A method and apparatus for controlling an unmanned vehicle

Abstract

The application discloses a kind of method and device for controlling unmanned vehicle, it is related to intelligent driving technical field.The specific embodiment of the method includes: obtaining the space environment information of the space where unmanned vehicle is currently located, according to space environment information, target scene template is selected, the scene constraint condition included in target scene template is used to analyze the space environment information;According to the result of analysis, the unmanned vehicle is regulated;By dynamically selecting target scene template according to the difference of space environment information in the process of vehicle driving to regulate unmanned vehicle, the flexibility of controlling unmanned vehicle is improved, and the accuracy of controlling unmanned vehicle is improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent driving technology, and in particular to a method and apparatus for controlling unmanned vehicles. Background Technology

[0002] With the development of autonomous driving technology, vehicles and equipment equipped with autonomous driving technology are being widely used in logistics, smart cities and other scenarios.

[0003] The decision-making method used for vehicle action commands is an important part of autonomous driving technology. At present, the decision-making methods are usually based on pre-set rules or pre-trained models. The existing decision-making methods have a strong coupling with geospatial features. In real traffic scenarios with a large number of state dimensions and many uncertain factors, the accuracy of vehicle commands determined by the decision-making methods is low. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a method and apparatus for controlling an autonomous vehicle, which can acquire spatial environment information of the space where the autonomous vehicle is currently located, select a target scene template based on the spatial environment information, and analyze the spatial environment information using the scene constraints included in the target scene template; by dynamically selecting the target scene template according to the different spatial environment information during the vehicle's driving process to regulate the autonomous vehicle, the flexibility and accuracy of controlling the autonomous vehicle are improved.

[0005] To achieve the above objectives, according to one aspect of the present invention, a method for controlling an autonomous vehicle is provided, comprising: acquiring spatial environment information of the space currently occupied by the autonomous vehicle; selecting a target scene template for the autonomous vehicle from multiple types of scene templates based on the spatial environment information and scene constraint conditions included in multiple types of scene templates, wherein the scene template further includes scene constraint conditions; analyzing the spatial environment information based on the scene constraint conditions included in the target scene template; and adjusting the autonomous vehicle according to the analysis results.

[0006] Optionally, the method for controlling the autonomous vehicle includes: the spatial environment information includes one or more of the following: geographical range information, environmental object information, autonomous vehicle status, environmental object status, and driving constraint elements of the geographical range.

[0007] Optionally, in the method for controlling an autonomous vehicle, the step of selecting a target scene template for the autonomous vehicle from multiple types of scene templates includes: searching for target scene constraints satisfied by the spatial environment information from the scene constraints included in the multiple types of scene templates; and selecting a target scene template to which the target scene constraints belong for the autonomous vehicle.

[0008] Optionally, in the method for controlling an autonomous vehicle, the step of finding the target scene constraint condition satisfied by the spatial environment information includes: matching the spatial environment information with one or more scene elements included in the scene constraint condition; and determining the scene constraint condition as the target scene constraint condition when the spatial environment information is successfully matched with all the scene elements included in the scene constraint condition.

[0009] Optionally, the method for controlling the autonomous vehicle includes: each type of scene template corresponds to a driving scenario; the scene constraints included in each type of scene template include: constraints in the state space containing specific state features of the corresponding driving scenario; and constraints in the action space containing a specific coordinate system of the corresponding driving scenario.

[0010] Optionally, in the method for controlling an autonomous vehicle, the analysis of the spatial environment information includes: determining the state space of the autonomous vehicle based on the spatial environment information and the constraints of the state space included in the target scene template; determining the action space of the autonomous vehicle based on the state of the autonomous vehicle included in the spatial environment information and the constraints of the action space included in the target scene template; and generating control commands for the autonomous vehicle based on the state space of the autonomous vehicle and the action space of the autonomous vehicle.

[0011] Optionally, the method for controlling an autonomous vehicle, wherein determining the state space of the autonomous vehicle includes: for the case where the driving scenario is a lane driving scenario, converting the geographical range information included in the spatial environment information into Frenet coordinates, determining the Frenet coordinates of the autonomous vehicle, the Frenet coordinates of environmental objects, and the associated lanes included in the geographical range information, and determining the driving state of the autonomous vehicle and environmental objects in their respective lanes; for the case where the driving scenario is a drivable area driving scenario, converting the geographical range information included in the spatial environment information into Cartesian coordinates, determining the Cartesian coordinates of the autonomous vehicle, the Cartesian coordinates of environmental objects, the target point of the autonomous vehicle, and the boundary point list of the drivable area included in the geographical range information, and calculating the driving state of the autonomous vehicle in the drivable area.

[0012] Optionally, in the method for controlling an autonomous vehicle, determining the action space of the autonomous vehicle includes: in the case of driving in a lane, using the Frenet coordinate system and the current driving state of the autonomous vehicle included in the spatial environment information to determine the driving state of the vehicle and the target location to be reached within a future set time range; in the case of driving in a drivable area, using the Cartesian coordinate system and the current driving state of the autonomous vehicle included in the spatial environment information to determine the driving state of the vehicle and the target location to be reached within a future set time range.

[0013] Optionally, the method for controlling the autonomous vehicle further includes: acquiring the actions performed by the autonomous vehicle based on control; rewarding the target scene template according to a preset evaluation strategy when the actions performed by the autonomous vehicle meet the set action evaluation constraints; penalizing the target scene template according to a preset evaluation strategy when the actions performed by the autonomous vehicle do not meet the set action evaluation constraints; and adjusting the target scene template based on the reward and / or penalty results.

[0014] Optionally, the method for controlling the unmanned vehicle further includes: dynamically adjusting the target scene template based on the spatial environment information acquired in real time.

[0015] To achieve the above objectives, according to a second aspect of the present invention, an apparatus for controlling an unmanned vehicle is provided, characterized in that it includes: an information acquisition module, a template determination module, and a vehicle control module; wherein,

[0016] The information acquisition module is used to acquire spatial environment information of the space where the unmanned vehicle is currently located;

[0017] The template determination module is used to select a target scene template for the autonomous vehicle from multiple types of scene templates based on the spatial environment information and the scene constraints included in the multiple types of scene templates, wherein the scene template further includes scene constraints.

[0018] The vehicle control module is used to analyze the spatial environment information based on the scene constraints included in the target scene template; and to control the unmanned vehicle according to the analysis results.

[0019] Optionally, the device for controlling the autonomous vehicle includes: the spatial environment information includes one or more of the following: geographical range information, environmental object information, autonomous vehicle status, environmental object status, and driving constraint elements of the geographical range.

[0020] Optionally, in the device for controlling the autonomous vehicle, the step of selecting a target scene template for the autonomous vehicle from multiple types of scene templates includes: searching for target scene constraints satisfied by the spatial environment information from the scene constraints included in the multiple types of scene templates; and selecting the target scene template to which the target scene constraints belong for the autonomous vehicle.

[0021] Optionally, in the device for controlling the autonomous vehicle, the step of finding the target scene constraint condition satisfied by the spatial environment information includes: matching the spatial environment information with one or more scene elements included in the scene constraint condition; and determining the scene constraint condition as the target scene constraint condition when the spatial environment information is successfully matched with all the scene elements included in the scene constraint condition.

[0022] Optionally, the device for controlling the unmanned vehicle includes: each type of scene template corresponds to a driving scenario; each type of scene template includes scene constraints, including: state space constraints containing specific state features of the corresponding driving scenario; and action space constraints containing a specific coordinate system of the corresponding driving scenario.

[0023] Optionally, in the device for controlling the autonomous vehicle, the analysis of the spatial environment information includes: determining the state space of the autonomous vehicle based on the spatial environment information and the constraints of the state space included in the target scene template; determining the action space of the autonomous vehicle based on the state of the autonomous vehicle included in the spatial environment information and the constraints of the action space included in the target scene template; and generating control commands for the autonomous vehicle based on the state space of the autonomous vehicle and the action space of the autonomous vehicle.

[0024] Optionally, in the device for controlling the autonomous vehicle, determining the state space of the autonomous vehicle includes: for the case where the driving scenario is a lane driving scenario, converting the geographical range information included in the spatial environment information into Frenet coordinates, determining the Frenet coordinates of the autonomous vehicle, the Frenet coordinates of environmental objects, and the associated lanes included in the geographical range information, and determining the driving state of the autonomous vehicle and environmental objects in their respective lanes; for the case where the driving scenario is a drivable area driving scenario, converting the geographical range information included in the spatial environment information into Cartesian coordinates, determining the Cartesian coordinates of the autonomous vehicle, the Cartesian coordinates of environmental objects, the target point of the autonomous vehicle, and the boundary point list of the drivable area included in the geographical range information, and calculating the driving state of the autonomous vehicle in the drivable area.

[0025] Optionally, in the device for controlling the autonomous vehicle, determining the autonomous vehicle's action space includes: in the case of driving in a lane, using the Frenet coordinate system and the current driving state of the autonomous vehicle included in the spatial environment information to determine the vehicle's driving state and target location within a future set time range; in the case of driving in a drivable area, using the Cartesian coordinate system and the current driving state of the autonomous vehicle included in the spatial environment information to determine the vehicle's driving state and target location within a future set time range.

[0026] Optionally, the device for controlling the autonomous vehicle further includes: acquiring actions performed by the autonomous vehicle based on control; rewarding the target scene template according to a preset evaluation strategy when the actions performed by the autonomous vehicle meet the set action evaluation constraints; penalizing the target scene template according to a preset evaluation strategy when the actions performed by the autonomous vehicle do not meet the set action evaluation constraints; and adjusting the target scene template based on the reward and / or penalty results.

[0027] Optionally, the device for controlling the unmanned vehicle further includes: dynamically adjusting the target scene template based on the spatial environment information acquired in real time.

[0028] To achieve the above objectives, according to a third aspect of the present invention, an electronic device for controlling an autonomous vehicle is provided, characterized in that it includes: one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors perform any of the methods described above for controlling an autonomous vehicle.

[0029] To achieve the above objectives, according to a fourth aspect of the present invention, a computer-readable medium is provided having a computer program stored thereon, characterized in that, when the program is executed by a processor, it implements any of the methods described above for controlling an unmanned vehicle.

[0030] One embodiment of the above invention has the following advantages or beneficial effects: it can acquire spatial environment information of the current location of the autonomous vehicle, select a target scene template based on the spatial environment information, analyze the spatial environment information using the scene constraints included in the target scene template, regulate the autonomous vehicle based on the analysis results, and improve the flexibility and accuracy of controlling the autonomous vehicle by dynamically selecting the target scene template to regulate the autonomous vehicle according to the different spatial environment information during the vehicle's driving process.

[0031] The further effects of the aforementioned unconventional alternative methods will be explained below in conjunction with specific implementation methods. Attached Figure Description

[0032] The accompanying drawings are provided to better understand the invention and are not intended to unduly limit the scope of the invention. Wherein:

[0033] Figure 1 This is a flowchart illustrating a method for controlling an unmanned vehicle according to an embodiment of the present invention;

[0034] Figure 2 This is a structural schematic diagram of a lane driving scenario provided by an embodiment of the present invention;

[0035] Figure 3 This is a schematic diagram of a drivable area scenario provided by an embodiment of the present invention;

[0036] Figure 4 This is a schematic diagram of a device for controlling an unmanned vehicle according to an embodiment of the present invention;

[0037] Figure 5 This is an exemplary system architecture diagram in which embodiments of the present invention can be applied;

[0038] Figure 6 This is a schematic diagram of the structure of a computer system suitable for implementing terminal devices or servers of the present invention. Detailed Implementation

[0039] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0040] Autonomous decision-making capability is the core of autonomous driving technology. Current rule-based decision-making and vehicle control methods suffer from decision failures due to the high dimensionality and uncertainty of real-world traffic scenarios. Autonomous driving systems first analyze the safety of the action commands output by the current control method based on experience or rules. If a dangerous action is detected, such as failure to brake when close to the vehicle ahead, the system adjusts the current action commands according to the model or rules to avoid danger. However, this method of judging whether the reinforcement learning output meets safety requirements through rules or models limits the flexibility of reinforcement learning. Another method is to improve the reliability of self-learning strategies by adjusting the policy iteration process. This method considers sampling during policy iteration, but the pursuit of confidence significantly reduces the policy improvement speed, and the final performance of the strategy still depends entirely on the reinforcement learning result. Therefore, the reliability of the strategy may decrease due to the difficulty in estimating the performance of the reinforcement learning strategy.

[0041] In real-world environments, geospatial features are typically complex and varied, with differences in road curvature, width, and driving characteristics across different locations. Designing decision-making methods for autonomous driving requires a clear problem description, including state space and action space. Furthermore, the differences in geospatial features can render pre-designed or pre-trained strategies unsuitable for new geospatial environments, resulting in low versatility of the decision-making methods.

[0042] In view of this, such as Figure 1 As shown, this embodiment of the invention provides a method for controlling an unmanned vehicle, which may include the following steps:

[0043] Step S101: Obtain spatial environment information of the current location of the autonomous vehicle.

[0044] Specifically, controlling an autonomous vehicle requires acquiring real-time information about one or more of the controlled autonomous vehicles, namely, the spatial environment information of the current location. This spatial environment information includes one or more of the following: geographical range information, environmental object information, autonomous vehicle status, environmental object status, and driving constraints within the geographical range.

[0045] Specifically, the geographic range information includes the geographic location of the autonomous vehicle to be controlled, such as the lane it is in, the drivable area it is in, and the geographic information of its location; the environmental object information includes other objects besides itself that the autonomous vehicle to be controlled can perceive, such as other autonomous vehicles, other mobile devices, pedestrians, animals, obstacles, and traffic equipment; the autonomous vehicle status and the environmental object status can be running, stopped, etc.; the driving constraint elements of the geographic range include traffic rule elements associated with the geographic range, such as traffic lights, stop lines, speed limit signs, and pedestrians that suddenly approach.

[0046] Furthermore, the geographical range information of autonomous vehicles can be obtained through vehicle navigation network information. This information indicates the current location of the autonomous vehicle and the planned routes and connection point sequences that the vehicle needs to traverse to reach its destination. For autonomous driving, the vehicle navigation network information needs to provide highly detailed lane-level reference trajectories. The generation of these reference trajectories primarily utilizes graph search methods to generate paths based on high-precision maps, lane-level navigation maps, the vehicle's current location, and its destination.

[0047] Furthermore, the autonomous vehicle's environmental object information, its state, and the road structure information associated with the environment are all considered. It's understandable that the fundamental requirement for autonomous driving decision-making is to reach the target location without collision within the drivable area. However, in real-world road environments, the influence of geographically specific driving constraints (such as traffic rules) needs to be considered to constrain and control the autonomous vehicle's actions.

[0048] Furthermore, methods for acquiring spatial environment information can collect data from map modules, navigation modules, various types of sensors (e.g., radar) or devices (cameras, inertial navigation systems, etc.) contained in autonomous vehicles.

[0049] Step S102: Based on the spatial environment information and the scene constraints included in the various types of scene templates, select a target scene template for the autonomous vehicle from the various types of scene templates, wherein the scene template further includes scene constraints.

[0050] Specifically, various types of scene templates can be set for autonomous vehicles, and each type of scene template includes scene constraints. For example, when the scene template is a lane driving scene template, the scene constraints for the lane driving scene template are as follows: 1) There are one or more lanes in the geographical area (e.g., within the area corresponding to the location) where the autonomous vehicle to be controlled is located; 2) The sensors of the autonomous vehicle determine that the environmental objects within the perception range are vehicles, that is, only vehicles are driving in the lanes, and other environmental objects (e.g., pedestrians) are far away from the lanes; 3) One or more vehicles corresponding to the environmental objects are in normal driving state, that is, driving in the lanes or changing lanes, and there is no situation of stopping laterally in the middle of the road or driving across the lanes for a long time. Under the condition that the obtained spatial environmental information meets the above scene constraints, the target scene template used by the autonomous vehicle is determined to be the lane driving scene template. In addition to the lane driving scene template, there is also a drivable area scene template. If there are scene constraints that are not met, the drivable area scene template can be selected as the scene template. That is, selecting a target scene template for the autonomous vehicle from multiple types of scene templates includes: searching for the target scene constraints that the spatial environment information satisfies from the scene constraints included in the multiple types of scene templates; and selecting the target scene template to which the target scene constraints belong for the autonomous vehicle.

[0051] Further, the step of finding the target scene constraint conditions satisfied by the spatial environment information includes: matching the spatial environment information with one or more scene elements included in the scene constraint conditions; and determining the scene constraint conditions as the target scene constraint conditions when the spatial environment information successfully matches all the scene elements included in the scene constraint conditions. Specifically, based on the geographic range information, environmental object information, autonomous vehicle state, environmental object state, and geographic range driving constraint element information of the autonomous vehicle to be controlled, matching is performed with the scene elements included in the scene constraint conditions, and the scene constraint conditions are determined as the target scene constraint conditions based on the matching results. For example, if the conditions satisfying the scene constraint conditions of the lane driving scene template are determined based on the geographic range information, environmental object information, autonomous vehicle state, environmental object state, and geographic range driving constraint element information of the autonomous vehicle to be controlled, then the scene constraint conditions are determined as the target scene constraint conditions (i.e., the constraint conditions of the lane driving scene).

[0052] Furthermore, the geographical range information of autonomous vehicles is related to the road structure. For the lanes and drivable areas included in the road structure, corresponding scene elements can be set. For example, the definition of a lane is as shown in formula (1):

[0053]

[0054] in, This represents the information set for a lane, where i is the lane's number on the current road. The numbering rule is to start counting from 0 and increment sequentially; n id This indicates the road number of the current lane on the navigation map; the lane centerline point column corresponding to the lane centerline includes three parts of information, c k w represents the vector indicating the position of the center point of the lane centerline. k θ represents the lane width near that point. k v is the tangential direction angle at that point. min,k ,v max,k This represents the minimum and maximum speed limits near the centerline point of lane number k in the current lane. η s The (0,1) signal indicates whether the current vehicle needs to stop at the end of the road, η n This represents the target location that the vehicle needs to reach at the current end of the road. Lanes provide a method for constructing the state space of the environment, namely, constructing a state space and action space with the lane centerline as the reference frame. Since autonomous vehicles need to travel along lane lines, their action space will also be constrained by the lane lines.

[0055] Furthermore, for the drivable area scenario, the drivable area can be the drivable area outside the lane. The drivable area is defined by the area boundary, which can be formed by a series of points connected in sequence, with the points connected by straight lines, as shown in formula (2).

[0056]

[0057] Where, j k Connecting the points at the boundaries of the drivable area, each point is connected end-to-end (e.g., counter-clockwise) to form a drivable area. The drivable area constrains the state of the autonomous vehicle by limiting the boundary range of the vehicle's state. A state set can be set to represent the set of states of the autonomous vehicle within the drivable area. In particular, although the vehicle may be traveling in a lane in some situations, its decision-making logic is independent of the lane. For example, in scenarios where a pedestrian suddenly crosses the road and the autonomous vehicle needs to make emergency avoidance maneuvers, the vehicle does not need to consider lane constraints, but only the drivable area information. Lane information and drivable area information are usually stored in a map, so this static information can be directly obtained from the data provided by the map template, improving the efficiency of obtaining the geographic range information of the autonomous vehicle.

[0058] Furthermore, each type of scene template corresponds to a driving scenario; each type of scene template includes scene constraints, including: state space constraints containing specific state features of the corresponding driving scenario; and action space constraints containing a specific coordinate system of the corresponding driving scenario.

[0059] The driving scenarios of this invention are illustrated using lane driving scenarios and drivable area scenarios as examples. Specifically, embodiments of this invention provide lane driving scenario templates and drivable area scenario templates, meaning each type of scenario template corresponds to one driving scenario. Each type of scenario template includes scenario constraints, which include state space constraints for specific state characteristics of the driving scenario and action space constraints for a specific coordinate system of the corresponding driving scenario. The state space constraints and action space constraints are explained below:

[0060] For scenario constraints that include state space constraints for specific state features of the driving scenario, and for the case of a lane driving scenario, the scenario constraints for the lane driving scenario include state space constraints for specific state features of the lane driving scenario. For example, the state space definition (i.e., the state space constraints) for the lane driving scenario is shown in formula (3):

[0061] s T ={q e (t), {q i (t)}} (3)

[0062] in, It represents the driving state (i.e., specific state characteristics) of the vehicle (i.e., the driverless vehicle to be controlled) in a multi-lane coordinate system (e.g., the Frenet coordinate system); This represents the state of the surrounding environment and vehicles (i.e., specific state characteristics) in a multi-lane coordinate system. It is understandable that perceived environmental objects (vehicles) are not necessarily all within the lanes. Therefore, after calculating the multi-lane Frenet coordinate system, objects far from the lanes can be removed. Furthermore, based on the scenario constraints for determining the lane driving scenario, if pedestrians or other environmental objects are perceived within the surrounding lanes, the multi-lane template cannot be used to construct the decision problem.

[0063] For scenarios where the driving scenario is a drivable area, the constraints of the drivable area scenario include the constraints of the state space of the specific state characteristics of the driving scenario. For example, the state space definition of the drivable area scenario (i.e., the constraints of the state space) is shown in formula (4):

[0064]

[0065] Where, qe (t) represents the kinematic state of the vehicle (the unmanned vehicle to be controlled), q i (t) represents the kinematic state (i.e., specific state characteristics) of the environmental object. The kinematic state includes the vehicle's or environmental object's position, velocity, and attribute information (used to calculate the state transition equations), etc. n The target location of the vehicle represents the trajectory the vehicle needs to find within the drivable area to reach that point. For situations such as forced avoidance or road repairs, the target point c is adjusted accordingly. n This is used to issue action commands to the vehicle to control it. As shown in formula (2). This is a list of boundary points for the drivable area, representing the area where the vehicle can drive in the absence of environmental objects.

[0066] For the case of a lane-based driving scenario, the constraints of the action space in a specific coordinate system of the driving scenario are shown in Equation (5):

[0067]

[0068] Among them, a t This represents the lateral and longitudinal acceleration of the vehicle (the unmanned vehicle to be controlled) in the future time interval Δt, and can then be integrated to calculate the position and speed that the vehicle (unmanned vehicle) needs to reach after time Δt. The method for obtaining the lateral and longitudinal acceleration can be the data measured by the inertial navigation system; the coordinate system used is the Frenet coordinate system (i.e., a specific coordinate system); preferably, according to the definition of lane information as shown in formula (1), vehicles traveling in the lane need to ensure that their speed is within the specified range, and also need to decelerate according to traffic lights or stop lines when approaching the end of the road. Therefore, when generating control actions, the action space needs to be reduced. The reduction principles may include: according to the vehicle speed limit, all actions that cause the vehicle to exceed the speed limit or fall below the specified speed will be deleted from the action space; all actions that cause the vehicle to leave the lane boundary (such as crossing double yellow lines, or illegally entering the bicycle lane, etc.) will be deleted; if it is necessary to stop at the end of the road, all actions that prevent the vehicle from stopping at the end of the road will be deleted. This improves the effectiveness and accuracy of determining vehicle actions.

[0069] For the drivable area scenario, the constraints of the action space in the specific coordinate system of the drivable area scenario are shown in Equation (6):

[0070]

[0071] Among them, action a tThis represents the acceleration in the Cartesian coordinate system (i.e., a specific coordinate system) over a future time interval Δt. By integrating, the target position and velocity of the autonomous vehicle after time Δt can be calculated. The acceleration can be obtained from the data measured by the inertial navigation system.

[0072] Based on the geographical range information, environmental object information, autonomous vehicle status, environmental object status, and driving constraint elements of the geographical range described by formulas (1)-(6), Figure 2 A schematic diagram of a lane driving scenario is shown, in which a vehicle... Represents the driving state (i.e., specific state characteristics) of the vehicle (i.e., the autonomous vehicle to be controlled) in a multi-lane coordinate system (e.g., the Frenet coordinate system); environmental objects Represents the state (i.e., specific state characteristics) of the surrounding environment and vehicles in a multi-lane coordinate system; η n Represents a navigation constraint; η s This represents another type of navigation constraint (such as traffic information numbers). Figure 3 A schematic diagram of a drivable area scenario is shown; in which, Figure 2 , Figure 3 The "autonomous vehicle" mentioned herein refers to the unmanned vehicle controlled by the method of controlling unmanned vehicles described in this invention, in order to distinguish it from environmental objects.

[0073] Step S103: Analyze the spatial environment information based on the scene constraints included in the target scene template; adjust the unmanned vehicle according to the analysis results.

[0074] Specifically, the description of the scenario constraints for different types of driving scenarios is consistent with the description of step S102, and will not be repeated here.

[0075] Furthermore, the spatial environment information is analyzed, including:

[0076] Based on the spatial environment information and the constraints of the state space included in the target scene template, the state space of the autonomous vehicle is determined; based on the autonomous vehicle state included in the spatial environment information and the constraints of the action space included in the target scene template, the action space of the autonomous vehicle is determined; based on the state space of the autonomous vehicle and the action space of the autonomous vehicle, control commands are generated for the autonomous vehicle. Specifically, in the case of a lane driving scenario, based on spatial environment information and the constraints of the state space included in the target scenario template; when the target scenario template is a lane driving scenario template, the state space of the autonomous vehicle in the lane driving scenario is determined using the constraints of the state space as shown in formula (3); when the target scenario template is a drivable area driving scenario template, the state space of the autonomous vehicle in the drivable area driving scenario is determined using the constraints of the state space as shown in formula (4); further, based on the autonomous vehicle state included in the spatial environment information and the constraints of the action space included in the target scenario template, the action space of the autonomous vehicle is determined: for the case of a lane driving scenario, based on the autonomous vehicle state and the constraints of the action space as shown in formula (5), the action space of the autonomous vehicle in the lane driving scenario is determined; for the case of a drivable area driving scenario, based on the autonomous vehicle state and the constraints of the action space as shown in formula (6), the action space of the autonomous vehicle in the drivable area driving scenario is determined. Furthermore, based on the state space and action space of the autonomous vehicle, control commands are generated for the autonomous vehicle, such as: stop, decelerate (to speed 1), accelerate (to speed 2), etc.

[0077] Further, determining the state space of the autonomous vehicle includes: for the case of a lane driving scenario, converting the geographical range information included in the spatial environment information into Frenet coordinates, determining the Frenet coordinates of the autonomous vehicle, the Frenet coordinates of environmental objects, and the associated lanes included in the geographical range information, and determining the driving state of the autonomous vehicle and environmental objects in their respective lanes; specifically, still taking formula (3) as an example, s t ={q e (t),{q i(t)}}, the coordinate system for the lane driving scenario is constructed based on the multi-lane Frenet coordinate system. When using it, coordinates in the Cartesian plane coordinate system can be converted to coordinates along the lane direction (longitudinal) and perpendicular to the lane direction (lateral), that is, the geographical range information included in the spatial environment information is converted into Frenet coordinates. When using the Frenet coordinate system, the lateral coordinates of a vehicle traveling along the current lane remain unchanged, and only the longitudinal coordinates change. The lateral coordinate value of the vehicle indicates which lane the vehicle is in. That is, the Frenet coordinates of the autonomous vehicle, the Frenet coordinates of environmental objects, and the associated lanes included in the geographical range information are determined; thereby utilizing s t The driving status of the autonomous vehicle and environmental objects within their respective lanes is determined. Autonomous vehicles operating in the multi-lane Frenet coordinate system can disregard information such as road curvature and width, improving the accuracy of vehicle control and reducing its complexity.

[0078] For the case where the driving scenario is a drivable area driving scenario, the geographical range information included in the spatial environment information is converted into Cartesian coordinates. The Cartesian coordinates of the autonomous vehicle, the Cartesian coordinates of the environmental objects, the target point of the autonomous vehicle, and the boundary point series of the drivable area included in the geographical range information are determined, and the driving state of the autonomous vehicle in the drivable area is determined. Specifically, taking formula (4) as an example, the state space of the drivable area scenario template is defined in the Cartesian coordinate system. That is, the geographical range information included in the spatial environment information is converted into Cartesian coordinates. For example, according to the Cartesian coordinates of the autonomous vehicle, the Cartesian coordinates of the environmental objects, the target point of the autonomous vehicle, and the boundary point series of the drivable area indicated by formula (4), the state space of the drivable area scenario template is defined using s t Determine the driving status of the unmanned vehicle in the drivable area.

[0079] Further, determining the action space of the autonomous vehicle includes: in the case of lane driving scenario, using the Frenet coordinate system and the current driving state of the autonomous vehicle included in the spatial environment information to determine the driving state of the vehicle within a future set time range and the target position to be reached; specifically, the action space of the lane driving scenario is also defined in the multi-lane Frenet coordinate system, still taking formula (5) as an example, a tThis represents the current driving state of the autonomous vehicle, the lateral and longitudinal accelerations of the autonomous vehicle within a set time range of Δt in the future, and can then be integrated to calculate the position and speed (i.e., driving state) that the vehicle needs to reach after time Δt. In other words, the driving state of the vehicle and the target position to be reached within a set time range in the future are determined by using the Frenet coordinate system and the current driving state of the autonomous vehicle, which includes the spatial environment information.

[0080] In the scenario of driving within the drivable area, the driving state of the vehicle and the target location to be reached within a future set time range are determined using a Cartesian coordinate system and the current driving state of the autonomous vehicle, including the spatial environment information. Specifically, taking formula (6) as an example, a t This represents the acceleration within a set time range of Δt in the Cartesian coordinate system based on the current driving state of the autonomous vehicle. By integrating, the target position and velocity (i.e., driving state) after time Δt can be calculated.

[0081] Furthermore, based on the state space and action space of the autonomous vehicle, control commands are generated for the autonomous vehicle; wherein, the control commands may be: stop, move forward, turn left, turn right, turn around, etc.

[0082] Preferably, the actions performed by the autonomous vehicle based on control are obtained; if the actions performed by the autonomous vehicle meet the set action evaluation constraints, the target scenario template is rewarded according to a preset evaluation strategy; if the actions performed by the autonomous vehicle do not meet the set action evaluation constraints, the target scenario template is penalized according to a preset evaluation strategy; and the target scenario template is adjusted based on the reward and / or penalty results. Specifically, for the driving scenario corresponding to the lane driving template, the preset evaluation strategy of the lane driving template is set according to whether the vehicle meets the driving constraints, where the driving constraints are the action evaluation constraints, including safety constraints, navigation constraints, etc. The formula of the preset evaluation strategy is shown in formula (7) for example:

[0083]

[0084] Z sThe set of all states that satisfy the set action evaluation constraints is defined. If the actions executed by the autonomous vehicle based on the control do not satisfy the set action evaluation constraints (e.g., failure to reach the target location, execution of control commands inconsistent with preset commands, etc.), a penalty strategy is set, for example, making r(s) -1; otherwise, a reward strategy is set, making r(s) 0. The validity and accuracy of the lane driving template (target scenario template) are determined by calculating the reward or penalty results based on the target scenario template. For the driving scenario corresponding to the drivable area scenario template, its preset evaluation strategy is similar to formula (7), and the Z corresponding to the drivable area scenario template... s The state set represents the constraints of the motion evaluation, where the state set for motion evaluation can be a set of states indicating that the vehicle can reach the target point, a set of states indicating that the vehicle has not collided, a set of states indicating that the vehicle is driving within the drivable area, etc.

[0085] Furthermore, the target scene template is dynamically adjusted based on the real-time acquired spatial environment information. It is understood that during the operation of the autonomous vehicle, the method of the invention is used to acquire the spatial environment information of the space where the autonomous vehicle is located in real time, thereby dynamically adjusting the target scene template, that is, dynamically selecting either a lane driving template or a drivable area template, to achieve the effect of controlling the autonomous vehicle. Therefore, the method for controlling autonomous vehicles according to the present invention improves the versatility, flexibility, and efficiency of controlling autonomous vehicles.

[0086] like Figure 4 As shown, this embodiment of the invention provides a device 400 for controlling an unmanned vehicle, including: an information acquisition module 401, a template determination module 402, and a vehicle control module 403; wherein,

[0087] The information acquisition module 401 is used to acquire spatial environment information of the space where the unmanned vehicle is currently located.

[0088] The template determination module 402 is used to select a target scene template for the autonomous vehicle from multiple types of scene templates based on the spatial environment information and the scene constraints included in the multiple types of scene templates, wherein the scene template further includes scene constraints.

[0089] The vehicle control module 403 is used to analyze the spatial environment information based on the scene constraints included in the target scene template; and to control the unmanned vehicle according to the analysis results.

[0090] This invention also provides an electronic device for controlling an unmanned vehicle, comprising: one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method provided in any of the above embodiments.

[0091] This invention also provides a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the method provided in any of the above embodiments.

[0092] Figure 5 An exemplary system architecture 500 is shown, which can be applied to a method for controlling an autonomous vehicle or an apparatus for controlling an autonomous vehicle according to embodiments of the present invention.

[0093] like Figure 5 As shown, system architecture 500 may include terminal devices 501, 502, and 503, a network 504, and a server 505. Network 504 serves as the medium for providing communication links between terminal devices 501, 502, and 503 and server 505. Network 504 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0094] Users can use terminal devices 501, 502, and 503 to interact with server 505 via network 504 to receive or send messages, etc. Various client applications, such as map applications, navigation applications, and sensor applications, can be installed on terminal devices 501, 502, and 503.

[0095] Terminal devices 501, 502, and 503 can be various electronic devices with displays and supporting various client applications, including but not limited to smartphones, tablets, laptops, desktop computers, autonomous vehicles, etc.

[0096] Server 505 can be a server that provides various services, such as a backend management server that supports client applications used by users through terminal devices 501, 502, and 503. The backend management server can process the received spatial environment information data and feed the analysis results back to the terminal devices.

[0097] It should be noted that the method for controlling an unmanned vehicle provided in the embodiments of the present invention is generally executed by the unmanned vehicle 501, and correspondingly, the device for controlling the unmanned vehicle is generally installed in the unmanned vehicle 501; in addition, the method for controlling an unmanned vehicle provided in the embodiments of the present invention can also be executed by the server 505, which controls the unmanned vehicle by sending control commands. In this case, the device for controlling the unmanned vehicle can be installed in the server 505.

[0098] It should be understood that Figure 5 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0099] The following is for reference. Figure 6 It shows a schematic diagram of the structure of a computer system 600 suitable for implementing a terminal device of the present invention. Figure 6 The terminal device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0100] like Figure 6 As shown, the computer system 600 includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 602 or programs loaded from storage section 608 into random access memory (RAM) 603. The RAM 603 also stores various programs and data required for the operation of the system 600. The CPU 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0101] The following components are connected to I / O interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to I / O interface 605 as needed. A removable medium 611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 610 as needed so that computer programs read from it can be installed into storage section 608 as needed.

[0102] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 609, and / or installed from removable medium 611. When the computer program is executed by central processing unit (CPU) 601, it performs the functions defined above in the system of this invention.

[0103] It should be noted that the computer-readable medium shown in this invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0104] 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 the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that 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 consecutively indicated 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 a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0105] The modules and / or units described in the embodiments of the present invention can be implemented in software or hardware. The described modules and / or units can also be housed in a processor. For example, a processor can be described as including an information acquisition module, a template determination module, and a vehicle control module. The names of these modules do not necessarily limit the module itself; for example, the information acquisition module can also be described as "a module for acquiring spatial environment information of the space where the autonomous vehicle is currently located."

[0106] In another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiments; or it may exist independently and not assembled into the device. The computer-readable medium carries one or more programs, which, when executed by the device, cause the device to include: acquiring spatial environment information of the space currently occupied by the autonomous vehicle; selecting a target scene template for the autonomous vehicle from multiple types of scene templates based on the spatial environment information and scene constraints included in multiple types of scene templates, wherein the scene template further includes scene constraints; analyzing the spatial environment information based on the scene constraints included in the target scene template; and controlling the autonomous vehicle based on the analysis results.

[0107] The embodiments of the present invention can acquire spatial environment information of the current location of the autonomous vehicle, select a target scene template based on the spatial environment information, analyze the spatial environment information using the scene constraints included in the target scene template, and regulate the autonomous vehicle based on the analysis results. By dynamically selecting the target scene template to regulate the autonomous vehicle according to the different spatial environment information during the vehicle's driving process, the flexibility and accuracy of controlling the autonomous vehicle are improved.

[0108] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for controlling an unmanned vehicle, characterized in that, include: Obtain spatial environment information of the current location of the autonomous vehicle; Based on the spatial environment information and the scene constraints included in various types of scene templates, a target scene template is selected for the autonomous vehicle from various types of scene templates, wherein the scene template further includes scene constraints. Based on the scene constraints included in the target scene template, the spatial environment information is analyzed; and the autonomous vehicle is adjusted according to the analysis results. The scenario templates include lane driving scenario templates and drivable area scenario templates; if it is determined that the scenario constraints corresponding to the lane driving scenario template are not met, the drivable area scenario template is selected; wherein, the drivable area is the driving area outside the lane, and its decision logic is independent of the lane; The coordinate system of the lane driving scenario template is constructed based on the multi-lane Frenet coordinate system; the state space of the drivable area scenario template is defined in the Cartesian coordinate system.

2. The method according to claim 1, characterized in that, The space environment information includes one or more of the following: Geographic range information, environmental object information, autonomous vehicle status, environmental object status, and driving constraints within the geographic range.

3. The method according to claim 1, characterized in that, The step of selecting a target scene template for the autonomous vehicle from multiple types of scene templates includes: From the various types of scene templates, find the target scene constraints that the spatial environment information satisfies; Select the target scene template to which the target scene constraints belong for the autonomous vehicle.

4. The method according to claim 3, characterized in that, The target scene constraints that the search for the spatial environment information must satisfy include: Match the spatial environment information with one or more scene elements included in the scene limiting conditions; If the spatial environment information completely matches all the scene elements included in the scene limiting conditions, then the scene limiting conditions are determined to be the target scene limiting conditions.

5. The method according to claim 1, characterized in that, Each type of scene template corresponds to a specific driving scenario; Each type of scene template includes the following scene constraints: The constraints of the state space containing the specific state characteristics of the corresponding driving scenario; It includes constraints on the action space in a specific coordinate system corresponding to the driving scenario.

6. The method according to claim 5, characterized in that, The analysis of the spatial environment information includes: Based on the spatial environment information and the constraints of the state space included in the target scene template, the state space in which the autonomous vehicle is located is determined. Based on the autonomous vehicle status included in the spatial environment information and the constraints of the action space included in the target scene template, the action space of the autonomous vehicle is determined. Based on the state space and action space of the autonomous vehicle, control commands are generated for the autonomous vehicle.

7. The method according to claim 6, characterized in that, Determining the state space of the autonomous vehicle includes: For the case where the driving scenario is a lane driving scenario, the geographical range information included in the spatial environment information is converted into Frenet coordinates, the Frenet coordinates of the autonomous vehicle, the Frenet coordinates of the environmental objects and the associated lanes included in the geographical range information are determined, and the driving status of the autonomous vehicle and the environmental objects in their respective lanes is determined. For the case where the driving scenario is a driving scenario in a drivable area, the geographical range information included in the spatial environment information is converted into Cartesian coordinates. The Cartesian coordinates of the autonomous vehicle, the Cartesian coordinates of the environmental objects, the target point of the autonomous vehicle, and the boundary point list of the drivable area included in the geographical range information are determined, and the driving state of the autonomous vehicle in the drivable area is calculated.

8. The method according to claim 6, characterized in that, Determining the motion space of the autonomous vehicle includes: In the lane driving scenario, the driving status of the vehicle and the target location to be reached within a future set time range are determined by using the Frenet coordinate system and the current driving status of the autonomous vehicle, which includes the spatial environment information. In the scenario of driving in the drivable area, the driving status of the vehicle and the target location to be reached within a future set time range are determined by using the Cartesian coordinate system and the current driving status of the autonomous vehicle, which includes the spatial environment information.

9. The method according to claim 1, characterized in that, Further includes: Obtain the actions executed by the autonomous vehicle based on the control; Under the condition that the actions performed by the autonomous vehicle meet the set action evaluation constraints, the target scene template is rewarded according to the preset evaluation strategy. If the actions performed by the autonomous vehicle do not meet the set action evaluation constraints, the target scene template is penalized according to the preset evaluation strategy. The target scenario template is adjusted based on the results of the rewards and / or penalties imposed on it.

10. The method according to claim 1, characterized in that, Further includes: The target scene template is dynamically adjusted based on the real-time acquired spatial environment information.

11. A device for controlling an unmanned vehicle, characterized in that, include: The module comprises an information acquisition module, a template determination module, and a vehicle control module; among which, The information acquisition module is used to acquire spatial environment information of the space where the unmanned vehicle is currently located; The template determination module is used to select a target scene template for the autonomous vehicle from multiple types of scene templates based on the spatial environment information and the scene constraints included in the multiple types of scene templates, wherein the scene template further includes scene constraints. The vehicle control module is used to analyze the spatial environment information based on the scene constraints included in the target scene template; and to control the unmanned vehicle according to the analysis results. The scenario templates include a lane driving scenario template and a drivable area scenario template. If the scenario constraints corresponding to the lane driving scenario template are not met, the drivable area scenario template is selected. The drivable area is a driving area outside the lane, and its decision logic is independent of the lane. The coordinate system of the lane driving scenario template is constructed based on the multi-lane Frenet coordinate system. The state space of the drivable area scenario template is defined in the Cartesian coordinate system.

12. An electronic device, characterized in that, include: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-10.

13. A computer-readable medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-10.

Images