer readable storage medium 600 includes a non-transitory computer-readable storage medium. The computer readable storage medium 600 has storage space for program code 610 that performs any method step of the above method. These program codes can be read from or written to one or more computer program products. The program code 610 can be compressed, for example, in an appropriate form.

[0140] In some embodiments, a computer program product or computer program is provided, the computer program product or computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the electronic device performs the steps in the above-mentioned method embodiments.

[0141] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A vehicle path planning method, characterized in that: The method comprises Acquire the driving-related parameters of the vehicle at the current moment, wherein the driving-related parameters include at least the vehicle positioning information, obstacle information, map information, navigation information, and predicted trajectory information of traffic participants around the vehicle; Inputting the driving related parameters into a pre-trained large language model to obtain an interactive text output by the large language model, wherein the interactive text is used to characterize the driving strategy of the vehicle after the current moment; Inputting the map information, the navigation information and the interactive text into a pre-trained path planning model to obtain a planned path output by the path planning model; Control the vehicle to travel along the planned path.

2. The method according to claim 1, characterized in that The path planning model at least includes a path generation unit, a path selection unit and a path tracking unit; The step of inputting the map information, the navigation information and the interactive text into a pre-trained path planning model to obtain a planned path output by the path planning model includes: Inputting the interactive text, the driving state parameters of the vehicle, the map information and the navigation information into the path generation unit to obtain a plurality of candidate paths output by the path generation unit; Inputting the plurality of candidate paths and the interactive text into the path selection unit, and obtaining the optimal path output by the path selection unit as the planned path; The planned path, the driving state parameters of the vehicle and the predicted trajectory information of the obstacle are input into a path tracking unit to obtain the planned path output by the path tracking unit and the control sequence corresponding to the planned path.

3. The method according to claim 2, characterized in that The control sequence at least includes an acceleration parameter and a steering wheel angle parameter of the vehicle; The controlling the vehicle to travel according to the planned path includes: Based on the acceleration parameter and the steering wheel angle parameter, the vehicle is controlled to travel along the planned path.

4. The method according to claim 1, characterized in that: There are multiple interactive texts, and there are multiple planned paths, and the multiple planned paths correspond one-to-one to the multiple interactive texts; Before controlling the vehicle to travel according to the planned path, the method further includes: Using a preset scoring strategy, scoring each of the planned paths to obtain a path score for each of the planned paths; Determine a planned path with the highest path score from the multiple planned paths as the target planned path; The controlling the vehicle to travel according to the planned path includes: Control the vehicle to travel along the target planned path.

5. The method according to claim 4, characterized in that The method of scoring each of the planned paths by using a preset scoring strategy to obtain a path score of each of the planned paths includes: Calculating the index score of each of the planned paths using multiple index calculation formulas to obtain multiple index scores of each of the planned paths, wherein the multiple index calculation formulas correspond one-to-one to multiple preset performance evaluation indicators, and the multiple preset performance evaluation indicators include at least a safety indicator, a comfort indicator, a communication efficiency indicator, and a compliance indicator; According to the weight parameters corresponding to each indicator score, the multiple indicator scores of each planned path are weighted and summed to obtain the comprehensive indicator score of each planned path as the path score of each planned path.

6. The method according to any one of claims 1 to 5, characterized in that: The training process of the large language model and the path planning model includes: Acquire a first sample set, the first sample set comprising a plurality of first sample data, each of the first sample data comprising a driving-related parameter of a vehicle at a historical moment and a historical driving path after the historical moment; Inputting each of the first sample data into a pre-trained initial language model to obtain a first predicted interactive text output by the pre-trained initial language model; Inputting the first predicted interactive text into an initial planning model to obtain a first planning path output by the initial planning model; Based on the degree of difference between the first planned path and the historical driving path, the model parameters of the pre-trained initial language model and the initial planning model are iteratively updated until a first preset condition is met, and the updated initial language model is obtained as the large language model, and the updated initial planning model is obtained as the path planning model.

7. The method according to claim 6, characterized in that The pre-training process of the initial language model includes: Acquire a second sample set, the second sample set including a plurality of second sample data, each of the second sample data including a driving-related parameter of a vehicle at a historical moment, each of the second sample data carrying a preset label, the preset label including a preset interactive text, the preset interactive text being used to characterize a preset driving strategy of the vehicle after the historical moment; Input each of the second sample data into the initial language model to obtain a second predicted interactive text output by the initial language model; Based on the degree of difference between the second predicted interaction text and the preset interaction text, the model parameters of the initial language model are iteratively updated until a second preset condition is met, thereby obtaining the pre-trained initial language model.

8. A vehicle path planning device, characterized in that: The device comprises: A parameter acquisition module, used to acquire the driving-related parameters of the vehicle at the current moment, wherein the driving-related parameters at least include the vehicle positioning information, obstacle information, map information, navigation information, and predicted trajectory information of traffic participants around the vehicle; A strategy generation module, used for inputting the driving related parameters into a pre-trained large language model to obtain an interactive text output by the large language model, wherein the interactive text is used for representing the driving strategy of the vehicle after the current moment; A path planning module, used to input the map information, the navigation information and the interactive text into a pre-trained path planning model to obtain a planned path output by the path planning model; The control module is used to control the vehicle to travel along the planned path.

9. A vehicle, characterized in that: The vehicle comprises: one or more processors; Memory; One or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs are configured to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program code, which can be called by a processor to execute the method according to any one of claims 1 to 7.

Images