Method, apparatus, device and storage medium for vehicle control

Abstract

Embodiments of the present disclosure provide a method, device, equipment and storage medium for vehicle control. The method comprises: determining mutual state prediction of a target vehicle and an obstacle vehicle at a second time based on at least a first decision behavior of the obstacle vehicle at a first time and observation data of the target vehicle at the first time; the observation data comprises mutual state of the target vehicle and the obstacle vehicle at the first time and a second decision behavior of the target vehicle at a third time, the second time is after the first time, and the third time is before the first time; determining a fourth decision behavior of the obstacle vehicle at the second time based on at least the mutual state prediction at the second time and the second decision behavior; determining at least one decision behavior of the target vehicle after the first time based on at least the mutual state prediction at the second time and the fourth decision behavior; and controlling a driving process of the target vehicle after the first time through the at least one decision behavior. Thus, the safety of the target vehicle can be improved.

Description

TECHNICAL FIELD

[0001] Example embodiments of the present disclosure generally relate to the field of computers, and in particular, to a method, apparatus, device and storage medium for vehicle control. BACKGROUND

[0002] In a complex urban road, an autonomous driving vehicle (target vehicle) inevitably interacts with other traffic participants (for example, other vehicles). In the interaction with other traffic participants, the autonomous driving vehicle and the other traffic participants are respectively two independent individuals, and there is no correlation between the two. Therefore, in the control of the autonomous driving vehicle, there is a certain safety hazard. How to make the autonomous driving vehicle travel more safely and improve the safety performance of the vehicle is worth attention. SUMMARY

[0003] In a first aspect of the present disclosure, a method for vehicle control is provided. The method can include: determining a mutual state prediction of a target vehicle and an obstacle vehicle at a second time based at least on a first decision behavior of the obstacle vehicle at a first time and observation data of the target vehicle at the first time, the observation data including mutual states of the target vehicle and the obstacle vehicle at the first time and a second decision behavior of the target vehicle at a third time, the second time being after the first time, and the third time being before the first time; determining a fourth decision behavior of the obstacle vehicle at the second time based at least on the mutual state prediction at the second time and the third decision behavior of the target vehicle at the first time; determining at least one decision behavior of the target vehicle after the first time based at least on the mutual state prediction at the second time and the fourth decision behavior; and controlling a driving process of the target vehicle after the first time through the at least one decision behavior.

[0004] In a second aspect of the present disclosure, an apparatus of vehicle control is provided. The apparatus can include: an inter-state prediction module, which can be configured to determine an inter-state prediction of a target vehicle and an obstacle vehicle at a second time based at least on a first decision behavior of the obstacle vehicle at a first time and observation data of the target vehicle at the first time, the observation data comprising an inter-state of the target vehicle and the obstacle vehicle at the first time and a second decision behavior of the target vehicle at a third time, the second time being after the first time, and the third time being before the first time. An obstacle vehicle decision behavior determination module, which can be configured to determine a fourth decision behavior of the obstacle vehicle at the second time based at least on the inter-state prediction at the second time and the third decision behavior of the target vehicle at the first time. A target vehicle decision behavior determination module, which can be configured to determine at least one decision behavior of the target vehicle after the first time based at least on the inter-state prediction at the second time and the fourth decision behavior. A target vehicle control module, which can be configured to control a driving process of the target vehicle after the first time by the at least one decision behavior.

[0005] In a third aspect of the present disclosure, an electronic device is provided. The device includes at least one processing unit; and at least one memory coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit. The instructions, when executed by the at least one processing unit, cause the electronic device to perform the method of the first aspect.

[0006] In a fourth aspect of the present disclosure, a computer-readable storage medium is provided. The computer-readable storage medium has stored thereon a computer program, which is executable by a processor to implement the method of the first aspect.

[0007] In a fifth aspect of the present disclosure, a computer program product is provided. The computer program product includes computer-executable instructions that, when executed by a processor, implement the method of the first aspect.

[0008] It should be understood that all statements herein made regarding the exemplary embodiments of the present disclosure are intended to encompass both the specific and generic features of the embodiments. Various modifications and changes can be made thereto without departing from the spirit and scope of the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0009] The above and other features, aspects, and advantages of embodiments of the present disclosure will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, in which like reference numerals denote like elements, and wherein:

[0010] Figure 1 A schematic diagram illustrating an example environment in which embodiments of the present disclosure can be implemented is shown;

[0011] Figure 2 A flowchart of a method of vehicle control is shown, in accordance with some embodiments of the present disclosure;

[0012] Figure 3 An example diagram of a mutual state prediction process is shown, in accordance with some embodiments of the present disclosure;

[0013] Figure 4 A schematic block diagram of an apparatus of vehicle control is shown, in accordance with some embodiments of the present disclosure; and

[0014] Figure 5 A block diagram of an electronic device that can implement one or more embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0015] Embodiments of the present disclosure will be described below in greater detail with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it is understood that the present disclosure can be embodied in various forms and should not be interpreted as being limited to the embodiments set forth herein, but rather, these embodiments are provided so that the present disclosure can be more thoroughly and completely understood. It should be understood that the drawings of the present disclosure and the embodiments are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0016] In the description of embodiments of the present disclosure, the term "comprising" and its conjugations should be understood to encompass the meanings of "consisting of" and "consisting essentially of", i.e., "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The term "some embodiments" should be understood as "at least some embodiments". Other explicit and implicit definitions can also be included below.

[0017] In this document, unless explicitly stated otherwise, performing a step "in response to" an event means that the step can be performed immediately in response to the event, but can also include one or more intermediate steps.

[0018] It can be understood that the data involved in the technical solutions of the present disclosure (including but not limited to the data itself, the obtaining, use, storage or deletion of the data) should comply with the requirements of relevant laws and regulations and relevant provisions.

[0019] It can be understood that, before using the technical solutions disclosed in the embodiments of the present disclosure, the type of information involved in the present disclosure, the scope of use, the scenario of use, etc. should be informed to the relevant users and the authorization of the relevant users should be obtained through appropriate means, wherein the relevant users can include any type of right subject, such as individuals, enterprises, groups.

[0020] For example, in response to receiving an active request of a user, a prompt information is sent to the relevant user to explicitly prompt the relevant user that the operation requested to be performed will require obtaining and using information of the relevant user, so that the relevant user can autonomously select whether to provide information to the software or hardware such as an electronic device, an application program, a server or a storage medium, etc. performing the operation of the technical solution of the present disclosure according to the prompt information.

[0021] As an optional but non-limiting implementation manner, in response to receiving an active request of a relevant user, the manner of sending a prompt information to the relevant user may, for example, be a pop-up window manner, and the prompt information may be presented in the form of text in the pop-up window. In addition, the pop-up window may also carry selection controls for the user to select "agree" or "disagree" to provide information to the electronic device.

[0022] It can be understood that the above notification and user authorization obtaining process is only illustrative and does not limit the implementation manner of the present disclosure, and other manners meeting relevant laws and regulations can also be applied to the implementation manner of the present disclosure.

[0023] As used herein, the term "model" can learn an association between respective inputs and outputs from training data, such that after training, a corresponding output can be generated for a given input. The generation of a model can be based on machine learning techniques. Deep learning is a machine learning algorithm that processes inputs and provides corresponding outputs by using multiple layers of processing units. A neural network model is one example of a model based on deep learning. In this document, a "model" can also be referred to as a "machine learning model", a "learning model", a "machine learning network", or a "learning network", which terms are used interchangeably herein.

[0024] Figure 1 A schematic diagram of an environment 100 in which embodiments of the present disclosure can be implemented is shown. The environment 100 can include a target vehicle 110, an obstacle vehicle 120. The target vehicle 110 refers to a vehicle that is of interest and whose behavior is to be controlled, which can be embodied as a vehicle with automatic driving function or assisted driving function. The obstacle vehicle 120 refers to any vehicle that can hinder the normal driving of the ego vehicle or affect the safety of the ego vehicle in passing relative to the target vehicle 110. Although Figure 1 One obstacle vehicle 120 is shown in the figure, but in actual applications, there can be multiple obstacle vehicles, and in such cases, the embodiments of the present disclosure can still be similarly applied.

[0025] The target vehicle 110 can rely on the vehicle control device 130 to automatically drive. The vehicle control device 130 can be a head unit of the target vehicle, or a server device. The server device can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud service, cloud database, cloud computing, cloud function, cloud storage, network service, cloud communication, middleware service, domain name service, security service, content distribution network, and big data and artificial intelligence platform. The server device may, for example, include a computing system / server such as a mainframe, an edge computing node, a computing device in a cloud environment, and the like.

[0026] The vehicle control device 130 determines the decision behavior of the target vehicle 110 in dependence on external data. The external data can include sensing results from the target vehicle 110 and / or external data sources. The target vehicle 110 is usually configured with acquisition devices for capturing elements in the physical environment, sensing devices for perceiving objects in the physical environment, and the like (not shown in the figure). For example, the acquisition devices or sensing devices can include optical cameras that can capture images in the physical environment, lidar that can capture point cloud data, and the like. The external data sources can be embodied as devices for acquiring sensing data in the environment in which the vehicle is located, such as roadside cameras, roadside lidar, and the like.

[0027] It should be understood that the structure and function of the various elements in the environment 100 are described for illustrative purposes only, without implying any limitation on the scope of the present disclosure.

[0028] For the control of an autonomous vehicle, it is generally necessary to linearly predict the motion trend of an obstacle vehicle in the next few seconds using the current speed of the obstacle vehicle to obtain a predicted trajectory of the obstacle vehicle. The predicted trajectory is used to determine whether there is a risk of collision between the target vehicle (ego vehicle) and the obstacle vehicle in the future. Once the predicted trajectory of the obstacle vehicle intersects with the driving trajectory of the target vehicle, the brake control of the target vehicle is triggered.

[0029] In the above control process, the target vehicle and the obstacle vehicle are usually treated as two independent individuals without correlation between each other. However, in actual scenarios, the target vehicle and the obstacle vehicle dynamically change in the interaction process as the states of the two vehicles and environmental disturbances change. Therefore, treating the target vehicle and the obstacle vehicle as two independent individuals cannot reflect the actual interaction process, which may lead to misjudgment of the target vehicle. For example, if the target vehicle adopts a too conservative strategy, it may cause more sudden braking rear-end problems, which may also reduce the efficiency of the target vehicle driving as a whole.

[0030] In embodiments of the present disclosure, an improved scheme for controlling a target vehicle is proposed. The scheme proposes that: based at least on a first decision behavior of an obstacle vehicle at a first time and observation data of the target vehicle at the first time, a mutual state prediction of the target vehicle and the obstacle vehicle at a second time is determined; the observation data comprises mutual states of the target vehicle and the obstacle vehicle at the first time and a second decision behavior of the target vehicle at a third time, the second time being after the first time, and the third time being before the first time. Based at least on the mutual state prediction at the second time and the third decision behavior of the target vehicle at the first time, a fourth decision behavior of the obstacle vehicle at the second time is determined. Based at least on the mutual state prediction at the second time and the fourth decision behavior, at least one decision behavior of the target vehicle after the first time is determined. The target vehicle is controlled in a driving process after the first time through the at least one decision behavior.

[0031] Through the above process, the target vehicle and the obstacle vehicle are taken as a whole. Based on the prediction of the mutual state between the target vehicle and the obstacle vehicle, the future dynamic relationship between the two can be foreseen in advance, so as to obtain a two-vehicle interactive decision stochastic optimal control model. Based on this prediction, the vehicle control device can optimize the control strategy of the target vehicle, so that it can more accurately cope with possible dangerous situations in the subsequent driving process. Thus, through the prediction of the mutual state, the target vehicle can more effectively avoid the obstacle vehicle, thereby improving the overall driving safety and the rationality of the decision.

[0032] Figure 2 An example flow 200 of a method of vehicle control according to some embodiments of the present disclosure is shown. For ease of discussion, the flow 200 will be described with reference to the environment of Figure 1 In the application environment 110, the control of the target vehicle can be performed by the vehicle control device 130.

[0033] At block 201, the vehicle control device 130 determines a mutual state prediction of the target vehicle and the obstacle vehicle at a second time based at least on a first decision behavior of an obstacle vehicle at a first time and observation data of the target vehicle at the first time; the observation data comprises mutual states of the target vehicle and the obstacle vehicle at the first time and a second decision behavior of the target vehicle at a third time, the second time being after the first time, and the third time being before the first time.

[0034] The vehicle control device 130 takes the target vehicle 110 and the obstacle vehicle 120 as a whole. By comprehensively considering the mutual influence between the decision behaviors of the target vehicle 110 and the obstacle vehicle 120, as well as the random disturbance caused by sensor errors, network delays, environmental disturbances and other factors, a control strategy for the target vehicle 110 is determined.

[0035] The first decision behavior of the obstacle vehicle 120 can correspond to a control instruction of the obstacle vehicle 120. For the first decision behavior of the obstacle vehicle 120, the vehicle control device 130 cannot obtain, but needs to determine by using the observation data and a kinematic model of the obstacle vehicle. Taking the current time as the first time as an example, the observation data at the first time can be represented as H k .

[0036] The observation data H k at the first time can include the mutual state of the target vehicle 110 and the obstacle vehicle 120 at the first time. The mutual state at the first time can be represented as x k . The mutual state x k at the first time is determined based on the first state of the target vehicle 110 and the second state of the obstacle vehicle 120, and the mutual state x k can be represented as The vector dimension can be represented as.

[0037] The first state of the target vehicle 110 and the second state of the obstacle vehicle 120 have the same dimension. Taking the first state as an example, the first state may indicate the position, speed, direction, etc. of the target vehicle 110. Correspondingly, the mutual state x k of the target vehicle 110 and the obstacle vehicle 120 at the first time can also indicate the relative position, relative speed, and relative direction, etc. of the target vehicle 110 and the obstacle vehicle 120.

[0038] The observation data H k at the first time can also include the second decision behavior of the target vehicle 110 at the third time, and the third time is a time before the first time. The second decision behavior can be represented as According to the observation data H k at the first time, the second decision behavior of the target vehicle 110 at the third time, and the observation data history H k-1 , the first decision behavior of the obstacle vehicle 120 at the first time can be obtained, and the first decision behavior can be represented as a vector dimension. In summary, the observation data H k at the first time can be represented as

[0039] At least based on the first decision behavior of the obstacle vehicle 120 at the first time and the observation data H k of the target vehicle 110 at the first time, and the decision behavior of the target vehicle 110 at the first time The mutual state prediction at the second time instance can be determined based on the state prediction of the target vehicle 110 and the state prediction of the obstacle vehicle 120. The mutual state prediction at the second time instance can be represented as x k+1 The first decision behavior of the obstacle vehicle 120 at the first time instance The second decision behavior of the obstacle vehicle 120 at the second time instance can be determined in the course of the calculation at the second time instance.

[0040] The mutual state prediction at the second time instance can include the state prediction of the target vehicle 110 and the state prediction of the obstacle vehicle 120, from which the mutual state prediction can be derived based on the state prediction of the target vehicle 110 and the state prediction of the obstacle vehicle 120. The specific prediction process of the mutual state prediction at the second time instance will be described in detail later. Through the above process, the current state of the target vehicle 110 and the obstacle vehicle 120 as a whole jointly affects the next state of the two vehicles.

[0041] At block 202, the vehicle control device 130 determines the fourth decision behavior of the obstacle vehicle 120 at the second time instance based on at least the mutual state prediction at the second time instance and the third decision behavior of the target vehicle at the first time instance.

[0042] The vehicle control device 130 determines the fourth decision behavior of the obstacle vehicle 120 at the second time instance based on the mutual state prediction at the second time instance and the third decision behavior of the target vehicle at the first time instance The prediction of the behavior of the obstacle vehicle 120 can be updated by the Bayesian method in combination with the previous observation data. Specifically, this process involves updating the probability distribution of the behavior parameters of the obstacle vehicle 120 under different behavior modes.

[0043] Through the prediction of the behavior of the obstacle vehicle 120, the likelihood under different behavior modes can be determined. That is, in the case of knowing the behavior mode and parameters of the obstacle vehicle 120, the probability distribution of the decision behavior of the obstacle vehicle 120 is updated. Based on the updated behavior mode and behavior parameters of the obstacle vehicle 120, the fourth decision behavior of the obstacle vehicle 120 at the second time instance is calculated. The fourth decision behavior is derived by evaluating the probability distribution of the possible behavior of the obstacle vehicle 120, which is used to predict the future action of the obstacle vehicle 120 so that the target vehicle 110 can make appropriate responses when making decisions.

[0044] The above process combines the observation data of the target vehicle 110, the current and predicted state of the obstacle vehicle 120, and the control decision of the target vehicle 110, from which the possible behavior of the obstacle vehicle 120 at the future time instance can be gradually derived.

[0045] At block 203, the vehicle control device 130 determines at least one decision behavior of the target vehicle 110 after the first time instance based on the mutual state prediction at the second time instance and the fourth decision behavior.

[0046] The vehicle control device 130 determines the cost of each possible decision behavior of the target vehicle 110 based on the state prediction at the current time instance and the decision behavior of the obstacle vehicle 120 at the third time instance. This includes the direct cost of different decisions made by the target vehicle 110 in the current state (such as safety, efficiency, etc.) and the expected cost that may arise in the future.

[0047] By evaluating the cost of each possible decision behavior, the decision behavior sequence that minimizes the overall cost is selected. The overall cost here includes the direct cost at the current time instance and the expected cost at the future time instance. The expected cost is calculated based on the current state and the predicted future behavior of the obstacle vehicle 120, taking into account the state changes and corresponding costs that may occur in the future.

[0048] By dynamic programming method, the optimal decision behavior sequence that the target vehicle 110 should take throughout the decision-making process is determined. This process takes into account the cost of each possible decision behavior and selects the decision sequence that minimizes the total cost, thereby optimizing the control strategy of the target vehicle 110.

[0049] At block 204, the vehicle control device 130 can control the target vehicle 110 to drive after the first time instance through at least one decision behavior. In the process of controlling the target vehicle 110, the vehicle control device 130 adopts a prediction method that considers the target vehicle 110 and the obstacle vehicle 120 as a whole. First, by analyzing the observation data and historical data of the target vehicle 110 and the obstacle vehicle 120 at a certain time instance, the vehicle control device 130 predicts the behavior of the obstacle vehicle 120 at future time instances based on these data. This prediction involves the decision behavior of the obstacle vehicle 120 at different time nodes, thereby providing a targeted control strategy for the target vehicle 110.

[0050] The process of mutual state prediction of the target vehicle 110 and the obstacle vehicle 120 at the second time instance is described in detail below. The vehicle control device 130 determines the state change trend between the target vehicle 110 and the obstacle vehicle 120 based on the mutual state of the target vehicle 110 and the obstacle vehicle 120 at the first time instance. The first influence degree of the first decision behavior on the mutual state is determined. The second influence degree of the third decision behavior on the mutual state is determined. At least based on the state change trend, the first influence degree and the second influence degree, the mutual state prediction of the target vehicle 110 and the obstacle vehicle 120 at the second time instance is determined.

[0051] The mutual state prediction of the target vehicle 110 and the obstacle vehicle 120 at the second time can be represented as follows:

[0052]

[0053] Figure 3 An example flow 300 for mutual state prediction according to some embodiments of this disclosure is shown. k It can indicate the mutual status of the target vehicle 110 and the obstacle vehicle 120 at the first moment. k+1 It can indicate the mutual state prediction of target vehicle 110 and obstacle vehicle 120 at a second time. The second time is after the first time.

[0054] Combination Figure 3 As shown, the mutual state prediction 301 at the second time step is represented as x. k+1 The time interval 305 between the first and second moments can be represented as 1 / z. The mutual state prediction x at the second moment... k+1 After a time interval of 1 / z, the mutual states will revert to the state of the previous time step. By calculating the mutual states of the previous time step, the state change trend 306 can be obtained. Therefore, the state change trend 306 between the target vehicle 110 and the obstacle vehicle 120 can be continuously updated iteratively.

[0055] The trend of state change 306 can be represented as (This can be used to represent the dimension of a vector), and the trend of this state change can be determined by a joint kinematic model of the target vehicle 110 and the obstacle vehicle 120. The process of determining the joint kinematic model will be detailed later.

[0056] The decision-making behavior of the target vehicle 110 at the first moment can be indicated. The decision-making behavior of the target vehicle 110 at the first moment can be determined by the kinematic model of the target vehicle 110. The kinematic model of the target vehicle 110 can adopt a bicycle model. The bicycle model can be linearly represented as... It can be used to represent the dimension of matrices and vectors.

[0057] Using the control input matrix 302 of the obstacle vehicle 120, the degree of influence of the first decision behavior on the mutual state can be obtained. The decision-making behavior of vehicle 120 in the obstacle can be indicated. In this case, the degree of the second influence on the mutual state. (Combined) Figure 3 As shown, the control input matrix 302 of the obstacle vehicle 120 can be represented as B o (.), Their parameters are all time-varying, and their magnitudes depend on the mutual states. It can represent the dimension of a matrix.

[0058] The first influence degree of the third decision behavior of the target vehicle 110 on the mutual state can be obtained by using the control input matrix 303 of the target vehicle 110. The first influence degree of the third decision behavior of the target vehicle 110 on the mutual state can be indicated in the case that the third decision behavior of the target vehicle 110 is The first influence degree of the third decision behavior of the target vehicle 110 on the mutual state can be indicated in the case that the third decision behavior of the target vehicle 110 is Figure 3 The control input matrix 303 of the target vehicle 110 can be represented as B e (.), The parameters of the control input matrix 303 are all time-varying, and the size depends on the mutual state. The dimensions of the control input matrix 303 can be represented as

[0059] In addition, the random disturbance 304 is also included. As shown in Figure 3 The random disturbance 304 can be represented as d k The random disturbance 304 can be represented as the random disturbance caused by factors such as sensor errors, network delays, environmental disturbances, model errors, etc. It is assumed here that the random disturbance d k obeys a Gaussian distribution, and its mean is 0 and its variance is∑ d The parameters of the random disturbance are estimated by statistics.

[0060] Through the above process, the complex dynamic interaction between the target vehicle 110 and the obstacle vehicle 120 can be captured. Through this prediction, the decision behavior of the vehicle can be updated and adjusted in real time during the driving of the vehicle, ensuring that the target vehicle 110 can make a reasonable response to avoid potential dangers.

[0061] The determination process of the state change trend between the target vehicle 110 and the obstacle vehicle 120 will be described in detail below. Based on the first state of the target vehicle 110 at the first time, the vehicle control device 130 determines the first state change trend corresponding to the target vehicle 110 by using the target vehicle decision behavior model corresponding to the target vehicle 110. The first state is part of the mutual state. Based on the second state of the obstacle vehicle 120 at the first time, the vehicle control device 130 determines the second state change trend corresponding to the obstacle vehicle 120 by using the obstacle vehicle decision behavior model corresponding to the obstacle vehicle. The second state is part of the mutual state. Based on the first state change trend and the second state change trend, the state change trend between the target vehicle 110 and the obstacle vehicle 120 is determined.

[0062] As mentioned earlier, the target vehicle decision behavior model corresponding to the target vehicle 110 can be a bicycle model, which can be linearly represented as In the target vehicle decision behavior model, A e may correspond to the kinematic model matrix. Based on the first state of the target vehicle 110 at the first time With the target vehicle decision behavior model, a first state change trend corresponding to the target vehicle 110 can be determined.

[0063] The kinematic model of the obstacle vehicle can be represented as The kinematic model is typically a nonlinear function fitted by a neural network based on sensor data such as lidar or vision data. Based on the second state of the obstacle vehicle 120 at the first time point With the obstacle vehicle decision behavior model, a second state change trend corresponding to the obstacle vehicle 120 can be determined.

[0064] The state change trend between the target vehicle 110 and the obstacle vehicle 120 corresponds to the first state of the target vehicle 110 at the first time point and the second state of the obstacle vehicle 120 at the first time point The state change trend between the target vehicle 110 and the obstacle vehicle 120 is determined by using the decision behavior models of the target vehicle 110 and the obstacle vehicle 120 respectively to determine their respective state updates, and then integrating the results into a unified expression f(x k ).

[0065] By integrating the decision behavior models of the target vehicle 110 and the obstacle vehicle 120, the dynamic behavior of both and their mutual influence can be more comprehensively described. This unified dynamic modeling method makes the prediction and decision more accurate and consistent, and can better reflect the interaction between the two in the actual driving process, providing a solid foundation for developing more accurate control strategies.

[0066] The determination process of the fourth decision behavior of the obstacle vehicle 120 at the second time point is described in detail below. The fourth decision behavior at the second time point is completed based on the first decision behavior of the obstacle vehicle 120 at the first time point and the mutual state prediction of the target vehicle 110 and the obstacle vehicle 120 at the second time point. That is, the entire determination of the decision behavior of the obstacle vehicle 120 is an iterative process. The principle of each iteration is the same. Taking the fourth decision behavior at the second time point as an example, the entire iteration process is described. The vehicle control device 130 determines the probability of at least one behavior parameter of the obstacle vehicle 120 based on the observation data at the first time point and the mutual state prediction. Based on the observation data at the first time point and the third decision behavior of the target vehicle 110, the probability of at least one behavior mode of the obstacle vehicle 120 is determined. Based on the probability of at least one behavior parameter and the probability of at least one behavior mode, the fourth decision behavior of the obstacle vehicle 120 at the second time point is determined.

[0067] The vehicle control device 130 determines the probability of at least one behavior parameter of the obstacle vehicle 120 based on the observation data H k at the first time point and the historical observation data H k-1The first behavior parameter θ and the first behavior mode M of the obstacle vehicle 120 can be determined using a prior distribution. The prior distribution can be indicative of being based on observation data. The observation data can be based on the behavior parameter θ of the obstacle vehicle 120 in the behavior mode M M The first decision behavior of the obstacle vehicle 120 at the first time instant can be inferred

[0068] As the observation data H k The observation data can be updated to H k+1 The observation data H k+1 The observation data H

[0069] Based on the updated observation data, a probability of at least one behavior parameter of the obstacle vehicle 120 and a probability of at least one behavior mode of the obstacle vehicle 120 can be determined. Exemplarily, the behavior mode can include a conservative behavior mode, a normal behavior mode, a risky behavior mode, etc. The behavior parameter can include a left turn, a right turn, an acceleration, a deceleration, a lane changing, etc.

[0070] Based on the above process, a fourth decision behavior of the obstacle vehicle 120 can be determined based on the probability of at least one behavior parameter of the obstacle vehicle 120 and the probability of at least one behavior mode of the obstacle vehicle 120.

[0071] The determination process of the probability of at least one behavior parameter and the probability of at least one behavior mode will be introduced in turn. First, the determination process of the probability of behavior parameter will be introduced. The first decision behavior of the obstacle vehicle 120 at the first time instant is determined based on the first behavior parameter and the first behavior mode of the obstacle vehicle 120. The obstacle vehicle 120 can correspond to a plurality of behavior parameters, and the determination process of the probability of each behavior parameter in the plurality of behavior parameters is the same, and the determination process of the probability of a given behavior parameter in the plurality of behavior parameters will be taken as an example. The vehicle control device 130 determines a first probability of the occurrence of the mutual state prediction between the target vehicle 110 and the obstacle vehicle 120 at the second time instant based on the first behavior parameter, the first behavior mode, the observation data at the first time instant, and the third decision behavior. The prior probability of the first behavior parameter of the obstacle vehicle 120 is obtained. Based on the first behavior parameter, the first behavior mode, and the observation data at the first time instant, a second probability of the occurrence of the mutual state prediction between the target vehicle 110 and the obstacle vehicle 120 at the second time instant is determined. Based on the first probability, the prior probability of the first behavior parameter, and the second probability, the probability of the given behavior parameter at the second time instant is determined.

[0072] The probability of the given behavior parameter can be represented as follows:

[0073]

[0074] p(θ M′ |H k+1 M) can be represented as the updated obstacle vehicle 120 corresponding to the given behavior parameter θ. M′ The posterior probability. That is, based on the new observation data H. k+1 And the first behavior pattern M, determines that the obstacle vehicle 120 has a given behavior parameter θ. M′ The probability of.

[0075] This can be represented as the first behavioral parameter θ of the obstacle vehicle 120. M First behavioral pattern M, first moment observation data H k The third decision-making behavior of target vehicle 110 In this case, new observation data x is determined. k+1 The first probability.

[0076] p(θ M |H k M) can be represented as H based on the observation data at the first moment. k In the case of the first behavior mode M, the first behavior parameter θ of the obstacle vehicle 120 is determined. M The prior probability.

[0077] It is a normalization factor, which can be expressed as the third decision-making behavior of the target vehicle 110 in the first behavior mode M. And the observation data H at the first moment k In this case, new observation data x is determined. k+1 The second probability. The normalization factor typically integrates all possible behavioral parameters.

[0078] Using the first probability and the first behavioral parameter θ M The prior probability and the second probability can be used to obtain the updated given behavior parameter θ. M′ The probability of.

[0079] The process of determining the probability of a behavior mode is described as follows. The obstacle vehicle 120 can correspond to a plurality of behavior modes, and the process of determining the probability of each behavior mode is the same, and the process of determining the probability of a given behavior mode in the plurality of behavior modes is described as an example. The vehicle control device 130 determines a third probability that the mutual state prediction between the target vehicle 110 and the obstacle vehicle 120 at the second time occurs, based on the third decision behavior, the first behavior mode, and the observation data at the first time. The prior probability of the first behavior mode of the obstacle vehicle 120 is obtained. A fourth probability that the mutual state prediction between the target vehicle 110 and the obstacle vehicle 120 at the second time occurs is determined, based on the third decision behavior of the target vehicle 110 and the observation data at the first time. The probability of the given behavior mode at the second time is determined, based on the third probability, the prior probability of the first behavior mode, and the fourth probability.

[0080] The probability of the given behavior mode can be expressed as follows:

[0081]

[0082] p(M' | H k+1 ) can represent the posterior probability of the behavior mode M' of the obstacle vehicle 120 based on the new observation data H k+1 . That is, based on the new observation data H k+1 , the probability that the obstacle vehicle 120 has the given behavior mode M' is determined.

[0083] p(x k+1 | H k , M, a) can represent the third probability that the new observation data x k+1 is determined based on the observation data H k at the first time, the first behavior mode M, and the third decision behavior a of the target vehicle 110. This represents the change in the state of the obstacle vehicle 120 that can be caused by the behavior of the target vehicle 110 under the first behavior mode M.

[0084] p(M | H k ) can represent the prior probability that the obstacle vehicle 120 has the first behavior mode M based on the observation data H k at the first time.

[0085] Z is a normalization factor, and can be expressed as the fourth probability that the new observation data x k+1 is determined based on the observation data H k at the first time and the third decision behavior a of the target vehicle 110.

[0086] The updated probability of the given behavior mode M' can be obtained by the third probability, the prior probability of the first behavior mode M, and the fourth probability.

[0087] After determining the probability of the at least one behavior parameter and the probability of the at least one behavior mode, the vehicle control device 130 can determine a probability distribution of the at least one behavior parameter and the at least one behavior mode based on the probability of the at least one behavior parameter and the probability of the at least one behavior mode. Based on the probability distribution, the vehicle control device 130 can determine the fourth decision behavior of the obstacle vehicle 120.

[0088] Based on the probability of the at least one behavior parameter and the probability of the at least one behavior mode, a joint distribution of a given behavior parameter and a given behavior mode can be obtained. The joint distribution can be represented as follows:

[0089] F k ′ +1 = p(θ M′ |H k+1 ; M)p(M′|H k+1 ) (4)

[0090] After updating the probability of the given behavior mode and the probability of the given behavior parameter of the obstacle vehicle 120, the probability of the given behavior parameter θ k+1 of the obstacle vehicle 120 under the given behavior mode M’ based on the observation data H M′ at the second time can be obtained. This probability reflects the likelihood of the obstacle vehicle 120 taking the given behavior θ M′ under the given behavior mode M’.

[0091] Taking into account all behavior parameters of the obstacle vehicle 120 under the given behavior mode M’, they can be represented as follows:

[0092] F k+1 = {F k ′ +1 (θ M′ , M′)| θ M′ ∈ θ M , M′∈ M} (5)

[0093] Combining all possible behavior modes and behavior parameters, a complete probability distribution can be constructed. This probability distribution describes all possible behaviors of the obstacle vehicle 120 under different behavior modes and behavior parameters. This means that the vehicle control device 130 has obtained a comprehensive behavior prediction distribution by all possible behavior modes and behavior parameters of the obstacle vehicle 120.

[0094] Based on the complete probability distribution, the fourth decision behavior of the obstacle vehicle 120 can be obtained. The determination of the fourth decision behavior can be represented as follows:

[0095]

[0096] It can represent the behavior corresponding to the i-th behavior parameter under behavior pattern M. This can represent the probability of that behavior. That is, It can be the behavior with the highest probability of behavior parameters determined based on expression (5) when the probability of behavior pattern is the highest.

[0097] After determining the fourth decision action of the obstacle vehicle 120 at the second moment, the vehicle control device 130 can determine a first cost based on the mutual state of the target vehicle 110 and the obstacle vehicle 120 at the first moment and the third decision action of the target vehicle 110. Based on the predicted mutual state of the target vehicle 110 and the obstacle vehicle 120 at the second moment and the fourth decision action, a second cost is determined. Based on the first and second costs, at least one decision action of the target vehicle 110 after the first moment is determined. The at least one decision action can be represented as follows:

[0098]

[0099] Based on the initial state Based on the mutual state x of the target vehicle 110 and the obstacle vehicle 120 at the first moment k The probability distribution F obtained by calculating equation (6) k+1 and the previously established probability distribution F k Combining the cost function The first cost in Second cost E[V] k+1 (x k+1 ,F k+1 )], and the final state x N The termination value R T (x N This allows us to obtain every possible decision-making behavior. The cost. The cost of the terminated state is known to be V. N (x N ,F N ) = R T (x N In other words, when the final state x is reached... N In this case, the cost value is directly derived from the objective function R. T (x N The final state x is determined by (the state of x). N That is, the mutual state of the target vehicle 110 and the obstacle vehicle 120 at the end time, F N In the final state x N The probability distribution of the behavior patterns and behavioral parameters of obstacle vehicle 120.

[0100] By minimizing the cost function, the vehicle control device 130 can determine an optimal decision behavior sequence that minimizes the overall decision cost, which can be represented as

[0101]

[0102] The first cost and the second cost can be determined based on multiple factors, for example, can be related to the distance of the target vehicle 110 to the obstacle vehicle 120, road boundaries, dangerous areas, and other objects. The closer the distance, the higher the cost. In addition, it can also be related to the continuity and smoothness of the path and speed. If the path is discontinuous or the speed changes drastically, it will lead to an increase in efficiency cost.

[0103] The first cost can indicate the immediate value of taking a decision behavior at the current state The second cost can indicate the expected future cost based on the mutual state prediction x k+1 at the second time and the probability distribution F k+1 under the calculation formula (6), which is a comprehensive consideration of all possible future states, including the expected cost of future safety, efficiency, and other factors. By considering the expected future cost, it can ensure that the current decision not only minimizes the immediate cost, but also optimizes the overall cost in the future.

[0104] Through the above process, based on the mutual state prediction of the target vehicle 110 and the obstacle vehicle 120, the behavior pattern and the joint probability distribution of the behavior parameters of the obstacle vehicle 120, the vehicle control device 130 can predict the future cost that can be generated under different decision behavior sequences. By minimizing the expected future cost, the vehicle control device 130 can select an optimal decision behavior sequence, thereby achieving intelligent control of the target vehicle 110 in a dynamic and complex driving environment. This decision mechanism helps to improve the safety and rationality of the decision of the target vehicle 110, especially in the face of complex traffic scenarios, which can more effectively avoid potential risks.

[0105] Figure 4 A schematic structural block diagram of a vehicle control device 400 according to some embodiments of the present disclosure is shown. The device 400 may, for example, be implemented in or included in the vehicle control device 130. Various modules / components in the device 400 can be implemented by hardware, software, firmware, or any combination thereof.

[0106] As shown in the figure, the device 400 comprises a mutual state prediction module 401 configured to determine a mutual state prediction of the target vehicle and the obstacle vehicle at a second time based on at least a first decision behavior of the obstacle vehicle at a first time and observation data of the target vehicle at the first time, the observation data comprising a mutual state of the target vehicle and the obstacle vehicle at the first time and a second decision behavior of the target vehicle at a third time, the second time being after the first time, and the third time being before the first time. An obstacle vehicle decision behavior determination module 402 configured to determine a fourth decision behavior of the obstacle vehicle at the second time based on at least the mutual state prediction at the second time and the third decision behavior of the target vehicle at the first time. A target vehicle decision behavior determination module 403 configured to determine at least one decision behavior of the target vehicle after the first time based on at least the mutual state prediction at the second time and the fourth decision behavior. A target vehicle control module 404 configured to control a driving process of the target vehicle after the first time by the at least one decision behavior.

[0107] In some embodiments of the present disclosure, the mutual state prediction module 401 can be specifically configured to determine a state change trend between the target vehicle and the obstacle vehicle based on a mutual state of the target vehicle and the obstacle vehicle at the first time. Determine a first influence degree of the first decision behavior on the mutual state. Determine a second influence degree of the third decision behavior on the mutual state. Determine the mutual state prediction of the target vehicle and the obstacle vehicle at the second time based on at least the state change trend, the first influence degree and the second influence degree.

[0108] In some embodiments of the present disclosure, the mutual state prediction module 401 can be further specifically configured to determine a first state change trend corresponding to the target vehicle based on a first state of the target vehicle at the first time by using a target vehicle decision behavior model corresponding to the target vehicle, the first state being part of the mutual state. Determine a second state change trend corresponding to the obstacle vehicle based on a second state of the obstacle vehicle at the first time by using an obstacle vehicle decision behavior model corresponding to the obstacle vehicle, the second state being part of the mutual state. Determine the state change trend between the target vehicle and the obstacle vehicle based on the first state change trend and the second state change trend.

[0109] In some embodiments of the present disclosure, the obstacle vehicle decision behavior determination module 402 can be configured to determine a probability of at least one behavior parameter of the obstacle vehicle based on the observation data at the first time and the mutual state prediction at the second time. Determine a probability of at least one behavior mode of the obstacle vehicle based on the observation data at the first time and the third decision behavior of the target vehicle. Determine the fourth decision behavior of the obstacle vehicle at the second time based on the probability of the at least one behavior parameter and the probability of the at least one behavior mode.

[0110] In some embodiments of the present disclosure, the first decision behavior of the obstacle vehicle is determined based on the first behavior parameter and the first behavior mode of the obstacle vehicle, and based on this, the obstacle vehicle decision behavior determination module 402 can be further configured to: for a given behavior parameter in the at least one behavior parameter, determine a first probability of occurrence of a mutual state prediction between the target vehicle and the obstacle vehicle at the second time based on the first behavior parameter, the first behavior mode, the observation data at the first time, and the third decision behavior. Obtain a prior probability of the first behavior parameter of the obstacle vehicle. Determine a second probability of occurrence of a mutual state prediction between the target vehicle and the obstacle vehicle at the second time based on the first behavior parameter, the first behavior mode, and the observation data at the first time. Determine the probability of the given behavior parameter at the second time based on the first probability, the prior probability of the first behavior parameter, and the second probability.

[0111] In some embodiments of the present disclosure, the first decision behavior of the obstacle vehicle is determined based on the first behavior parameter and the first behavior mode of the obstacle vehicle, and based on this, the obstacle vehicle decision behavior determination module 402 can be further configured to: for a given behavior mode in the at least one behavior mode, determine a third probability of occurrence of a mutual state prediction between the target vehicle and the obstacle vehicle at the second time based on the third decision behavior, the first behavior mode, and the observation data at the first time. Obtain a prior probability of the first behavior mode of the obstacle vehicle. Determine a fourth probability of occurrence of a mutual state prediction between the target vehicle and the obstacle vehicle at the second time based on the third decision behavior of the target vehicle and the observation data at the first time. Determine the probability of the given behavior mode at the second time based on the third probability, the prior probability of the first behavior mode, and the fourth probability.

[0112] In some embodiments of the present disclosure, the obstacle vehicle decision behavior determination module 402 can be further configured to: determine a probability distribution of the at least one behavior parameter and the at least one behavior mode based on the probability of the at least one behavior parameter and the probability of the at least one behavior mode. Determine the fourth decision behavior of the obstacle vehicle based on the probability distribution.

[0113] In some embodiments of the present disclosure, the fourth decision behavior of the obstacle vehicle is determined based on the probability of the at least one behavior parameter and the probability of the at least one behavior mode corresponding to the obstacle vehicle, and based on this, the target vehicle decision behavior determination module 403 can be configured to: determine a first cost based on the mutual state of the target vehicle and the obstacle vehicle at the first time and the third decision behavior of the target vehicle. Determine a second cost based on the mutual state prediction of the target vehicle and the obstacle vehicle at the second time, the probability of the at least one behavior parameter and the probability of the at least one behavior mode corresponding to the obstacle vehicle. Determine at least one decision behavior of the target vehicle after the first time based on the first cost and the second cost.

[0114] Figure 5A block diagram of an electronic device 500 in which one or more embodiments of the present disclosure may be implemented is shown. It should be understood that... Figure 5 The electronic device 500 shown is merely exemplary and should not be construed as limiting the functionality and scope of the embodiments described herein. Figure 5 The illustrated electronic device 500 may include or be implemented as Figure 1 Vehicle control equipment 130, or Figure 4 Device 400.

[0115] like Figure 5 As shown, electronic device 500 is in the form of a general-purpose electronic device. Components of electronic device 500 may include, but are not limited to, one or more processors or processing units 510, memory 520, storage device 530, one or more communication units 540, one or more input devices 550, and one or more output devices 560. Processing unit 510 may be a physical or virtual processor and is capable of performing various processes according to programs stored in memory 520. In a multiprocessor system, multiple processing units execute computer-executable instructions in parallel to improve the parallel processing capability of electronic device 500.

[0116] Electronic device 500 typically includes multiple computer storage media. Such media can be any accessible media that is accessible to electronic device 500, including but not limited to volatile and non-volatile media, removable and non-removable media. Memory 520 can be volatile memory (e.g., registers, cache, random access memory (RAM)), non-volatile memory (e.g., read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory), or some combination thereof. Storage device 530 can be removable or non-removable media and can include machine-readable media, such as flash drives, disks, or any other media that can be used to store information and / or data and can be accessed within electronic device 500.

[0117] Electronic device 500 may further include additional removable / non-removable, volatile / non-volatile storage media. Although not explicitly stated... Figure 5 As shown, disk drives for reading from or writing to removable, non-volatile disks (e.g., "floppy disks") and optical disk drives for reading from or writing to removable, non-volatile optical disks can be provided. In these cases, each drive can be connected to a bus (not shown) via one or more data media interfaces. Memory 520 may include computer program product 525 having one or more program modules configured to perform various methods or actions of various embodiments of this disclosure.

[0118] Communication unit 540 enables communication with other electronic devices via a communication medium. Additionally, the functionality of components of electronic device 500 can be implemented using a single computing cluster or multiple computing machines capable of communicating via communication connections. Therefore, electronic device 500 can operate in a networked environment using logical connections to one or more other servers, network personal computers (PCs), or another network node.

[0119] Input device 550 can be one or more input devices, such as a mouse, keyboard, trackball, etc. Output device 560 can be one or more output devices, such as a monitor, speaker, printer, etc. Electronic device 500 can also communicate with one or more external devices (not shown) via communication unit 540 as needed. These external devices include storage devices, display devices, etc., and can communicate with one or more devices that enable user interaction with electronic device 500, or with any device that enables electronic device 500 to communicate with one or more other electronic devices (e.g., network card, modem, etc.). Such communication can be performed via input / output (I / O) interface (not shown).

[0120] According to an exemplary implementation of this disclosure, a computer-readable storage medium is provided that stores computer-executable instructions thereon, wherein the computer-executable instructions are executed by a processor to implement the methods described above. According to an exemplary implementation of this disclosure, a computer program product is also provided, which is tangibly stored on a non-transitory computer-readable medium and includes computer-executable instructions, which are executed by a processor to implement the methods described above.

[0121] According to an exemplary implementation of this disclosure, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform... Figure 2 The methods provided are among the various optional methods available in the code, so they will not be elaborated upon here.

[0122] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatuses, devices, and computer program products implemented according to this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0123] These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processing unit of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0124] Computer-readable program instructions can be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions that execute on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0125] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction, which contains one or more executable instructions for implementing the specified logical function. In some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0126] Various implementations of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed implementations. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described implementations. The terminology used herein is chosen to best explain the principles, practical applications, or improvements to technology in the market, or to enable others skilled in the art to understand the various implementations disclosed herein.

Claims

1. A method of vehicle control, comprising: determining a mutual state prediction of a target vehicle and an obstacle vehicle at a second time based at least on a first decision behavior of the obstacle vehicle at a first time and observation data of the target vehicle at the first time, the observation data comprising mutual states of the target vehicle and the obstacle vehicle at the first time and a second decision behavior of the target vehicle at a third time, the second time being after the first time, and the third time being before the first time; determining a fourth decision behavior of the obstacle vehicle at the second time based at least on the mutual state prediction at the second time and a third decision behavior of the target vehicle at the first time; determining at least one decision behavior of the target vehicle after the first time based at least on the mutual state prediction at the second time and the fourth decision behavior; and controlling a driving process of the target vehicle after the first time by the at least one decision behavior. 2.The method of claim 1, wherein the determining the mutual state prediction of the target vehicle and the obstacle vehicle at a second time comprises: determining a state change tendency between the target vehicle and the obstacle vehicle based on mutual states of the target vehicle and the obstacle vehicle at the first time; determining a first influence degree of the first decision behavior on the mutual states; determining a second influence degree of the third decision behavior on the mutual states; and determining the mutual state prediction of the target vehicle and the obstacle vehicle at the second time based at least on the state change tendency, the first influence degree and the second influence degree. 3.The method of claim 2, wherein the determining the state change tendency between the target vehicle and the obstacle vehicle comprises: determining a first state change tendency corresponding to the target vehicle based on a first state of the target vehicle at the first time by using a target vehicle decision behavior model corresponding to the target vehicle, the first state being part of the mutual states; determining a second state change tendency corresponding to the obstacle vehicle based on a second state of the obstacle vehicle at the first time by using an obstacle vehicle decision behavior model corresponding to the obstacle vehicle, the second state being part of the mutual states; and determining the state change tendency between the target vehicle and the obstacle vehicle based on the first state change tendency and the second state change tendency. 4.The method of claim 1, wherein the determining the fourth decision behavior of the obstacle vehicle at the second time comprises: determining a probability of at least one behavior parameter of the obstacle vehicle based on the observation data at the first time and the mutual state prediction at the second time; determining a probability of at least one behavior mode of the obstacle vehicle based on the observation data at the first time and the third decision behavior; and determining the fourth decision behavior of the obstacle vehicle at the second time based on the probability of the at least one behavior parameter and the probability of the at least one behavior mode. ​ ​ ​ ​ 5. The method of claim 4, wherein the first decision behavior of the obstacle vehicle is determined based on a first behavior parameter and a first behavior pattern of the obstacle vehicle, and determining the probability of the at least one behavior parameter of the obstacle vehicle comprises: For a given behavior parameter of the at least one behavior parameter, determining, based on the first behavior parameter, the first behavior pattern, the observation data at the first time instant, and the third decision behavior, a first probability of occurrence of a mutual state prediction between the target vehicle and the obstacle vehicle at a second time instant; obtaining a prior probability of the first behavior parameter of the obstacle vehicle; determining, based on the first behavior parameter, the first behavior pattern, and the observation data at the first time instant, a second probability of occurrence of a mutual state prediction between the target vehicle and the obstacle vehicle at a second time instant; and determining, based on the first probability, the prior probability of the first behavior parameter, and the second probability, a probability of the given behavior parameter at the second time instant.

6. The method of claim 4, wherein the first decision behavior of the obstacle vehicle is determined based on the first behavior parameter and the first behavior pattern of the obstacle vehicle, and determining the probability of at least one behavior pattern of the obstacle vehicle comprises: For a given behavior pattern of the at least one behavior pattern, determining, based on the third decision behavior, the first behavior pattern, and the observation data at the first time instant, a third probability of occurrence of a mutual state prediction between the target vehicle and the obstacle vehicle at a second time instant; obtaining a prior probability of the first behavior pattern of the obstacle vehicle; determining, based on the third decision behavior and the observation data at the first time instant, a fourth probability of occurrence of a mutual state prediction between the target vehicle and the obstacle vehicle at a second time instant; and determining, based on the third probability, the prior probability of the first behavior pattern, and the fourth probability, a probability of the given behavior pattern at the second time instant.

7. The method of claim 4, wherein determining the fourth decision behavior of the obstacle vehicle at the second time instant based on the probability of the at least one behavior parameter and the probability of the at least one behavior pattern comprises: determining, based on the probability of the at least one behavior parameter and the probability of the at least one behavior pattern, a probability distribution of the at least one behavior parameter and the at least one behavior pattern; and determining, based on the probability distribution, the fourth decision behavior of the obstacle vehicle.

8. The method of claim 1, wherein the fourth decision behavior of the obstacle vehicle is determined based on the probability of the at least one behavior parameter and the probability of the at least one behavior pattern corresponding to the obstacle vehicle, and determining the at least one decision behavior of the target vehicle after the first time instant comprises: determining a first cost based on the mutual state of the target vehicle and the obstacle vehicle at the first time instant and the third decision behavior; determining a second cost based on the mutual state prediction of the target vehicle and the obstacle vehicle at a second time instant, the probability of the at least one behavior parameter and the probability of the at least one behavior pattern corresponding to the obstacle vehicle; and determining the at least one decision behavior of the target vehicle after the first time instant based on the first cost and the second cost. comprises: ​ 9. An apparatus for vehicle control, characterized by ​ An inter-state prediction module configured to determine an inter-state prediction of the target vehicle and the obstacle vehicle at a second time instance based on at least a first decision behavior of the obstacle vehicle at a first time instance and observation data of the target vehicle at the first time instance, the observation data comprising an inter-state of the target vehicle and the obstacle vehicle at the first time instance and a second decision behavior of the target vehicle at a third time instance, the second time instance being after the first time instance, and the third time instance being before the first time instance; An obstacle vehicle decision behavior determination module configured to determine a fourth decision behavior of the obstacle vehicle at the second time instance based on at least the inter-state prediction at the second time instance and a third decision behavior of the target vehicle at the first time instance; A target vehicle decision behavior determination module configured to determine at least one decision behavior of the target vehicle after the first time instance based on at least the inter-state prediction at the second time instance and the fourth decision behavior; and A target vehicle control module configured to control a driving process of the target vehicle after the first time instance by the at least one decision behavior.

10. An electronic device, comprising: comprising: at least one processing unit; and at least one memory that is coupled to the at least one processing unit and stores instructions for execution by the at least one processing unit, the instructions, when executed by the at least one processing unit, cause the electronic device to perform the method according to any one of claims 1 to 8. The computer program is executable by a processor to implement the method according to any one of claims 1 to 8.

11. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer executable instructions, when executed by a processor, implement the method according to any one of claims 1 to 8.

12. A computer program product comprising computer executable instructions, characterised in that, ​

Images