A method and device for predicting the trajectory of construction vehicles based on physical laws constraints.

Abstract

The application provides a construction vehicle trajectory prediction method and device based on physical law constraints. The method comprises: obtaining known trajectory and state information of a target vehicle, the state information comprising driving behavior information of the target vehicle and / or action information of the environment on the target vehicle; inputting the known trajectory and the state information into a trained trajectory prediction model to obtain a predicted trajectory of the target vehicle output by the trained trajectory prediction model; wherein the trained trajectory prediction model is used to determine and output the predicted trajectory according to the known trajectory and the state information. The prediction model of the application combines the known trajectory and the state information to predict the trajectory of the target vehicle, introduces physical constraints to accurately simulate the actual behavior of the target vehicle, improves the accuracy of the trajectory prediction of the construction vehicle, and ensures that the prediction result conforms to the physical law in the actual traffic scene.

Description

Technical Field

[0001] This application belongs to the field of computer technology, and in particular relates to a method and device for predicting the trajectory of construction vehicles based on physical laws. Background Technology

[0002] Currently, trajectory prediction is being used to improve driving safety and address driving safety issues. Specifically, it involves accurately predicting the future trajectory of a target vehicle to enhance driving safety and enable risk assessment and decision-making for the vehicle itself.

[0003] Trajectory prediction typically uses a construction vehicle trajectory prediction model to predict the trajectory of construction vehicles based on their trajectory data. However, this model relies on a single data-driven approach, resulting in low prediction accuracy. Summary of the Invention

[0004] This application provides a method, device, electronic device, readable storage medium, and computer program product for predicting the trajectory of construction vehicles based on physical law constraints, which can solve the problem of low prediction accuracy.

[0005] In a first aspect, embodiments of this application provide a method for predicting the trajectory of construction vehicles based on physical law constraints, including:

[0006] Acquire the known trajectory and status information of the target vehicle, wherein the status information includes the driving behavior information of the target vehicle and / or the effect of the environment on the target vehicle;

[0007] The known trajectory and the state information are input into the trained trajectory prediction model to obtain the predicted trajectory of the target vehicle output by the trained trajectory prediction model.

[0008] The trained trajectory prediction model is used to determine and output the predicted trajectory based on the known trajectory and the state information.

[0009] Secondly, embodiments of this application provide a construction vehicle trajectory prediction device, comprising:

[0010] The acquisition module is used to acquire the known trajectory and status information of the target vehicle, wherein the status information includes the driving behavior information of the target vehicle and / or the effect information of the environment on the target vehicle;

[0011] The prediction module is used to input the known trajectory and the state information into the trained trajectory prediction model to obtain the predicted trajectory of the target vehicle output by the trained trajectory prediction model.

[0012] The trained trajectory prediction model is used to determine and output the predicted trajectory based on the known trajectory and the state information.

[0013] Thirdly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method as described in any one of the first aspects above.

[0014] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method as described in any one of the first aspects above.

[0015] Fifthly, embodiments of this application provide a computer program product that, when run on an electronic device, causes the electronic device to perform the method described in any one of the first aspects above.

[0016] The beneficial effects of the embodiments in this application compared with the prior art are:

[0017] This application embodiment obtains the known trajectory and state information of the target vehicle, including the target vehicle's driving behavior information and / or the environment's effect on the target vehicle; the known trajectory and state information are input into a trained trajectory prediction model to obtain the predicted trajectory of the target vehicle output by the trained trajectory prediction model; wherein, the trained trajectory prediction model is used to determine and output the predicted trajectory based on the known trajectory and state information, so that the prediction model combines the known trajectory and state information to predict the trajectory of the target vehicle, introduces physical constraints to accurately simulate the actual behavior of the target vehicle, improves the accuracy of construction vehicle trajectory prediction, and ensures that the prediction results conform to the physical laws in actual traffic scenarios.

[0018] It is understood that the beneficial effects of the second to fifth aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic flowchart of the first embodiment of the construction vehicle trajectory prediction method provided in this application;

[0021] Figure 2 This is a schematic diagram of the structure of a trained trajectory prediction model provided in an embodiment of this application;

[0022] Figure 3 This is a schematic diagram of the second process of the construction vehicle trajectory prediction method provided in one embodiment of this application;

[0023] Figure 4 This is a schematic diagram of the structure of a construction vehicle trajectory prediction device provided in one embodiment of this application;

[0024] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0025] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0026] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0027] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0028] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0029] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0030] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0031] In one embodiment, Figure 1 This is a schematic flowchart of the first embodiment of the construction vehicle trajectory prediction method provided in this application. Figure 1 As shown, the method includes:

[0032] S11: Obtain the known trajectory and status information of the target vehicle.

[0033] The state information includes the target vehicle's driving behavior information and / or the environment's influence on the target vehicle. The selection of state information requires in-depth analysis of the physical and behavioral characteristics of the construction vehicle. State information influences the target vehicle's movement. Driving behavior information may include information on optimally reaching the desired location and safe following behavior in actual traffic scenarios. Information on the environment's influence on the target vehicle may include information on collision avoidance caused by obstacles and lane-keeping requirements caused by lane constraints in actual traffic scenarios. State information represents the physical laws governing actual traffic scenarios.

[0034] The trajectory is known to be sequential data, including historical location points over a period of time.

[0035] S12: Input the known trajectory and state information into the trained trajectory prediction model to obtain the predicted trajectory of the target vehicle output by the trained trajectory prediction model.

[0036] The trained trajectory prediction model is used to determine and output the predicted trajectory based on the known trajectory and state information.

[0037] The expression for the model's input data is X = [V track V physics ], V track Given the trajectory, V physics This is status information.

[0038] In the application, the trained trajectory prediction model processes historical location points to learn the past movement patterns of the target vehicle; the trained trajectory prediction model processes state information to learn the physical laws of the actual traffic scenario. By learning past movement patterns and physical laws, the future trajectory of the target vehicle is predicted.

[0039] Understandably, existing prediction models use a data-driven approach to predict the trajectories of construction vehicles. This method lacks a physical explanation of the underlying mechanisms, leading to a lack of trust in the decision-making results. This embodiment processes state information using a trained trajectory prediction model, learns the physical laws of real-world traffic scenarios, provides the physical meaning of the data, enhances the interpretability of the model, and improves the trustworthiness of the decision-making results.

[0040] The model in this embodiment combines data-driven and physics-driven approaches, enabling it to accurately predict the trajectory of target vehicles in complex and ever-changing traffic scenarios, while reducing the model's over-reliance on sample size and feature generalization in the data-driven approach.

[0041] Furthermore, the known trajectory is sequence data, including historical location points over a period of time, which enables the model to learn the past motion patterns of the target vehicle and improve the model's performance.

[0042] This embodiment acquires the known trajectory and state information of the target vehicle, including the target vehicle's driving behavior information and / or the environment's effect on the target vehicle; the known trajectory and state information are input into a trained trajectory prediction model to obtain the predicted trajectory of the target vehicle output by the trained trajectory prediction model; wherein, the trained trajectory prediction model is used to determine and output the predicted trajectory based on the known trajectory and state information, so that the prediction model combines the known trajectory and state information to predict the trajectory of the target vehicle, introduces physical constraints to accurately simulate the actual behavior of the target vehicle, improves the accuracy of construction vehicle trajectory prediction, and ensures that the prediction results conform to the physical laws in the actual traffic scene.

[0043] In one embodiment, Figure 2 This is a schematic diagram of the structure of a trained trajectory prediction model provided in an embodiment of this application. For example... Figure 2 As shown, the trained trajectory prediction model includes a trained trajectory encoding module, a trained social force rule module, a trained cascade module, and a trained trajectory decoding module.

[0044] In one possible implementation, the trained trajectory prediction model can be SF-GRU. SF-GRU is based on the GRU (Gated Recurrent Unit) structure and incorporates social force rules. The trajectory encoding module is a GRU encoder, and the trajectory decoding module is a GRU decoder.

[0045] Step S12 includes:

[0046] S121: Input the known trajectory into the trained trajectory encoding module, encode the known trajectory through the trained trajectory encoding module to obtain the final hidden state; and input the final hidden state into the trained cascade module;

[0047] The expression for the known trajectory is V. track =[X i,t-T+1 ,…,X i,j ,…,X i,t The known trajectory includes T location coordinates, X... i,j =(x i,j ,y i,j Let be the x and y components of the target vehicle i in time j, and T be the memory time range of the known trajectory. track ∈R C×M×T C represents the number of features, which is determined by the x and y coordinates and has a value of 2. M represents the number of trajectories, and T represents the length of the trajectory. The coordinates can be Cartesian coordinates or other coordinate systems.

[0048] In one possible implementation, a reset gate in the trained GRU encoder is used to determine the amount of previous state information to be forgotten. The formula for the reset gate is: r t =σ(W r V track +U r h t-1 The update gate in the trained GRU encoder is responsible for determining the amount of new state information to be introduced into the current state. The formula for the update gate is: z t =σ(W z V track +U z h t-1 Here, σ is the sigmoid activation function, and its output is a value between [0,1], enabling fine-grained control over information flow. Based on the reset gate and update gate, the formula for calculating the hidden state at time t (1≤t≤T) is h. t =(1-z) t )h t-1 +z t h~ t h~ t =tanh(WV track +U(r t ⊙h t-1 )), h t-1 h is the previously stored content. t The candidate storage content is represented by W and U, which are the weight matrices learned by the model. After calculating the sequence data, the final hidden state is obtained, which represents the information of the entire sequence data.

[0049] By using a GRU encoder to process the data, the potential information and features of the known trajectory of the target vehicle can be fully mined to accurately process the known trajectory. This can effectively solve problems such as gradient messages and improve the stability and accuracy of the model when processing time series data.

[0050] S122: Input the state information into the trained social force rule module, and determine the social force acting on the target vehicle based on the state information through the trained social force rule module; and input the social force into the trained cascade module.

[0051] Among them, social force rules characterize the physical rules in actual traffic scenarios. Social force rules are determined based on actual traffic scenarios, describing the physical meaning implied by state information and the interactive behavior of target vehicles (physical interactions between construction vehicles and interactions between construction vehicles and the environment). Specifically, these may include driving force rules for the desired location at the next time step, attraction rules, repulsion rules, boundary force rules, etc., based on the surrounding environment.

[0052] In application, the model uses the trained social force rule module to supplement and enhance the state information, better simulate the actual situation of the target vehicle on the road, and more accurately and in detail describe the social forces experienced by the target vehicle on the road.

[0053] In one possible implementation, in a real-world traffic scenario, the target vehicle will proceed to the desired location in an optimal manner while maintaining driving safety. The attraction rule for the desired location at the next time step is that, without external interference, the target vehicle will proceed along the expected path at a set speed. During this process, the target vehicle will experience a certain response time, adjusting through acceleration and deceleration to achieve a smooth transition from the current speed to the desired speed, and the target vehicle's acceleration will not exceed its maximum capacity limit. Under this rule, the target vehicle will be subject to a driving force, which propels the target vehicle towards the desired location. Correspondingly, social forces include the driving force.

[0054] Correspondingly, the driving behavior information includes the target vehicle's maximum acceleration, desired speed, direction of desired speed, initial speed, and response time, where the response time is the time required for the initial speed to change to the desired speed.

[0055] Step S121 includes:

[0056] S21: Input the maximum acceleration, desired velocity, direction of desired velocity, first velocity, and response time into the trained social force rule module. The trained social force rule module determines the driving force acting on the target vehicle based on the maximum acceleration, desired velocity, direction of desired velocity, first velocity, and response time.

[0057] In application, the trained social force rule module utilizes the formula Calculate the driving force.

[0058] Here, 'i' refers to the target vehicle. As the driving force, a1 is the first coefficient of the force, which is set according to the scenario. For maximum acceleration, Let be the expected velocity at time t. Let v be the direction of the expected velocity at time t. i (t) represents the first velocity (actual velocity) at time t, τ i For response time.

[0059] In one possible implementation, in a real-world traffic scenario, when a target vehicle needs to follow the vehicle in front, it will safely follow, exhibiting safe following behavior. The attraction rules of the surrounding environment guide the target vehicle to maintain a safe distance and speed while following the vehicle in front. Under these rules, the target vehicle is attracted, and this attraction drives it to follow the vehicle in front; social forces include attraction.

[0060] Correspondingly, the driving behavior information also includes the target vehicle's first maximum deceleration, the driver's reaction time, the direction of attraction, the second speed of the vehicle in front, the second maximum deceleration of the vehicle in front, and the safe distance between the target vehicle and the vehicle in front. The direction of attraction is determined based on the direction of travel of the vehicle in front, which is the vehicle that the target vehicle is following.

[0061] In one possible implementation, the attraction direction can be the same as the direction of travel of the vehicle in front, which can describe the situation where the target vehicle is affected by the vehicle in front.

[0062] For example, if the vehicle in front is traveling straight, the attraction direction is the same as the direction of its straight-going movement. If the vehicle in front turns right, the attraction direction is the same as the direction of its right turn. When the vehicle in front is preparing to turn right, two driving directions are involved. Before turning right, the target vehicle follows the current driving direction of the vehicle in front, and the attraction direction is the same as its current driving direction. When turning right, the target vehicle follows the vehicle in front, and the attraction direction is the same as the direction of travel after the vehicle in front turns right.

[0063] Step S121 includes:

[0064] S22: Input the first maximum deceleration, driver's reaction time, first speed, attraction direction, second speed, second maximum deceleration, and safe distance into the trained social force rule module. Based on the first maximum deceleration, first speed, second speed, second maximum deceleration, and safe distance, determine the safe speed at which the target vehicle maintains a distance from the vehicle in front. Determine the attraction force based on the safe speed, reaction time, maximum acceleration, and attraction direction.

[0065] In application, the trained social force rule module utilizes formulas Calculate the safe speed. Where i+1 represents the vehicle in front. For safe speed, g i (t) represents the safe distance at time t, v i+1 (t) represents the second velocity (actual velocity) at time t, b i b is the first maximum deceleration i+1 For the second maximum deceleration, δ i To reflect time, the trained social force rule module utilizes the formula... Calculate attractiveness, where, For attraction, a2 is the second coefficient of force, and the second coefficient is set according to the scenario, n i,i+1 To attract direction.

[0066] In one possible implementation, in a real-world traffic scenario, when a target vehicle encounters an obstacle, it will avoid the obstacle to prevent a collision. The repulsive force rule of the surrounding environment dictates that the target vehicle should move away from nearby obstacles and maintain a safe distance. Under this rule, the target vehicle will experience a repulsive force, which propels the target vehicle away from the obstacle; social forces include this repulsive force.

[0067] The obstacles include nearby pedestrians, other construction vehicles, and static obstacles. Other construction vehicles can be vehicles following in the same lane or adjacent lanes. Multiple obstacles exist around the target vehicle, and multiple sets of action information corresponding to these obstacles are obtained.

[0068] Correspondingly, the action information includes the intensity of the repulsive force between the target vehicle and the obstacle in the environment, the first distance, the range of the repulsive force, and the first unit vector, which is the unit vector from the center of the obstacle to the center of the target vehicle.

[0069] The effective range of the repulsive force controls the rate at which the repulsive force decays with distance, defining its effective range. Generally, a larger effective range means the repulsive force can act at a greater distance; a smaller effective range means the repulsive force is effective at closer distances. When the initial distance is within the effective range of the repulsive force, the repulsive force is significant. When the initial distance approaches or exceeds the effective range of the repulsive force, the repulsive force decreases and approaches zero. The effective range of the repulsive force can be determined based on factors such as the type and size of the construction vehicle, its speed, and road conditions.

[0070] The specific range of the repulsive force can be determined by analyzing the historical trajectory data of construction vehicles. For example, a reasonable value can be fitted by statistically analyzing the interactions of construction vehicles under different conditions. Alternatively, a reasonable value can be predicted using a model.

[0071] The strength of the repulsive force can be determined through a physics-driven theoretical framework and data-driven experimental optimization. For example, the parameters can be fitted and optimized using the theoretical basis of the model's formulas and real data, or the reasonableness of the parameters can be verified through experiments and sensitivity analysis to determine a reasonable value.

[0072] Step S122 includes:

[0073] S23: Input the repulsive force intensity, first distance, repulsive force range, and first unit vector into the trained social force rule module. The trained social force rule module determines the repulsive force acting on the target vehicle based on the repulsive force intensity, first distance, repulsive force range, and first unit vector.

[0074] In application, the trained social force rule module utilizes the formula Calculate the repulsive force. The repulsive force between the target vehicle i and the obstacle j. The strength of the repulsive force. The first distance, The range of the repulsive force. It is the first unit vector.

[0075] In one possible implementation, in real-world traffic scenarios, a target vehicle typically travels near the lane centerline while maintaining a certain distance from the roadside. The boundary force rule of the surrounding environment governs lane-keeping requirements, dictating that the vehicle travels near the lane centerline and maintains a certain distance from the roadside. Under this rule, the target vehicle is subject to boundary forces, which drive it to remain on the lane centerline. These boundary forces are part of the social forces.

[0076] Correspondingly, the action information includes the boundary force strength between the lane boundary and the target vehicle, the second distance, the range of action of the boundary force, and the second unit vector, which is the unit vector from the lane boundary to the center of the target vehicle.

[0077] The second distance can be the vertical distance from the center of the target vehicle to the lane boundary. For example, if the target vehicle is traveling on the lane centerline and the center of the construction vehicle coincides with the lane centerline, the second distance is the vertical distance from the center of the target vehicle to any lane boundary. If the target vehicle is traveling on an irregular lane (curve, narrow section, etc.), the vertical distance from the target vehicle to the two boundaries is not the same and is not a constant value. The vertical distance is determined based on the position of the target vehicle and the selected boundary.

[0078] The strength of boundary forces can be determined through a physics-driven theoretical framework and data-driven experimental optimization. For example, parameters can be fitted and optimized using model formulas and real data, or reasonable values ​​can be determined by verifying the rationality of parameters through experiments and sensitivity analysis.

[0079] The range of action of boundary forces can be adjusted through data-driven optimization and experimental analysis, as well as by combining the needs of specific scenarios. Then, through model optimization, a balance between accuracy and applicability can be achieved in the range of action of boundary forces, thereby improving the actual performance of the model.

[0080] Step S122 includes:

[0081] Step S24: Input the boundary force intensity, second distance, boundary force range, and second unit vector into the trained social force rule module. The trained social force rule module determines the boundary force acting on the target vehicle based on the boundary force intensity, second distance, boundary force range, and second unit vector.

[0082] In application, the trained social force rule module utilizes the formula Calculate the boundary forces. For boundary forces, A ib For the boundary force intensity, d ib For the second distance, B ib n represents the range of action of the boundary force. ib It is the second unit vector.

[0083] It is understandable that when the target vehicle is subjected to at least one of the aforementioned social forces due to actual circumstances during its operation, the corresponding trained social force rule module will output at least one social force.

[0084] S123: After the final hidden state and social force are input into the trained cascade module, the final hidden state and social force are processed by matrix cascade through the trained cascade module to obtain cascade data, and the cascade data is input into the trained trajectory decoding module.

[0085] In application, the trained cascaded modules are used via the formula O = W tracks O tracks +W SF O SF +b performs matrix concatenation processing on the final hidden state and social forces, where W tracks W SF For learnable parameters, O tracks For the final hidden state, O SF Let b be the social force and 'b' be the bias coefficient. tracks O SF The data is in matrix form.

[0086] S124: After the cascaded data is input into the trained trajectory decoding module, the cascaded data is decoded by the trained trajectory decoding module to generate and output the predicted trajectory.

[0087] In one possible implementation, the trained GRU decoder utilizes the formula Generate predicted trajectories. The ReLU function is the activation function.

[0088] Specifically, predicting trajectories The expression is in, Let H be the predicted position at time j, and H be the predicted time range of the predicted trajectory.

[0089] This embodiment supplements and enhances state information through social force rules, enabling it to handle various social forces experienced by the target vehicle. This allows the model to more accurately simulate the actual behavior of the target vehicle in real and complex traffic environments, improving the accuracy of model predictions and obtaining more precise prediction results. Furthermore, the social force rules allow the model to intuitively and accurately consider interactions in real traffic scenarios, precisely describe changes in relevant variables during the prediction process, enhance model interpretability, provide reasons for predictions, and make the model's prediction results easier to understand and accept, thereby improving the safety of autonomous driving and the efficiency of traffic flow.

[0090] Furthermore, by introducing social force rules to predict trajectories, the real-time issues caused by overly complex deep learning architectures can be reduced, thus balancing the real-time performance and complexity of the model.

[0091] In one embodiment, Figure 3 This is a schematic diagram of the second process of the construction vehicle trajectory prediction method provided in one embodiment of this application. For example... Figure 3 As shown, before obtaining the known trajectory and status information of the target vehicle, the following steps are also included:

[0092] S31: Obtain training samples and labels.

[0093] The training samples include trajectory point samples and status information samples of each construction vehicle sample, and the labels include the actual location points of each construction vehicle sample.

[0094] S32: Using training samples, train the trajectory prediction model until the value of the model's loss function is less than the preset loss value, and obtain the trained trajectory prediction model.

[0095] The model loss function is used to calculate the difference between the output of the trajectory prediction model and the label.

[0096] In application, the model loss function can be the root mean square error function. Since the movement of construction vehicles includes both lateral and longitudinal dimensions, the corresponding model loss function calculates the lateral and longitudinal losses.

[0097] The formula for the model loss function is: in, For the output of the model, x t+H y t+H Where N is the true location point, and N is the total number of trajectory point samples in the training samples.

[0098] This embodiment obtains training samples and labels. The training samples include trajectory point samples and state information samples of each construction vehicle sample. The labels include the real location points of each construction vehicle sample. Using the training samples, the trajectory prediction model is trained until the value of the model loss function is less than the preset loss value. The trained trajectory prediction model is obtained, and a model that can perform prediction tasks and accurately predict the trajectory of construction vehicles is obtained.

[0099] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. Furthermore, the data collection in the above embodiments is compliant, and its use or implementation does not involve any infringement upon public interests.

[0100] For ease of explanation, only the parts related to the embodiments of this application are shown in the methods described in the above embodiments.

[0101] In one embodiment, Figure 4 This is a schematic diagram of the structure of a construction vehicle trajectory prediction device provided in one embodiment of this application. Figure 4 As shown, a construction vehicle trajectory prediction device includes:

[0102] The acquisition module 10 is used to acquire the known trajectory and status information of the target vehicle. The status information includes the driving behavior information of the target vehicle and / or the effect of the environment on the target vehicle.

[0103] The prediction module 11 is used to input the known trajectory and state information into the trained trajectory prediction model to obtain the predicted trajectory of the target vehicle output by the trained trajectory prediction model.

[0104] The trained trajectory prediction model is used to determine and output the predicted trajectory based on the known trajectory and state information.

[0105] In one embodiment, the prediction module is specifically configured to: input a known trajectory into a trained trajectory encoding module to encode the known trajectory and obtain a final hidden state; input the final hidden state into a trained cascade module; input state information into a trained social force rule module to determine the social forces acting on the target vehicle based on the state information; input the social forces into the trained cascade module, where the social force rules represent physical rules in the actual traffic scenario; after the final hidden state and social forces are input into the trained cascade module, the trained cascade module performs matrix cascade processing on the final hidden state and social forces to obtain cascaded data, and input the cascaded data into a trained trajectory decoding module; after the cascaded data is input into the trained trajectory decoding module, the trained trajectory decoding module decodes the cascaded data to generate and output a predicted trajectory. The trained trajectory prediction model includes the trained trajectory encoding module, the trained social force rule module, the trained cascade module, and the trained trajectory decoding module.

[0106] In one embodiment, the prediction module is specifically used to input the maximum acceleration, desired velocity, direction of desired velocity, first velocity, and response time into the trained social force rule module. The trained social force rule module determines the driving force acting on the target vehicle based on the maximum acceleration, desired velocity, direction of desired velocity, first velocity, and response time. The driving force is used to drive the target vehicle to move towards the desired position. The social force includes the driving force.

[0107] In one embodiment, the prediction module is specifically used to input a first maximum deceleration, the driver's reaction time, a indicated first speed, an attraction direction, a second speed, a second maximum deceleration, and a safe distance into a trained social force rule module. The trained social force rules determine a safe speed for the target vehicle to maintain a distance from the vehicle in front based on the first maximum deceleration, the indicated first speed, the second speed, the second maximum deceleration, and the safe distance. The attraction force is determined based on the safe speed, reaction time, maximum acceleration, and attraction direction. The attraction force is used to drive the target vehicle to follow the vehicle in front. Social forces include the attraction force.

[0108] In one embodiment, the prediction module is specifically used to input the repulsive force intensity, the first distance, the repulsive force range, and the first unit vector into the trained social force rule module. The trained social force rule module determines the repulsive force acting on the target vehicle based on the repulsive force intensity, the first distance, the repulsive force range, and the first unit vector. The repulsive force is used to drive the target vehicle away from the obstacle. Social forces include repulsive forces.

[0109] In one embodiment, the prediction module is specifically used to input the boundary force intensity, the second distance, the boundary force range, and the second unit vector into the trained social force rule module. The trained social force rule module determines the boundary force acting on the target vehicle based on the boundary force intensity, the second distance, the boundary force range, and the second unit vector. The boundary force is used to drive the target vehicle to the center line of the lane. The social force includes the boundary force.

[0110] In one embodiment, the device further includes a training module.

[0111] The training module is used to acquire training samples and labels. The training samples include trajectory point samples and state information samples of each construction vehicle sample, and the labels include the real location points of each construction vehicle sample. Using the training samples, the trajectory prediction model is trained until the value of the model loss function is less than the preset loss value, thus obtaining the trained trajectory prediction model. The model loss function is used to calculate the difference between the output of the trajectory prediction model and the label.

[0112] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 5 As shown, the electronic device 2 of this embodiment includes: at least one processor 20 ( Figure 5 (Only one is shown in the diagram), memory 21, and computer program 22 stored in said memory 21 and executable on said at least one processor 20, wherein said processor 20 executes said computer program 22 to implement the steps in any of the above method embodiments.

[0113] The electronic device 2 can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. The electronic device 2 may include, but is not limited to, a processor 20 and a memory 21. Those skilled in the art will understand that... Figure 5 This is merely an example of electronic device 2 and does not constitute a limitation on electronic device 2. It may include more or fewer components than shown in the figure, or combine certain components, or different components. For example, it may also include input / output devices, network access devices, etc.

[0114] The processor 20 can be a Central Processing Unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0115] In some embodiments, the memory 21 may be an internal storage unit of the electronic device 2, such as a hard disk or memory of the electronic device 2. In other embodiments, the memory 21 may be an external storage device of the electronic device 2, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the electronic device 2. Furthermore, the memory 21 may include both internal and external storage units of the electronic device 2. The memory 21 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory 21 can also be used to temporarily store data that has been output or will be output.

[0116] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0117] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0118] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps in the above-described method embodiments.

[0119] This application provides a computer program product that, when run on an electronic device, enables the electronic device to implement the steps described in the various method embodiments above.

[0120] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying the computer program code to a photographing device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some cases, the computer-readable medium cannot be an electrical carrier signal or a telecommunication signal.

[0121] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0122] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0123] In the embodiments provided in this application, it should be understood that the disclosed apparatus / network devices and methods can be implemented in other ways. For example, the apparatus / network device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0124] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0125] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions 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 this application, and should all be included within the protection scope of this application.

Claims

1. A construction vehicle trajectory prediction method based on physical law constraints, characterized by, The method comprises: obtaining known trajectory and state information of a target vehicle, the state information comprising driving behavior information and environmental influence information of the target vehicle; inputting the known trajectory and the state information into a trained trajectory prediction model to obtain a predicted trajectory of the target vehicle output by the trained trajectory prediction model; wherein the trained trajectory prediction model is configured to determine and output the predicted trajectory according to the known trajectory and the state information; the trained trajectory prediction model comprises a trained trajectory encoding module, a trained social force rule module, a trained cascade module and a trained trajectory decoding module; inputting the known trajectory and the state information into a trained trajectory prediction model to obtain a predicted trajectory of the target vehicle output by the trained trajectory prediction model, comprising: inputting the known trajectory into the trained trajectory encoding module to perform encoding processing on the known trajectory by the trained trajectory encoding module to obtain a final hidden state; and inputting the final hidden state into the trained cascade module; inputting the state information into the trained social force rule module to determine a social force acting on the target vehicle according to the state information by the trained social force rule module; and inputting the social force into the trained cascade module, wherein the social force rule represents physical rules in an actual traffic scene; after the final hidden state and the social force are input into the trained cascade module, performing matrix cascade processing on the final hidden state and the social force by the trained cascade module to obtain cascade data, and inputting the cascade data into the trained trajectory decoding module; after the cascade data is input into the trained trajectory decoding module, decoding the cascade data by the trained trajectory decoding module to generate and output the predicted trajectory; the driving behavior information comprises maximum acceleration, expected speed, direction of the expected speed, first speed and response time of the target vehicle, wherein the response time is the time required for the first speed to change to the expected speed; the inputting the state information into the trained social force rule module to determine a social force acting on the target vehicle according to the state information by the trained social force rule module, comprising: inputting the maximum acceleration, the expected speed, the direction of the expected speed, the first speed and the response time into the trained social force rule module to determine a driving force acting on the target vehicle according to the maximum acceleration, the expected speed, the direction of the expected speed, the first speed and the response time by the trained social force rule module, wherein the driving force is used to drive the target vehicle to move towards an expected position, and the social force comprises the driving force.

2. The method of claim 1, wherein, The driving behavior information further comprises a first maximum deceleration of the target vehicle, a reaction time of a driver, an attraction direction, a second speed of a front vehicle, a second maximum deceleration of the front vehicle, and a safety distance between the target vehicle and the front vehicle, the attraction direction being determined according to a driving direction of the front vehicle, the front vehicle being a vehicle followed by the target vehicle; The inputting the state information into the trained social force rule module comprises: The inputting the state information into the trained social force rule module comprises:

3. The method of claim 2, wherein, The action information comprises repulsion force strength, a first distance, a repulsion force action range, and a first unit vector between the target vehicle and an obstacle in an environment, the first unit vector being a unit vector with a center of the obstacle pointing to a center of the target vehicle; The inputting the state information into the trained social force rule module comprises: The inputting the state information into the trained social force rule module comprises:

4. The method of claim 3, wherein, The action information comprises boundary force strength, a second distance, a boundary force action range, and a second unit vector between a lane boundary and the target vehicle, the second unit vector being a unit vector with the lane boundary pointing to a center of the target vehicle; The inputting the state information into the trained social force rule module comprises: inputting the boundary force intensity, the second distance, the boundary force action range, and the second unit vector into the trained social force rule module, and determining, by the trained social force rule module, a boundary force acting on the target vehicle according to the boundary force intensity, the second distance, the boundary force action range, and the second unit vector, the boundary force being used to drive the target vehicle to be located on the lane center line, the social force including the boundary force.

5. The method according to any one of claims 1 to 4, characterized in that, Before the obtaining of the known trajectory and the state information of the target vehicle, the method further includes: obtaining training samples and labels, the training samples including trajectory point samples and state information samples of each construction vehicle sample, and the labels including true position points of each construction vehicle sample; training a trajectory prediction model by using the training samples until a value of a model loss function is less than a preset loss value, to obtain a trained trajectory prediction model; wherein the model loss function is used to calculate a difference value between an output result of the trajectory prediction model and the labels.

6. A construction vehicle trajectory prediction apparatus characterized by comprising: The method includes: a obtaining module configured to obtain known trajectory and state information of a target vehicle, the state information including driving behavior information of the target vehicle and action information of an environment on the target vehicle; a prediction module configured to input the known trajectory and the state information into a trained trajectory prediction model to obtain a predicted trajectory of the target vehicle output by the trained trajectory prediction model; wherein the trained trajectory prediction model is used to determine and output the predicted trajectory according to the known trajectory and the state information, the trained trajectory prediction model including a trained trajectory encoding module, a trained social force rule module, a trained cascade module, and a trained trajectory decoding module, the driving behavior information including maximum acceleration, expected speed, direction of the expected speed, first speed, and response time of the target vehicle, the response time being a time required for the first speed to change to the expected speed; the prediction module is specifically configured to input the known trajectory into the trained trajectory encoding module, perform encoding processing on the known trajectory by the trained trajectory encoding module to obtain a final hidden state, input the final hidden state into the trained cascade module, input the state information into the trained social force rule module to determine, by the trained social force rule module, a social force acting on the target vehicle according to the state information, input the social force into the trained cascade module, the social force rule representing physical rules in an actual traffic scene, after the final hidden state and the social force are input into the trained cascade module, perform matrix cascade processing on the final hidden state and the social force by the trained cascade module to obtain cascade data, and input the cascade data into the trained trajectory decoding module, after the cascade data are input into the trained trajectory decoding module, perform decoding on the cascade data by the trained trajectory decoding module to generate and output the predicted trajectory. inputting the state information into the trained social force rule module, and determining, by the trained social force rule module, a social force acting on the target vehicle according to the state information, including: inputting the maximum acceleration, the desired speed, a direction of the desired speed, the first speed, and the response time into the trained social force rule module, and determining, by the trained social force rule module, a driving force acting on the target vehicle according to the maximum acceleration, the desired speed, the direction of the desired speed, the first speed, and the response time, the driving force being used to drive the target vehicle to move towards the desired position, the social force including the driving force.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor implements the method of any one of claims 1-5 when executing the computer program.

8. A computer-readable storage medium storing a computer program, the computer-readable storage medium comprising: The computer program, when executed by a processor, implements the method of any one of claims 1-5.

Images