Vehicle obstacle avoidance method, computer readable storage medium and program product

Abstract

The invention provides a vehicle obstacle avoidance method, a computer readable storage medium and a program product. A vehicle is connected with an unmanned aerial vehicle. The method comprises the following steps: determining road information of a road according to the unmanned aerial vehicle under the condition of determining that an obstacle exists in the road on which the vehicle runs; driving state information of the vehicle is obtained, global drift feature points in a drift space are determined according to the driving state information and the road information, a drift path is planned according to the global drift feature points, and the drift space is used for representing a parameter feasible region of the vehicle for effectively executing drift action; and controlling the vehicle to run along the drift path so as to prevent the vehicle from colliding with the obstacle.

Description

Technical Field

[0001] This specification relates to the field of automobiles, and particularly to a vehicle obstacle avoidance method, a computer-readable storage medium, and a program product. Background Art

[0002] With the continuous development of intelligent driving technology, more and more manufacturers have successively assisted drifting or even automatic drifting functions. However, in extreme working conditions such as continuous U-shaped or S-shaped curves, due to the performance of forward perception devices such as millimeter-wave radars and monocular cameras, the vehicle's perception ability of curves is limited, and it is impossible to achieve accurate road condition prediction and corresponding drift path planning, resulting in a significant reduction in the vehicle's effective detection distance and forcing the vehicle to reduce speed in exchange for a safety margin, severely limiting the ability to achieve efficient obstacle avoidance through drifting. Therefore, how to improve the vehicle's obstacle avoidance ability based on drifting has become a difficult problem that the industry continues to solve. Summary of the Invention

[0003] In view of this, this specification provides a vehicle obstacle avoidance method, a readable storage medium, and a program product to solve the deficiencies in the related art.

[0004] Specifically, this specification is implemented through the following technical solutions:

[0005] According to the first aspect of this specification, a vehicle obstacle avoidance method is provided, and the vehicle is connected to a drone; the method includes:

[0006] In the case of determining that there is an obstacle in the road where the vehicle is traveling, determining the road information of the road according to the drone;

[0007] Obtaining the driving state information of the vehicle, and planning a drift path according to the driving state information and the road information;

[0008] Controlling the vehicle to travel along the drift path to avoid the vehicle colliding with the obstacle.

[0009] According to the second aspect of this specification, a vehicle obstacle avoidance device is provided, and the vehicle is connected to a drone; the device includes:

[0010] A road information determination unit, configured to determine the road information of the road according to the drone in the case of determining that there is an obstacle in the road where the vehicle is traveling;

[0011] A drift path planning unit, configured to obtain the driving state information of the vehicle, determine global drift feature points in the drift space according to the driving state information and the road information, and plan a drift path according to the global drift feature points, where the drift space is used to represent the parameter feasible region for the vehicle to effectively perform drift actions;

[0012] A vehicle control unit for controlling the vehicle to travel along the drift path to avoid collision between the vehicle and the obstacle.

[0013] According to the third aspect of this specification, a computer-readable storage medium is provided, on which a computer program is stored. When the program is executed by a processor, the steps of the method described in the first aspect are implemented.

[0014] According to the fourth aspect of this specification, a computer program product includes a computer program / instructions. When the computer program / instructions are executed by a processor, the steps of the method described in the first aspect are implemented.

[0015] In this application, the vehicle is interconnected with the drone, thereby realizing the collaborative perception of the two. Among them, when it is determined that there are obstacles in the road where the vehicle travels, the drone can break through the vision and detection distance limitations of the vehicle-mounted sensors, obtain the road information of the entire road where the vehicle travels, and at the same time, combine the driving state information of the vehicle to determine the global drift feature point from the drift space used to characterize the parameter feasible region for the vehicle to effectively execute the drifting action, and plan a more accurate drift path according to the global drift feature point, so that the vehicle can bypass the above-mentioned obstacles in an active drifting posture without significantly reducing the vehicle speed, thereby improving the obstacle avoidance success rate while significantly reducing the safety speed reduction requirements caused by blocked perception, and ultimately effectively improving the vehicle's obstacle avoidance ability based on drifting, making it applicable to more extreme working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of this specification, the drawings required for use in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of this specification. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0017] Figure 1 is a schematic diagram of the architecture of a vehicle obstacle avoidance system shown in an open embodiment of this specification;

[0018] Figure 2 is a flowchart of a vehicle obstacle avoidance method shown in an open embodiment of this specification;

[0019] Figure 3 is a schematic diagram of an obstacle avoidance curvature shown in an open embodiment of this specification;

[0020] Figure 4 is a schematic diagram of a global drift feature point and a local trajectory point shown in an open embodiment of this specification;

[0021] Figure 5 It is a flowchart showing the planning situation of a drift path disclosed in the embodiments of this specification;

[0022] Figure 6 It is a schematic structural diagram of an electronic device shown in the embodiments of this specification;

[0023] Figure 7 It is a block diagram of a vehicle obstacle avoidance device shown in the embodiments of this specification. Detailed implementation manners

[0024] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with this specification. On the contrary, they are merely examples of devices and methods consistent with some aspects of this specification.

[0025] The terms used in this specification are only for the purpose of describing specific embodiments and are not intended to limit this specification. The singular forms "a", "the" and "said" used in this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0026] It should be understood that although the terms first, second, third, etc. may be used in this specification to describe various information, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of this specification, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "when" or "while" or "in response to determining".

[0027] The embodiments of a vehicle obstacle avoidance method of this specification will be described in detail below with reference to the drawings.

[0028] Figure 1 It is a schematic architecture diagram of a vehicle obstacle avoidance system disclosed in the embodiments of this specification. As Figure 1 shown, the system may include a drone 11 and a vehicle 12.

[0029] The drone 11 serves as an aerial perception node of the vehicle obstacle avoidance system. It can collect road information in front of the vehicle 12 by carrying vision sensors such as multispectral cameras or lidar, including but not limited to the three-dimensional coordinates of obstacles, the curvature radius of curves, and the road adhesion coefficient. It can also achieve data transmission with the vehicle 12 through low-latency communication links such as 5G or Dedicated Short Range Communication (DSRC). Specifically, the flight altitude and viewing angle of the drone 11 can be dynamically adjusted to cover the blind spots of vehicle-mounted sensors, such as areas blocked by mountains or buildings within the curve where the vehicle is located, thereby ensuring that the vehicle 12 can obtain the obstacle distribution and road geometric parameters in advance before entering the curve, providing high-precision and high-frequency environmental information for drift path planning.

[0030] The vehicle 12, as a vehicle for performing drift obstacle avoidance, is equipped with a chassis domain controller and a corresponding dynamics calculation module. The dynamics calculation module can be built into the above-mentioned chassis domain controller in the form of logic software or physical hardware, or independently set outside the chassis domain controller and connected to each other. Therefore Figure 1 the specific position is not defined herein. In addition, the dynamics calculation module can plan a drift path based on the road information of the drone 11 and the driving state of the vehicle. Finally, the chassis domain controller can determine corresponding control instructions according to the above drift path to control the vehicle to drive along the above drift path, so as to avoid collisions between the vehicle and the above obstacles by means of drifting without the vehicle decelerating.

[0031] Before formally introducing the solution of this specification, it is necessary to combine Figure 3 with Figure 4 to give a unified introduction to the technical concepts involved in the vehicle during drifting, so as to facilitate the understanding of the subsequent content. Among them, Figure 3 there is an obstacle in the road where the vehicle in [Figure 14] is traveling. The possible subsequent driving paths of the vehicle can be represented by the curves K1 - K4 in the figure. If the vehicle maintains the original curvature represented by curve K1, it will inevitably collide with the above obstacle; at the same time, Figure 4 there is also an obstacle in the road where the vehicle in [Figure 16] is traveling. The vehicle avoids the obstacle based on the drift path.

[0032] Road curvature: The geometric bending degree of the road where the vehicle is traveling, such as Figure 3 the curvature represented by the road edge of the curve where the vehicle in [Figure 21] is traveling.

[0033] Obstacle avoidance curvature: The minimum curvature of the path required for the vehicle to avoid collisions with obstacles, which can represent the limit trajectory for the vehicle to maintain a safe distance from the obstacles, as shown in Figure 3As shown by curve K2. The path corresponding to this curvature needs to satisfy vehicle dynamics constraints, and the specific implementation methods vary according to vehicle characteristics: for example, for a conventional steering system, it is often only possible to achieve this by reducing the vehicle speed and adjusting the direction; while for a highly maneuverable vehicle, it can quickly pass through by combining automatic drifting technology, thus ensuring both obstacle avoidance ability and passing efficiency at the same time.

[0034] Maximum drifting curvature: The maximum curvature of a bend that a vehicle can stably pass through during drifting. During drifting, the drifting behavior is usually accompanied by sideslip and yaw, that is, both the sideslip angle β and the yaw angular velocity γ are non-zero, corresponding to the curvature characterized by curve K3 in Figure 3 Compared with the obstacle avoidance curvature corresponding to curve K2 in Figure 3 For the curvature based on which the vehicle travels, the distance between the vehicle and the obstacle can be greater than the safety distance.

[0035] Minimum drifting curvature: The minimum curvature of a bend that a vehicle can stably pass through during drifting, as shown by curve K4 in Figure 3 Since the curvature corresponding to curve K4 is less than the obstacle avoidance curvature corresponding to curve K2, if the vehicle drifts with the curvature corresponding to curve K4, it will inevitably collide with the obstacle.

[0036] Maximum conventional steering curvature: The maximum curvature of the path required for a vehicle to stably pass through an obstacle during normal driving, and the maximum conventional steering curvature is less than the minimum drifting curvature. In Figure 3 assuming that when the vehicle is driving normally, it can at most steer with the curvature characterized by curve K5 based on Automatic Emergency Steering (AES) to avoid an obstacle. Then, limited by the steering ability, even after avoiding the obstacle, it will not be able to change the steering in time, resulting in the vehicle finally colliding with the road edge or directly driving out of the lane, thus triggering accidents such as rollover.

[0037] Drifting path: That is, the path passed by the vehicle during drifting. The drifting curvature of this path is between the above maximum drifting curvature and minimum drifting curvature. For example, Figure 4 the drifting path of the vehicle in

[0038] Drifting starting point: A global drifting feature point in the drifting space, representing the initial pose of the vehicle when it triggers drifting from the normal driving state. At this time, the vehicle hardly experiences sideslip and yaw, that is, the sideslip angle β≈0 and the yaw angle γ≈0. For example, Figure 4 point A in

[0039] Steady-state drifting point: A global drifting feature point in the drifting space, representing the intermediate state of the vehicle maintaining drifting in the drifting path. The sideslip angle β and the corresponding expected value β s , the yaw angular velocity γ and the corresponding expected value γ sConsistent, at this time the vehicle maintains a fixed drift curvature, for example Figure 4 Points B and C in Figure 4 , and any point between point B and point C on the drift path.

[0040] Drift end point: A global drift feature point in the drift space, representing the target pose for the vehicle to exit the drift and resume normal driving. At this time, the vehicle hardly experiences sideslip and yaw, that is, β≈0, γ≈0. For example Figure 4 Point D in Figure 4 .

[0041] Figure 2 is a flowchart of a vehicle obstacle avoidance method shown in an exemplary embodiment disclosed in this specification. The vehicle is connected to a drone, and specifically may include the following steps:

[0042] Step 202, when it is determined that there is an obstacle on the road where the vehicle is traveling, determine the road information of the road according to the drone.

[0043] When it is detected that there is an obstacle on the vehicle's driving road, the road information can be obtained in real time through the drone, which may include vehicle information, obstacle information, road geometry information, and road surface condition information, so as to make up for the limitations of the field of view and detection distance of in-vehicle sensors. Among them, the above-mentioned drone can obtain the curved road image in the road through the camera image sensor, and then perform semantic analysis to determine the corresponding features, such as vehicle type, position, driving speed, etc. as the above-mentioned vehicle information; for another example, obstacle type, position, shape and size, obstacle avoidance curvature, etc. as the above-mentioned obstacle information; for another example, road length, road width, road curvature, etc. as the above-mentioned road geometry information; at the same time, the ground image information can be extracted, feature extraction can be performed, and then the adhesion coefficient of the ground can be analyzed and obtained through technologies such as artificial intelligence as the above-mentioned road surface condition information. Among them, the above-mentioned driving speed can be calculated and determined by the displacement distance of the vehicle in the images taken at preset intervals.

[0044] Of course, the above-mentioned obstacle can be determined by the in-vehicle sensor of the vehicle, or determined by the sensor of the drone, or comprehensively determined by combining the sensing results of both. This specification does not limit this.

[0045] Step 204, obtain the driving state information of the vehicle, determine the global drift feature points in the drift space according to the driving state information and the road information, and plan the drift path according to the global drift feature points. The drift space is used to represent the parameter feasible region for the vehicle to effectively perform drift actions.

[0046] When the drone determines the road information of the above-mentioned road, it can combine it with the driving state information of the vehicle to determine the global drift feature points in the drift space, and can generate a drift path based on the global drift feature points, so as to ensure that the vehicle in the subsequent steps can drift based on the drift path. Among them, the driving state information can be obtained or calculated by the vehicle through on-vehicle sensors, and specifically includes the minimum drift curvature and the maximum drift curvature theoretically supported by the vehicle at the corresponding time point, the actual adhesion coefficient of the ground, as well as the actual driving speed, sideslip angle, yaw angular velocity, and the activation state of the drift mode and other information.

[0047] Among them, the above-mentioned drift space refers to the parameter feasible region in which the vehicle can safely and stably perform drift actions under the constraints of vehicle dynamics and environmental conditions. It defines the set of trajectories and control inputs for each possible effective drift under a series of specific scenarios and vehicle states such as curve curvature and obstacle distribution. As a multi-dimensional parameter space, the above-mentioned drift space usually can include the following key dimensions: speed, sideslip angle, yaw angular velocity, front wheel steering angle, rear wheel speed, turning radius, road curvature. Of course, for the physical constraints of the drift space, the boundary of the drift space can be jointly determined by vehicle dynamics and environmental conditions. For example: as the limit of the lateral force and longitudinal force of the tire, the vehicle needs to be at the elliptical edge corresponding to the preset tire friction ellipse when drifting; the yaw angular velocity γ needs to avoid exceeding the preset yaw angular velocity threshold to avoid spinning; the drift trajectory needs to meet the minimum safety distance from the road boundary and obstacle avoidance, etc.

[0048] Specifically, the planning steps of the drift path can be further divided into two stages, that is, the first stage: drift space construction and feature point search, and the second stage: local trajectory generation and actual path synthesis. For the convenience of description, the following is simply referred to as the two stages of global planning and local planning.

[0049] In one embodiment, the drift space can be determined according to the above-mentioned driving state information and the above-mentioned road information, and the global drift feature points of the drift path can be searched in the drift space. Among them, the above-mentioned global drift feature points can include the drift start point, steady-state drift point and drift end point of the above-mentioned drift path, which can be regarded as the global planning process. At the same time, local trajectory points between different global drift feature points can be generated, and the above-mentioned drift path can be generated according to the above-mentioned global drift feature points and local trajectory points, which can be regarded as the local planning process.

[0050] Those skilled in the art can understand that according to the actual drift scenario of the vehicle, the content of the global drift feature points and the specific method of generating local trajectory points will also change accordingly.

[0051] For example, when the above-mentioned global drift feature points include the drift starting point, steady-state drift point and drift end point of the above-mentioned drift path, generating local trajectory points between different global drift feature points can be regarded as generating local trajectory points between the above-mentioned drift starting point and the above-mentioned steady-state drift point, and / or between the above-mentioned steady-state drift point and the above-mentioned drift end point.

[0052] For example, in an out-of-control drift scenario, that is, when the vehicle control algorithm fails or is affected by external disturbances such as slippery road conditions, causing the system's adjustment capabilities to be exceeded, the drift cannot converge to a steady-state point, resulting in continued instability until a collision or stop; or in a short-term drift scenario, that is, a short, transient drift action such as emergency obstacle avoidance ends before reaching dynamic balance, resulting in the absence of a steady-state drift point, or there may be no steady-state drift point; the above-mentioned global drift feature point only includes the drift start point and drift end point of the above-mentioned drift path, but there is no steady-state drift point. At this time, generating local trajectory points between different global drift feature points can be regarded as generating local trajectory points between the above-mentioned drift start point and the above-mentioned drift end point.

[0053] For global planning, before actually planning the drift path, it is possible to pre-determine whether the vehicle really needs to drift to avoid obstacles. The judgment result can be divided based on the relationship between the maximum drift curvature, the minimum drift curvature and the obstacle avoidance curvature, and then the following three situations can be obtained.

[0054] In the first case, when the obstacle avoidance curvature of the vehicle determined based on the driving state information and the road information is less than the minimum drift curvature, a conventional non-drifting path may be planned based on the driving state information and the road information. For example, a linear quadratic regulator (LQR) is used for lateral control, and a proportional-integral-derivative (PID) is used for longitudinal control. The non-drifting path is generated based on smooth sampling of a polynomial curve, thereby achieving obstacle avoidance without drifting.

[0055] Second case: When the obstacle avoidance curvature of the vehicle determined according to the above driving state information and the above road information is not less than the minimum drift curvature and not greater than the maximum drift curvature, the above drift path can be planned according to the driving state information and the road information; at this time, a drift space under vehicle dynamics constraints can be constructed based on the vehicle driving state characterized by the driving state information and the road information. In this drift space, global drift feature points can be searched through an optimization and sampling algorithm, including a drift start point, a steady-state drift point, and a drift end point. Based on the above global drift feature points, the drift trajectory to be planned can be divided to obtain three connected trajectories: the entry drift segment, the steady-state drift segment, and the exit drift segment. At the same time, the global drift trajectory can also be obtained by combining the three trajectories. Of course, the above drift space can be established in advance based on known vehicle parameters to reduce the real-time calculation complexity and achieve dynamic update when the environmental mutation and model error are greater than the error threshold, that is, the preset parameters are continuously corrected by means of, for example, online learning or adaptive control to ensure drift stability.

[0056] It can be understood that the above drift space can be constructed based on the following Formulas 1-7:

[0057]

[0058] αf = actan(β + lf * r / V) - δf; (4)

[0059] γ = V * rho; (5)

[0060] Fyf = Fzf * μ * D * sin(C * actan(B * αf)); (6)

[0061]

[0062] Regarding Formula 1, as the yaw rate steady-state balance equation, it describes the balance of the yaw moment when the vehicle is in steady-state drift, that is, the change rate of the yaw rate is zero, which is used to constrain the distribution relationship of the lateral forces of the front and rear wheels during the drift process to ensure that the vehicle maintains a stable yaw rate γ. Among them, the front-wheel lateral force Fyf generates a positive yaw moment through the force arm Lf, and the rear-wheel lateral force Fyr generates a negative yaw moment through the force arm Lr. δ is the front-wheel steering angle, Lr is the distance from the center of mass to the rear axle, and Iz is the moment of inertia of the vehicle about the vertical axis, which is used to normalize the moment effect.

[0063] Regarding Formula 2, as the center-of-mass side slip angle steady-state balance equation, it describes the balance of the center-of-mass side slip angle β when the vehicle is in steady-state drift, that is, the change rate of the side slip angle Is zero, used to constrain the balance between the lateral force and the inertial force during the drift process, ensuring the stability of the center-of-mass sideslip angle β. Among them, the front-wheel lateral force Fyf is projected onto the vehicle coordinate system based on the geometric relationship between the front-wheel steering angle δf and the sideslip angle β, and mV is the product of the vehicle mass and speed, representing the inertial force.

[0064] Regarding Equation 3, which is the longitudinal velocity steady-state balance equation, it describes the balance of the longitudinal velocity of the vehicle during steady-state drift, and the longitudinal acceleration Is zero, used to constrain the balance of the longitudinal force during the drift process, ensuring the stability of the vehicle speed V. Among them, the longitudinal component of the front-wheel lateral force Fyf is determined based on the geometric relationship between the front-wheel steering angle δf and the sideslip angle β, and the longitudinal components of the rear-wheel lateral force Fyr and the resistance component of the rear-wheel longitudinal force Fxr are both determined based on the sideslip angle β.

[0065] Regarding Equation 4, which is the geometric relationship of the front-wheel sideslip angle, it defines the geometric origin of the front-wheel sideslip angle αf as the center-of-mass sideslip angle β, the additional sideslip angle lf*r / V generated by the yaw rate γ at the front axle, and the actual front-wheel steering angle δf. lf is the distance from the center of mass to the front axle, used to relate the vehicle motion state (β, γ, V) to the front-wheel sideslip angle αf and provide input for tire force calculation.

[0066] Regarding Equation 5, which is the kinematic relationship between the yaw rate and the curvature, it defines the direct relationship between the yaw rate γ, the vehicle speed V, and the path curvature ρ. When the vehicle travels along a path with curvature ρ at speed V, the yaw rate γ is equal to Vρ. It is used to couple the curvature ρ involved in path planning with the yaw rate γ involved in vehicle dynamics to ensure the consistency between the drift trajectory and vehicle motion.

[0067] Regarding Equation 6, which is the magic formula model of the front-wheel lateral force, it is a front-wheel lateral force model based on the Magic Formula, used to describe the non-linear relationship between the front-wheel lateral force Fyf and the sideslip angle αf. This relationship involves the front-wheel vertical load Fzf, the road surface adhesion coefficient μ, the stiffness factor B, the shape factor C, and the peak factor D. It is used to calculate the front-wheel lateral force, provide dynamic input for Equations 1-3, and constrain the physical limit of the tire force during the drift process.

[0068] Regarding Equation 7, which is the friction ellipse constraint of the rear-wheel lateral force, it provides a rear-wheel lateral force model based on the friction ellipse theory. Among them, the rear-wheel lateral force Fyr is limited by the maximum friction force between the tire and the ground (Fzrμ), and the influence of the rear-wheel longitudinal force Fxr needs to be deducted. It is used to constrain the coupling relationship between the rear-wheel longitudinal force and the lateral force to ensure that the tire force does not exceed the physical limit during the drift process.

[0069] In summary, Equations 1-3 define the steady-state drift conditions, namely, constant yaw rate, sideslip angle, and longitudinal velocity, which in turn constitute the dynamic boundaries of the drift feasible domain. Equations 4-5 establish a relationship between vehicle kinematics and geometric parameters, mapping path planning to vehicle states. Equations 6-7, based on the tire characteristic model, constrain the physical limits of tire forces during drift, ensuring the feasibility of the drift action. By combining Equations 1-7, given the curvature ρ, adhesion coefficient μ, and vehicle speed V, the vehicle state that satisfies steady-state drift can be solved, thereby constructing a multidimensional feasible domain of the drift state space.

[0070] In the third scenario, if the vehicle's obstacle avoidance curvature, determined based on the driving state information and the road information, exceeds the maximum drift curvature, the vehicle's speed may be reduced and the drift path replanned. Specifically, the vehicle's speed may be gradually reduced in fixed or dynamic steps. The vehicle dynamics model is then used to verify whether the reduced speed satisfies the drift conditions. If so, the drift trajectory is replanned based on the reduced speed, achieving the second scenario and resuming further deceleration.

[0071] In the above embodiment, through layered curvature threshold determination, this method achieves dynamic optimization of the vehicle's obstacle avoidance strategy and meets multi-dimensional vehicle driving needs. For example, in conventional curve scenarios with low curvature, the non-drift mode is preferentially adopted to reduce tire wear and energy consumption; in sharp curves with medium to high curvature or scenes with dense obstacles, the drift mode is used to maximize the utilization of the vehicle's dynamic limits to avoid driving efficiency losses caused by excessive deceleration; in extreme curve scenarios with ultra-high curvature, active deceleration is used to expand the drift feasibility boundary to ensure the safety of obstacle avoidance actions. The above multiple strategies improve the obstacle avoidance success rate and vehicle motion performance in complex road scenarios, while avoiding the risks of low driving efficiency or vehicle instability caused by the traditional method due to a single control mode.

[0072] For the specific drift space search process mentioned above, the random sampling capability of the Rapidly-exploring Random Tree Star (RRT*) algorithm and / or the optimization capability of the Sequential Quadratic Programming (SQP) algorithm can be used to find feasible solutions in the drift space to serve as the above-mentioned global drift feature points.

[0073] The following combination ​ The logic of the vehicle obstacle avoidance method for the above three situations is explained, such as ​ As shown, the method includes the following steps:

[0074] Step 502: Determine the maximum drift curvature and the minimum drift curvature of the vehicle according to the driving state information, and determine the obstacle avoidance curvature according to the road information.

[0075] In one embodiment, after the chassis domain controller of the vehicle receives the road information from the drone, it can hand over the driving state obtained by itself and the vehicle's own sensors to the vehicle dynamics calculation module for offline calculation in advance, so as to obtain the drift space. Under the determined adhesion coefficient, based on the known vehicle speed and curvature information, the sideslip angle β and yaw rate γ of other drift states can be obtained;

[0076] At the same time, based on the vehicle dynamics model, calculate the maximum drift curvature and minimum drift curvature at the current vehicle speed and adhesion coefficient, and calculate the obstacle avoidance curvature in combination with the obstacle position and road geometry information.

[0077] Step 504, determine whether the obstacle avoidance curvature is greater than the maximum drift curvature.

[0078] In one embodiment, the vehicle dynamics calculation module can compare the calculated obstacle avoidance curvature with the maximum drift curvature. If the judgment result is yes, it means that the drift ability at the current vehicle speed is not enough to avoid obstacles, and step 506 is executed; if the judgment result is no, step 508 is executed to further determine whether to enter the drift mode.

[0079] Step 506, reduce the vehicle speed.

[0080] In one embodiment, the chassis domain controller of the vehicle can reduce the vehicle speed by: sending a braking instruction to a braking system such as the Electronic Stability Program (ESP), adjusting the output of a power system such as a motor or an engine, etc., while ensuring that the deceleration process complies with the vehicle stability constraint, and re-trigger step 502 after deceleration to update the curvature parameters to verify the drift feasibility.

[0081] Step 508, determine whether the obstacle avoidance curvature is greater than the minimum drift curvature.

[0082] In one embodiment, the above-mentioned dynamics calculation module compares the obstacle avoidance curvature K2 with the minimum drift curvature K4 to ensure that the decision complies with the vehicle dynamics limit. Specifically, if the judgment result is yes, it means that the drift can cover the obstacle avoidance requirement under the current conditions, and step 512 is executed; if the judgment result is no, it means that there is no need to drift and normal steering can avoid obstacles, and step 510 is executed.

[0083] Step 510, plan a non-drift path according to the driving state information and road information.

[0084] In one embodiment, the above-mentioned chassis domain controller can generate a smooth steering trajectory based on LQR, adjust the vehicle speed based on PID to match the obstacle avoidance curvature requirement, and output a polynomial curve as the trajectory to ensure steady-state driving with β≈0 and γ≈0. This result can be sent to the steering and power execution system via the Controller Area Network (CAN) bus.

[0085] Step 512: Search for global drift feature points in the drift space through an optimization and sampling algorithm.

[0086] In one embodiment, the above-mentioned dynamics calculation module can randomly sample drift states such as V, β, γ, δ, etc. based on the drift space preset in step 502 to screen candidate points that meet the obstacle avoidance curvature and road boundaries, and optimize the front wheel steering angle δ and the rear wheel driving force Fxr of the candidate points to ensure that they satisfy the dynamic balance of the above formulas 1-3. Finally, the drift starting point [V = V0, β = 0, γ = 0; X = X0, Y = Y0, Ψ = Ψ0], the steady-state drift point [V = V0, β = βdrift, γ = γdrift; X = Xdrift, Y = Ydrift, Ψ = Ψdrift], and the drift ending point [V = V0, β = 0, γ = 0; X = Xend, Y = Yend, Ψ = Ψend] are obtained as global drift feature points. Among them, for the drift starting point, the vehicle enters the drift at the initial vehicle speed. Assuming that the vehicle speed remains constant during the drift, i.e., V = V0, the vehicle state before the drift trigger, at this time the vehicle is in steady-state driving without side slip and yaw, i.e., β = 0, γ = 0, the global coordinates of the drift starting position are X0, Y0, and the heading angle is Ψ0; for the steady-state drift point, assuming that the vehicle maintains the initial vehicle speed during the drift, i.e., V = V0, the vehicle is in a stable state during the drift, and the side slip angle and yaw angular velocity maintain dynamic balance, i.e., β = βdrift, γ = γdrift, the intermediate point in the drift path, corresponding to the vehicle centroid trajectory Xdrift, Ydrift, and the heading Ψdrift; for the drift ending point, the vehicle resumes the initial vehicle speed when the drift ends, i.e., V = V0, the steady-state driving state after the drift exits, without side slip and yaw, i.e., β = 0, γ = 0, the target position and heading angle at the end of the drift need to be aligned with the road geometry, for example, returning to the center of the lane after exiting the curve, and its global coordinates are Xend, Yend, and the heading angle is Ψend.

[0087] Step 514: Determine the global drift trajectory based on the global drift feature points.

[0088] In one embodiment, the chassis domain controller can segment the global drift feature points to obtain the drift segment, the steady-state drift segment, and the drift exit segment, and convert the segmented trajectory into specific steering, braking, and driving force commands, and implement the drift action through the actuator.

[0089] Those skilled in the art can understand that the actual determination process of the above-mentioned global drift feature points can be implemented by either the vehicle or the drone. That is, the vehicle can receive road information from the drone and calculate the above-mentioned drift space based on its own acquired driving state information and the road information, and search for the above-mentioned global drift feature points in the drift space; or, first, the drone receives the driving state information from the above-mentioned vehicle, calculates the above-mentioned drift space based on its own acquired road information and the above-mentioned driving state information, and searches for the above-mentioned global drift feature points in the drift space, and then the above-mentioned vehicle receives the global drift feature points obtained by the drone. The difference between the two lies in the allocation difference of computing resources. If the vehicle directly determines the global drift feature points, its built-in high-performance computing unit can be used to calculate the dynamic constraints in real time, but this depends on the low-latency communication of the drone to ensure the synchronous update of road information; if the drone determines the global drift feature points and then sends them to the vehicle, the environmental data and the received vehicle state can be directly fused from the aerial perspective, reducing the communication frequency, but limited by the on-board computing power, a lightweight model is usually used for approximate calculation. The two modes can be dynamically switched according to the network conditions and task requirements. For example, when avoiding obstacles emergently, the vehicle side is given priority to calculate the drift space and search for the global drift feature points to reduce the latency, while in the pre-planning of complex scenarios, the drone side calculates the drift space and searches for the global drift feature points offline to share the load. Of course, the driving state information of the vehicle can be further estimated within a preset error range from the data obtained by the sensors of the drone, and then uniformly transmitted to the vehicle after the global drift feature points are determined. This means that the above-mentioned steps of global planning are completely independently implemented by the drone, maximizing the reduction of the computing requirements on the vehicle side.

[0090] For local planning, based on the previously obtained global drift feature points, feedback linearization or kinematic recurrence methods can be used to generate a sequence of local trajectory points between the drift starting point and the steady-state drift point, and between the steady-state drift point and the drift ending point, and ensure the curvature continuity and dynamic feasibility of adjacent trajectory points; finally, the local trajectory points of each segment are spliced in chronological order to form a complete drift path. Specifically, in combination with ​ , after the vehicle obtains the global drift feature points A, B, C, and D, for the segment from normal driving to the drift starting point, that is, the segment before A, a fifth-degree polynomial curve can be used for fitting, for example, to connect the non-drift segment before the drift starting point and the drift segment after the drift starting point. For the segment from the drift starting point to the steady-state point, that is, the segment AB, and the segment from the steady-state point to the drift ending point, that is, the segment CD, linearized transition or kinematic recurrence methods can be used to splice the trajectory points of each segment. For the segment from the drift terminal to the subsequent driving segment, that is, the segment before D, a fifth-degree polynomial curve similar to the segment before A can also be used to restore to the steady-state control mode.

[0091] For example: In a U-shaped bend, when the vehicle detects an obstacle ahead and needs to avoid it urgently, the drift starting point (vehicle speed 80 km / h, β = 0°), the steady-state drift point (β = 15°, γ = 20° / s), and the drift ending point (β = 0°) are determined through global drift feature points; the local trajectory generation stage is divided into four segments: 1. Regular driving to the drift starting point: A fifth-degree polynomial curve fitting is adopted, with the initial curvature constrained to the current road curvature and the end curvature to the drift starting point curvature to ensure a smooth transition to the drift trigger state; 2. From the drift starting point to the steady-state drift point, based on the forward recursion of the vehicle dynamics model, the vehicle pose (X, Y, Ψ) and state variables (β, γ) are predicted every 0.05 seconds, and infeasible trajectory points are eliminated by combining with the friction ellipse constraint; 3. From the steady-state point to the drift ending point, the dynamic recursion method of the segment from the drift starting point to the steady-state drift point is followed, maintaining a steady drift with β = 15° and γ = 20° / s until β and γ are gradually reduced when approaching the end point; 4. From the drift ending point to subsequent driving, a fifth-degree polynomial curve is adopted, with the end curvature constrained to be consistent with the road curvature after exiting the bend, and the lateral acceleration change rate ≤ 1.5 m / s 3 , and at the same time, the Clothoid curve can be used for transition between each segment to ensure curvature continuity and control command smoothness.

[0092] In summary, the above process of global planning and local planning can define the vehicle control boundary through the drift space, and through the segmented optimization of global feature points and local trajectories, it can quickly generate a drift path that takes into account obstacle avoidance requirements and dynamic stability in complex road scenarios, avoiding the problem of insufficient real-time performance caused by excessive global planning calculation load in traditional methods. At the same time, through the smooth transition of segmented trajectories, the risk of sudden change in lateral acceleration during the drift process is significantly reduced.

[0093] Step 206, control the vehicle to travel along the drift path to avoid the vehicle colliding with the obstacle.

[0094] After planning the drift path, the drift trajectory can be executed through the vehicle chassis domain controller, and the steering angle and drive torque are adjusted in real time to control the vehicle to drift around the obstacle in an active skidding posture while maintaining driving stability.

[0095] In addition to the above local trajectory points, based on the target states in the global drift feature points, namely position Xref, Yref, heading angle Ψref, vehicle speed Vref, sideslip angle βref, and yaw rate γref, a vehicle state sequence and control commands between the drift start point and the drift end point can also be generated. Furthermore, based on the above vehicle state sequence and control commands, the vehicle can be controlled to travel along the drift path. Corresponding to the previous global planning and local planning, the above steps can be referred to as trajectory tracking control during the drift process based on the drift path. Among them, the vehicle state sequence can be generated through timestamp interpolation or kinematic recursion methods. This sequence contains the expected state quantities at discrete time nodes from the drift start point to the end, ensuring that parameters such as the curvature change rate and lateral acceleration between adjacent nodes meet the vehicle dynamics constraints. At the same time, according to the actual state of the current vehicle, that is, the deviation between the position Xact, heading angle Xact, Yact, heading angle Ψref, vehicle speed Vact, sideslip angle βact, yaw rate γact and the state sequence, a model predictive control (Model Predict Control, MPC) and feedforward-feedback execution control method are adopted to calculate the control commands in real time, namely the target front wheel steering angle and rear wheel torque, so that the vehicle approaches the drift path with a preset sideslip angle βref and yaw rate γref, that is, to ensure that the vehicle accurately tracks the drift path in real time.

[0096] In one embodiment, the front wheel steering angle adjustment amount and the rear wheel speed adjustment amount of the vehicle during driving can be determined based on model predictive control. At the same time, the current output torque of the vehicle is calculated according to the front wheel steering angle adjustment amount and the rear wheel speed adjustment amount to drive the vehicle to maintain drifting along the above drift path. Specifically, during the process of controlling the vehicle to travel along the drift path, the closed-loop tracking of the drift attitude can be realized through the model predictive control framework. Based on the real-time vehicle states including the sideslip angle β, yaw rate γ, vehicle speed V, etc. and the above trajectory point sequence of the drift path represented by X_ref, Y_ref, and Ψ_ref, a rolling optimization problem including the vehicle dynamics model and tire force constraints can be constructed within the prediction time domain, with the main goal of minimizing the trajectory tracking errors (lateral deviation e_L, heading deviation e_Ψ, sideslip angle deviation e_β, yaw rate eγ), adjusting the control weights of the front wheel steering angle and the rear wheel speed, and within the ability boundaries and response performance of the steering system and the drive and braking systems, in an augmented model predictive control manner, the optimal current front wheel steering angle control increment Δδ and rear wheel speed increment ΔW are solved in real time. According to the rear wheel target speed Wtarget = Wact + ΔW and the current vehicle state including the longitudinal vehicle speed, lateral vehicle speed, and yaw rate, the slip ratio kappa is calculated, and through the longitudinal-lateral coupling tire model, the rear wheel longitudinal force Fxr is calculated. The specific formula is as follows:

[0097] Fxr = Fzr * μ * D * sin(C * actam(B * kappa)) (8)

[0098] Combined with the PID feedback control of the target wheel speed and the actual wheel speed Wact, calculate the rear wheel end torque Twhl, and the specific formula is as follows:

[0099]

[0100] In summary, through the multi-objective optimization and rolling horizon correction of model predictive control, the above method can dynamically balance the trajectory tracking accuracy and vehicle stability requirements during the drifting process, significantly reducing the risk of out-of-control caused by tire force saturation or path mutation. Compared with traditional PID control, the above MPC can compensate for the drift attitude deviation in advance by explicitly considering the dynamic constraints and future state evolution, ensuring the controllability and safety of the vehicle in complex obstacle avoidance scenarios.

[0101] It is worth mentioning that in the scenario where the vehicle cooperates with multiple unmanned aerial vehicles, the above method further improves the road perception ability through a multi-unmanned aerial vehicle information fusion mechanism.

[0102] In one embodiment, when the vehicle is connected to multiple unmanned aerial vehicles, receive and splice the extended road information determined by at least a part of the multiple unmanned aerial vehicles, and the road range and / or road accuracy represented by the extended road information are greater than the road information determined by any one of the at least a part of the unmanned aerial vehicles. Among them, the vehicle can receive the local road information collected by at least two unmanned aerial vehicles respectively, and through the methods of time-space synchronization and coordinate system alignment, such as based on the Global Positioning System (GPS) timestamp and the Simultaneous Localization and Mapping (SLAM) feature point matching, fuse the multi-view data into extended road information. Specifically, the observation results of each unmanned aerial vehicle can be integrated through image stitching techniques such as three-dimensional point cloud registration or Scale-Invariant Feature Transform (SIFT) feature matching implemented by, for example, the Iterative Closest Point (ICP) algorithm, so that the extended road information covers a wider road range, such as a scenario where unmanned aerial vehicle A covers the entrance of a curve and unmanned aerial vehicle B covers the exit, and improves the perception accuracy of key areas such as the edges of obstacles. In addition, according to the real-time position and heading of the vehicle, the subset of unmanned aerial vehicles participating in the fusion can also be dynamically screened, such as the 3 unmanned aerial vehicles closest to the future trajectory of the vehicle, to balance the computational load and information quality. For the overlapping observation areas of multiple unmanned aerial vehicles, the data conflict can be eliminated by using the Kalman filter or weighted average algorithm, while for the blind areas, the information gaps can be filled by the complementary perspectives of multiple unmanned aerial vehicles.

[0103] This application can also improve the positioning robustness of the vehicle in complex scenarios based on a multi-source positioning fusion mechanism.

[0104] In one embodiment, the vehicle can determine its first positioning information based on the visual sensor of the drone; meanwhile, the vehicle is positioned according to the above first positioning information and / or the second positioning information determined by the vehicle. Among them, through the visual sensor carried by the drone, environmental features around the vehicle such as lane lines, traffic signs, and static obstacles can be captured in real time, and the first positioning information of the vehicle relative to the environmental coordinate system can be calculated based on visual odometry or feature matching algorithms, and its positioning accuracy can reach the centimeter level, which is especially suitable for providing continuous pose output in scenarios where the Global Navigation Satellite System (GNSS) fails. The second positioning information built into the vehicle can be generated by tightly coupled filtering fusion with the Inertial Measurement Unit (IMU), and the global coordinates and heading angle of the vehicle are output; the vehicle preferentially uses the second positioning information in open scenarios, while in scenarios with signal occlusion or high-dynamic scenarios with drift, it dynamically switches to the first positioning information or the fusion result of the two according to the confidence weight. Thus, the positioning reliability of the vehicle during the drift obstacle avoidance process is enhanced.

[0105] In addition, when the road information includes a road image, the vehicle can receive an expected drift trajectory drawn based on the above road image and control the vehicle to travel along the above expected drift trajectory to complete the drift action. Specifically, the vehicle can extract the bend boundary, obstacle contour, and drivable area through a semantic segmentation network based on the road image transmitted back by the drone, and reconstruct the three-dimensional geometry of the road; meanwhile, according to the vehicle dynamics constraints and the real-time vehicle speed, an expected drift trajectory is generated. Obviously, this method directly drives the generation and control of the drift trajectory through visual information, significantly improving the obstacle avoidance response speed and accuracy in complex scenarios. In addition, the above road image can also be used for the analysis and evaluation of drift trajectories in drift events, etc., and the application thereof in this specification is not limited.

[0106] ​ It is a schematic structural diagram of an electronic device in an exemplary embodiment. Please refer to ​, at the hardware level, the electronic device includes a processor 602, an internal bus 610, a network interface 604, a memory 606, and a non-volatile memory 608. Of course, it may also include other required hardware. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a detection device for risk codes at the logical level. Of course, in addition to the software implementation method, this specification does not exclude other implementation methods, such as logical devices or a combination of software and hardware, etc. That is to say, the execution subject of the following processing flow is not limited to each logical unit, and can also be hardware or logical devices.

[0107] ​ The block diagram of a vehicle obstacle avoidance device shown in the embodiments of this specification. Please refer to ​ , this device can be applied to devices such as ​ shown, to implement the technical solutions described in this specification, the vehicle is connected to the unmanned aerial vehicle; the device includes:

[0108] A road information determination unit 702, configured to determine the road information of the road according to the unmanned aerial vehicle when it is determined that there is an obstacle in the road where the vehicle is traveling;

[0109] A drift path planning unit 704, configured to obtain the driving state information of the vehicle, determine global drift feature points in the drift space according to the driving state information and the road information, and plan a drift path according to the global drift feature points, where the drift space is used to represent the parameter feasible region for the vehicle to effectively perform drift actions;

[0110] A vehicle control unit 706, configured to control the vehicle to travel along the drift path to avoid the vehicle colliding with the obstacle.

[0111] Optionally, the device further includes:

[0112] A local trajectory point generation unit, configured to generate local trajectory points between different global drift feature points, and generate the drift path according to the global drift feature points and the local trajectory points.

[0113] Optionally, the local trajectory point generation unit is specifically configured to:

[0114] When the global drift feature points include the drift start point and the drift end point of the drift path, generate local trajectory points between the drift start point and the drift end point;

[0115] When the global drift feature points include the drift start point, the steady-state drift point, and the drift end point of the drift path, generate local trajectory points between the drift start point and the steady-state drift point, and / or between the steady-state drift point and the drift end point.

[0116] Optionally, the vehicle control unit 706 is specifically configured to:

[0117] Generate a vehicle state sequence and control instructions between the drift start point and the drift end point according to the global drift feature points;

[0118] Control the vehicle to travel along the drift path based on the vehicle state sequence and the control instructions.

[0119] Optionally, the drift path planning unit 704 is specifically configured to:

[0120] Calculate the drift space according to the driving state information and the road information, and search for the global drift feature points in the drift space; or,

[0121] Receive the global drift feature points from the UAV, where the global drift feature points are obtained by the UAV calculating the drift space according to the road information and the driving state information acquired from the vehicle, and searching in the drift space.

[0122] Optionally, the device further includes any one of the following:

[0123] A non-drift path planning unit, configured to plan a non-drift path according to the driving state information and the road information when the obstacle avoidance curvature of the vehicle determined according to the driving state information and the road information is less than the minimum drift curvature;

[0124] A second drift path planning unit, configured to plan the drift path according to the driving state information and the road information when the obstacle avoidance curvature of the vehicle determined according to the driving state information and the road information is not less than the minimum drift curvature and not greater than the maximum drift curvature;

[0125] A drift path replanning unit, configured to reduce the vehicle speed and replan the drift path when the obstacle avoidance curvature of the vehicle determined according to the driving state information and the road information is greater than the maximum drift curvature.

[0126] Optionally, the device further includes:

[0127] A multi-UAV expansion unit, configured to receive and splice the expanded road information determined by at least a part of the multiple UAVs when the vehicle is connected to multiple UAVs, where the road range and / or road accuracy represented by the expanded road information is greater than the road information determined by any one of the at least a part of the UAVs.

[0128] Optionally, the device further includes:

[0129] A drone positioning unit, configured to determine first positioning information of the vehicle according to a vision sensor of the drone;

[0130] Position the vehicle according to the first positioning information and / or second positioning information determined by the vehicle.

[0131] Optionally, the vehicle control unit 706 is specifically configured to:

[0132] Determine an adjustment amount of the front-wheel steering angle and an adjustment amount of the rear-wheel rotational speed of the vehicle during driving based on model predictive control;

[0133] Calculate a current output torque of the vehicle according to the adjustment amount of the front-wheel steering angle and the adjustment amount of the rear-wheel rotational speed, so as to drive the vehicle to maintain drifting along the drifting path.

[0134] Optionally, the device further includes:

[0135] An expected drifting trajectory drawing unit, configured to receive an expected drifting trajectory drawn based on the road image when the road information includes a road image;

[0136] Control the vehicle to travel along the expected drifting trajectory to complete a drifting action.

[0137] For the implementation processes of the functions and roles of each unit in the above device, refer to the implementation processes of the corresponding steps in the above method for details, which will not be elaborated here.

[0138] For the device embodiment, since it basically corresponds to the method embodiment, refer to the partial description of the method embodiment for related parts. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution in this specification. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0139] Based on the same concept as the above method, this specification further provides an electronic device, including: a processor; a memory for storing processor-executable instructions; wherein, the processor runs the executable instructions to implement the steps of the method as described in any of the above embodiments.

[0140] Based on the same concept as the above method, this specification further provides a computer-readable storage medium, on which computer instructions are stored, and when the instructions are executed by a processor, the steps of the method as described in any of the above embodiments are implemented.

[0141] Based on the same concept as the above method, this specification also provides a computer program product, including computer programs / instructions, which, when executed by a processor, implement the steps of the method described in any of the above embodiments.

[0142] Embodiments of the subject matter and functional operations described in this specification can be implemented in the following: digital electronic circuits, tangible computer software or firmware, computer hardware including the structures disclosed in this specification and their structural equivalents, or a combination of one or more of them. Embodiments of the subject matter described in this specification can be implemented as one or more computer programs, i.e., one or more modules of computer program instructions encoded on a tangible non-transitory program carrier for execution by, or to control the operation of, a data processing apparatus. Alternatively or additionally, the program instructions can be encoded on a manually generated propagated signal, such as a machine-generated electrical, optical, or electromagnetic signal, which is generated to encode and transmit information to a suitable receiver apparatus for execution by the data processing apparatus. A computer storage medium can be a machine-readable storage device, a machine-readable storage substrate, a random or serial access memory device, or a combination of one or more of them.

[0143] The processes and logical flows described in this specification can be executed by one or more programmable computers executing one or more computer programs to perform the corresponding functions by operating on input data and generating output. The processes and logical flows can also be executed by dedicated logic circuits, such as FPGAs (Field Programmable Gate Arrays) or ASICs (Application Specific Integrated Circuits), and the apparatus can also be implemented as dedicated logic circuits.

[0144] Computers suitable for executing computer programs include, for example, general and / or special microprocessors, or any other type of central processing unit. Generally, the central processing unit will receive instructions and data from a read-only memory and / or a random access memory. The basic components of a computer include a central processing unit for implementing or executing instructions and one or more memory devices for storing instructions and data. Generally, a computer will also include one or more mass storage devices for storing data, such as magnetic disks, magneto-optical disks, or optical disks, etc., or the computer will be operatively coupled to such mass storage devices to receive data from them or transfer data to them, or both. However, a computer is not necessarily required to have such devices. In addition, a computer can be embedded in another device, such as a mobile phone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a global positioning system receiver, or a portable storage device such as a universal serial bus (USB) flash drive, to name just a few.

[0145] Computer-readable media suitable for storing computer program instructions and data include all forms of non-volatile memory, media, and memory devices, including, for example, semiconductor memory devices (such as EPROM, EEPROM, and flash memory devices), magnetic disks (such as internal hard disks or removable disks), magneto-optical disks, and CD-ROM and DVD-ROM disks. The processor and the memory may be supplemented by, or incorporated in, special purpose logic circuitry.

[0146] Although this specification contains many specific implementation details, these should not be construed as limiting the scope of any invention or the scope of what is claimed, but rather as mainly being used to describe the features of specific embodiments of a particular invention. Certain features that are described in multiple embodiments in this specification may also be implemented in combination in a single embodiment. On the other hand, the various features described in a single embodiment may also be implemented separately in multiple embodiments or in any suitable sub-combination. In addition, although features may operate in certain combinations as described above and even be claimed as such initially, one or more features from a claimed combination may in some cases be removed from that combination, and the claimed combination may be directed to a sub-combination or a variation of a sub-combination.

[0147] Similarly, although operations are depicted in the drawings in a particular order, this should not be construed as requiring that the operations be performed in the particular order shown or sequentially, or that all illustrated operations be performed, to achieve the desired result. In some cases, multitasking and parallel processing may be advantageous. In addition, the separation of the various system modules and components in the above embodiments should not be construed as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

[0148] Thus, specific embodiments of the subject matter have been described. In addition, the processes depicted in the drawings are not necessarily in the particular order or sequential order shown to achieve the desired result. In some implementations, multitasking and parallel processing may be advantageous.

[0149] The above is only the preferred embodiment of this specification and is not intended to limit this specification. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of this specification shall be included within the scope of protection of this specification.

Claims

1. A vehicle obstacle avoidance method, characterized in that, The vehicle is connected to a drone; the method includes: When it is determined that there is an obstacle on the road where the vehicle is traveling, determine the road information of the road according to the drone; Obtain the driving state information of the vehicle, and determine the global drift feature points in the drift space according to the driving state information and the road information, and plan a drift path according to the global drift feature points, where the drift space is used to represent the parameter feasible region for the vehicle to effectively perform drift actions; Control the vehicle to travel along the drift path to avoid the vehicle colliding with the obstacle.

2. The method according to claim 1, characterized in that, The planning of the drift path according to the global drift feature points includes: Generate local trajectory points between different global drift feature points, and generate the drift path according to the global drift feature points and the local trajectory points.

3. The method according to claim 2, wherein The generation of the local trajectory points between different global drift feature points includes When the global drift feature points include the drift start point and the drift end point of the drift path, generate the local trajectory points between the drift start point and the drift end point; When the global drift feature points include the drift start point, the steady-state drift point and the drift end point of the drift path, generate the local trajectory points between the drift start point and the steady-state drift point, and / or between the steady-state drift point and the drift end point.

4. The method according to claim 3, characterized in that The control of the vehicle to travel along the drift path includes: Generate a vehicle state sequence and control instructions between the drift start point and the drift end point according to the global drift feature points; Control the vehicle to travel along the drift path based on the vehicle state sequence and the control instructions.

5. The method according to claim 2, wherein The determination of the global drift feature points in the drift space according to the driving state information and the road information includes: Calculate the drift space according to the driving state information and the road information, and search for the global drift feature points in the drift space; or, Receive the global drift feature points from the drone, where the global drift feature points are obtained by the drone calculating the drift space according to the road information and the driving state information obtained from the vehicle, and searching in the drift space.

6. The method according to claim 1, characterized in that The method further includes one of the following: When the obstacle avoidance curvature of the vehicle determined according to the driving state information and the road information is less than the minimum drift curvature, plan a non-drift path according to the driving state information and the road information; When the obstacle avoidance curvature of the vehicle determined according to the driving state information and the road information is not less than the minimum drift curvature and not greater than the maximum drift curvature, plan the drift path according to the driving state information and the road information; When the obstacle avoidance curvature of the vehicle determined according to the driving state information and the road information is greater than the maximum drift curvature, reduce the vehicle speed and re-plan the drift path.

7. The method according to claim 1, characterized in that, The method further includes: When the vehicle is connected to multiple drones, receive and splice the extended road information determined by at least some of the multiple drones, where the road range and / or road accuracy represented by the extended road information are greater than the road information determined by any one of the at least some drones.

8. The method according to claim 1, characterized in that, The method further includes: Determine first positioning information of the vehicle according to the visual sensor of the drone; Position the vehicle according to the first positioning information and / or second positioning information determined by the vehicle.

9. The method according to claim 1, wherein The controlling the vehicle to travel along the drift path includes: Determine the front-wheel steering angle adjustment amount and the rear-wheel rotation speed adjustment amount of the vehicle during driving based on model predictive control; Calculate the current output torque of the vehicle according to the front-wheel steering angle adjustment amount and the rear-wheel rotation speed adjustment amount to drive the vehicle to maintain drifting along the drift path.

10. The method according to claim 1, characterized in that, The method further includes: When the road information includes a road image, receive an expected drift trajectory drawn based on the road image; Control the vehicle to travel along the expected drift trajectory to complete the drifting action.

11. A vehicle obstacle avoidance device, characterized in that, The vehicle is connected to a drone; the device includes: A road information determination unit, configured to determine the road information of the road according to the drone when it is determined that there is an obstacle in the road traveled by the vehicle; A drift path planning unit, configured to obtain the driving state information of the vehicle, determine global drift feature points in the drift space according to the driving state information and the road information, and plan a drift path according to the global drift feature points, where the drift space is used to represent the parameter feasible region for the vehicle to effectively perform the drifting action; A vehicle control unit, configured to control the vehicle to travel along the drift path to avoid the vehicle colliding with the obstacle.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 10.

13. A computer program product, characterized in that, Includes a computer program / instructions, which when executed by a processor implements the steps of the method according to any one of claims 1-10.

Images