An Adaptive Speed Planning Method, Device and Domain Controller

Abstract

The present invention discloses an adaptive speed planning method, device and domain controller. The method includes: determining whether an obstacle target sensed by the vehicle will collide with the vehicle on the planned path of the vehicle; selecting the obstacle target with the closest relative distance as the following vehicle target from the obstacle targets that will collide, where the relative distance is the distance from the current position of the vehicle when each obstacle target is on the planned path of the vehicle; planning a speed control scheme based on the relative distance between the vehicle and the following vehicle target, the state parameters of the vehicle and the state parameters of the following vehicle target, so as to control the vehicle to travel according to the speed control scheme, and the state parameters include speed, heading and position. The technical solution provided by the present invention realizes accurate obstacle avoidance for vehicle autonomous driving in a multi-obstacle scenario.

Description

Technical Field

[0001] The present invention relates to the field of intelligent driving, and in particular to an adaptive speed planning method, device and domain controller. Background Art

[0002] In an unmanned driving system, trajectory planning is divided into path planning and speed planning. Among them, speed planning can provide a reference speed and a reference acceleration sequence to a downstream control module. The longitudinal controller calculates throttle and brake control amounts based on the above speed planning information, and the drive-by-wire chassis executes the control instruction to achieve vehicle speed control. Common speed planning algorithms include: 1. Determine the appropriate speed and acceleration of the vehicle based on the relative position and relative speed between the vehicle and the following vehicle target directly in front of the route; 2. Predict the driving trajectory of the adjacent vehicle, and judge whether there will be a collision with the vehicle when the adjacent vehicle changes lanes. If a collision occurs, the vehicle outputs a control curve for early deceleration according to the relative distance and relative speed of the adjacent vehicle, and controls the speed and acceleration of the vehicle (reference can be made to the patent document CN108032858A).

[0003] However, the application scenarios of the above speed planning algorithms are relatively single. In the real scenario, traffic jams often occur, and there are many vehicles on the driving trajectory of the vehicle. For example, there are not only obstacle vehicles directly in front of the vehicle's route, but also more than one adjacent vehicle merging into the vehicle's lane. If speed planning is carried out according to the control method of the existing technology, the planning result has poor accuracy for the situation with many obstacles, and it is very easy to cause collision accidents. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide an adaptive speed planning method, device and domain controller, thereby achieving accurate obstacle avoidance for autonomous driving of vehicles in a multi-obstacle scenario.

[0005] According to a first aspect, an embodiment of the present invention provides an adaptive speed planning method, the method includes: judging whether an obstacle target sensed by the vehicle will collide with the vehicle on the planned path of the vehicle; selecting the obstacle target with the shortest relative distance as the following vehicle target from the obstacle targets that will collide, where the relative distance is the distance from the current position of the vehicle when each obstacle target is on the planned path of the vehicle; planning a speed control scheme based on the relative distance between the vehicle and the following vehicle target, the state parameters of the vehicle and the state parameters of the following vehicle target, so as to control the vehicle to travel according to the speed control scheme, and the state parameters include speed, heading and position.

[0006] Optionally, the method further includes: when the vehicle travels according to the speed control scheme, re-determining whether an obstacle target sensed by the vehicle will collide with the vehicle on the planned path of the vehicle at a first preset time; if there is an obstacle target that will collide, returning to the step of selecting the obstacle target with the shortest relative distance from the obstacle targets that will collide as the following vehicle target.

[0007] Optionally, the determination of whether an obstacle target sensed by the vehicle will collide with the vehicle on the planned path of the vehicle includes: predicting the headings of the vehicle at various positions on the planned path of the vehicle, and predicting the entry point where the current obstacle target cuts into the planned path of the vehicle; predicting the driving position and driving speed of the vehicle on the planned path of the vehicle when the current obstacle target cuts into the planned path of the vehicle; determining the entry speed of the current obstacle target at the entry point based on the headings of the vehicle at various positions on the planned path of the vehicle; using the collision detection distance, the entry speed, the driving speed, the length of the vehicle, and the length of the current obstacle target to determine whether the vehicle will collide with the current obstacle target, where the collision detection distance is the distance between the entry point and the driving position.

[0008] Optionally, the planning of the speed control scheme based on the relative distance between the vehicle and the following vehicle target, the state parameters of the vehicle, and the state parameters of the following vehicle target includes: calculating the relative speed between the vehicle and the following vehicle target based on the current state parameters of the vehicle and the currently predicted state parameters of the following vehicle target; determining the desired vehicle speed of the vehicle based on the current state parameters of the vehicle and the speed limit information of the vehicle path; performing linear programming calculations using the relative distance between the vehicle and the following vehicle target, the relative speed between the vehicle and the following vehicle target, the desired vehicle speed, and the current vehicle speed of the vehicle, and outputting a first acceleration control amount; calculating the predicted state parameters of the vehicle after a second preset time using the first acceleration control amount; using the predicted state parameters of the vehicle as the current state parameters of the vehicle, using the predicted state parameters of the following vehicle target after the second preset time as the currently predicted state parameters of the following vehicle target, and returning to the step of calculating the relative speed between the vehicle and the following vehicle target based on the current state parameters of the vehicle and the currently predicted state parameters of the following vehicle target for iterative calculations; generating the speed control scheme based on the predicted state parameters of the vehicle obtained in each iteration.

[0009] Optionally, if there is no collision between the obstacle target sensed by the vehicle and the vehicle on the planned path of the vehicle, the method further includes: performing speed planning on the vehicle based on the speed limit information of the vehicle path on the planned path of the vehicle.

[0010] Optionally, the speed planning of the vehicle based on the vehicle path speed limit information on the planned path of the vehicle includes: determining the desired vehicle speed based on the current state parameters of the vehicle and the vehicle path speed limit information; calculating the second acceleration control amount of the vehicle based on the distance between the vehicle and the next speed limit point, the desired vehicle speed, and the current vehicle speed of the vehicle; calculating the predicted state parameters of the vehicle after a second preset time using the second acceleration control amount; taking the predicted state parameters of the vehicle as the current state parameters of the vehicle, returning to the step of determining the desired vehicle speed of the vehicle based on the current state parameters of the vehicle and the vehicle path speed limit information, and performing iterative calculations; generating a constant speed control scheme based on the predicted state parameters of the vehicle obtained in each iteration to control the vehicle to travel according to the constant speed control scheme.

[0011] Optionally, the determination of the desired vehicle speed based on the current state parameters of the vehicle and the vehicle path speed limit information includes: calculating the distance to be traveled and the first deceleration distance based on the current state parameters of the vehicle and the vehicle path speed limit information, where the distance to be traveled is the distance between the vehicle and the next speed limit point, and the first deceleration distance is the distance required for deceleration when the current vehicle speed of the vehicle is higher than the speed limit value of the next speed limit point; when the first situation or the second situation is satisfied, the desired vehicle speed is equal to the speed limit value of the next speed limit point, and when the third situation or the fourth situation is satisfied, the desired vehicle speed is equal to the current speed limit value; where the first situation is that the speed limit value of the next speed limit point is not less than the current vehicle speed of the vehicle, and the distance to be traveled is not greater than zero; the second situation is that the speed limit value of the next speed limit point is less than the current vehicle speed of the vehicle, and the difference between the distance to be traveled and the first deceleration distance is less than a preset threshold; the third situation is that the speed limit value of the next speed limit point is not less than the current vehicle speed of the vehicle, and the distance to be traveled is greater than zero; the fourth situation is that the speed limit value of the next speed limit point is less than the current vehicle speed of the vehicle, and the difference between the distance to be traveled and the first deceleration distance is not less than a preset threshold.

[0012] Optionally, determining whether the vehicle will collide with the current obstacle target by using the collision detection distance, the cut-in speed, the driving speed, the vehicle length, and the current obstacle target length includes: when the cut-in speed is greater than or equal to the driving speed, determining whether the collision detection distance is not greater than the first safety distance. If the collision detection distance is not greater than the first safety distance, it is determined that the vehicle will collide with the current obstacle target; when the cut-in speed is less than the driving speed, determining whether the collision detection distance is not greater than the second safety distance. If the collision detection distance is not greater than the second safety distance, it is determined that the vehicle will collide with the current obstacle target; wherein, the first safety distance is calculated by summing the second deceleration distance when the obstacle target decelerates to the driving speed, the vehicle length, the current obstacle target length, and a preset safety threshold; the second safety distance is calculated by summing the vehicle length, the current obstacle target length, and a preset safety threshold.

[0013] According to a second aspect, an embodiment of the present invention provides an adaptive speed planning device, which includes: a collision detection unit, configured to determine whether an obstacle target sensed by the vehicle will collide with the vehicle on the planned path of the vehicle; a following target unit, configured to select the obstacle target with the shortest relative distance as the following target from the obstacle targets that will collide, where the relative distance is the distance from the current position of the vehicle when each obstacle target is on the planned path of the vehicle; a speed planning unit, configured to plan a speed control scheme based on the relative distance between the vehicle and the following target, the state parameters of the vehicle, and the state parameters of the following target, so as to control the vehicle to travel according to the speed control scheme, and the state parameters include speed, heading, and position.

[0014] According to a third aspect, an embodiment of the present invention provides a domain controller, which includes: a sensing processing unit, a decision-making processing unit, a control processing unit, and a communication unit. The sensing processing unit, the decision-making processing unit, the control processing unit, and the communication unit are communicatively connected to each other. Computer instructions are stored in the decision-making processing unit, and the decision-making processing unit executes the computer instructions to execute the method provided in any optional implementation manner of the first aspect.

[0015] According to a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, which stores computer instructions for causing the computer to execute the method in the first aspect or any optional implementation manner of the first aspect.

[0016] The technical solution provided by this application has the following advantages:

[0017] For the technical solution provided by this application, in a multi-obstacle scenario, when this vehicle senses obstacle targets such as nearby vehicles through sensing devices such as radars, it first determines whether the sensed obstacle targets will collide with this vehicle in the future. The obstacle targets sensed by this vehicle may already be on the planned path of this vehicle or may cut into the planned path of this vehicle in the future. After that, this vehicle selects the closest one to this vehicle when the obstacle target is on the path of this vehicle from all the sensed obstacle targets that will collide, as the following vehicle target, so that this vehicle performs adaptive speed regulation, avoiding collisions with other targets farther than the following vehicle target while not colliding with the following vehicle target. Thus, the success rate of obstacle avoidance for this vehicle in a multi-obstacle scenario is improved.

[0018] In addition, in one embodiment, when this vehicle is driving according to the planned speed control scheme, it also continuously performs collision detection. Considering that other vehicles that do not originally collide are very likely to have a rear-end collision or be rear-ended during the deceleration process of this vehicle. Therefore, once this vehicle re-determines that the obstacle targets sensed by this vehicle will have a new collision accident with this vehicle on the planned path of this vehicle, this vehicle will return to the step of selecting the following vehicle target, re-select the following vehicle target and re-generate a new speed control scheme, thereby realizing an accurate and reliable multi-obstacle avoidance function. In addition, in one embodiment of this application, regardless of whether a collision occurs, when this vehicle performs speed planning, the vehicle path speed limit information of each section on the planned path of this vehicle is introduced, so that the vehicle speed reaches the highest under the condition of not exceeding the above vehicle path speed limit information and not having a collision, saving driving time. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The features and advantages of the present invention will be more clearly understood by referring to the accompanying drawings. The drawings are schematic and should not be construed as imposing any limitation on the present invention. In the drawings:

[0020] Figure 1 The step schematic diagram of an adaptive speed planning method in one embodiment of the present invention is shown;

[0021] Figure 2 The flow schematic diagram of an adaptive speed planning method in one embodiment of the present invention is shown;

[0022] Figure 3 The schematic diagram of a multi-obstacle target scenario in one embodiment of the present invention is shown;

[0023] Figure 4 The structural schematic diagram of an adaptive speed planning device in one embodiment of the present invention is shown;

[0024] Figure 5 The structural schematic diagram of a domain controller in one embodiment of the present invention is shown. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0026] Please refer to Figure 1 , in one embodiment, an adaptive speed planning method specifically includes the following steps:

[0027] Step S101: Determine whether the obstacle target sensed by the vehicle itself will collide with the vehicle itself on the planned path of the vehicle itself.

[0028] Step S102: Select the obstacle target with the shortest relative distance as the following vehicle target from the obstacle targets that will collide. The relative distance is the distance between each obstacle target and the current position of the vehicle when the obstacle target is on the planned path of the vehicle itself.

[0029] Step S103: Plan a speed control scheme based on the relative distance between the vehicle itself and the following vehicle target, the state parameters of the vehicle itself, and the state parameters of the following vehicle target, so as to control the vehicle to travel according to the speed control scheme. The state parameters include speed, heading, and position.

[0030] Specifically, in the embodiments of the present invention, the state parameters of the vehicle itself and the state parameters of the obstacle targets are pre-obtained through sensing devices such as the positioning radar, sensing radar, camera, and signal collector on the vehicle itself, including but not limited to the speed, heading, and position of the vehicle itself and the obstacle targets. In the embodiments of the present invention, the planned path of the vehicle itself is also pre-obtained through the path planning unit, and the driving paths of the obstacle targets near the vehicle itself can also be obtained through real-time prediction by the cooperation of the above sensing devices and the decision-making processing unit in the vehicle domain controller. The above relevant data are regarded as known in the embodiments of the present invention. The specific acquisition methods can refer to the prior art and will not be elaborated here.

[0031] Based on this, in the embodiments of the present invention, through data such as the state parameters of the vehicle itself, the planned path of the vehicle itself obtained, the state parameters of each predicted obstacle target, and the paths of the predicted obstacle targets, it is possible to determine whether the vehicle itself and the nearby obstacle targets perceived by the vehicle itself will collide (the collision methods include but are not limited to the vehicle itself rear-ending the obstacle target and the obstacle target rear-ending the vehicle itself. Among them, the scenario where the vehicle itself travels at the current speed or with a stable acceleration plan and can safely overtake the obstacle target also belongs to the situation where no collision will occur). If the vehicle itself determines that at least one obstacle target will collide with the vehicle itself, the vehicle itself enters the adaptive speed planning process. After the vehicle itself enters the adaptive speed planning process, first analyze the number of vehicles that will collide currently. In the actual application scenario, the obstacle target (assumed to be an obstacle vehicle) may already be traveling on the planned path of the vehicle itself, or may merge from the side and cut into the planned path of the vehicle itself in the future. Therefore, based on the relative distance between the vehicle itself and the obstacle target (the relative distance includes the distance between the obstacle target that is already on the planned path of the vehicle itself and the vehicle itself, and the distance between the entry point of other obstacle targets that will cut into the planned path of the vehicle itself in the future and the vehicle itself), select the obstacle target with the closest relative distance as the following vehicle target. Then, perform linear programming calculations based on the relative distance between the vehicle itself and the following vehicle target, the state parameters of the vehicle itself, and the state parameters of the following vehicle target to achieve the regulation of acceleration and speed, so as to generate the speed control plan with the fastest speed under the condition of not colliding with the following vehicle target. Driving according to the above speed control plan, when the vehicle itself does not collide with the following vehicle target, other obstacle targets with relatively far distances and that may collide will not collide either, thereby greatly improving the stability and accuracy of automatic obstacle avoidance of intelligent vehicles in multi-obstacle scenarios.

[0032] Specifically, in one embodiment, an adaptive speed planning method provided by the embodiments of the present invention further includes the following steps:

[0033] Step 1: When the vehicle itself travels according to the speed control plan, re-determine whether the obstacle targets perceived by the vehicle itself and the vehicle itself will collide on the planned path of the vehicle itself at a first preset time.

[0034] Step 2: If there are obstacle targets that will collide, return to the step of selecting the obstacle target with the closest relative distance as the following vehicle target from the obstacle targets that will collide.

[0035] Specifically, in this embodiment, when the vehicle is traveling according to the planned speed control plan, it is considered that other vehicles that would not otherwise collide with the vehicle may rear-end or be rear-ended during the vehicle's deceleration. Therefore, the vehicle continuously performs collision detection at a first preset time. Once the vehicle re-determines that a perceived obstacle target will cause a new collision with the vehicle on its planned path, the vehicle returns to the step of selecting a following target, reselects a following target, and regenerates a new speed control plan. This embodiment iterates collision and rear-end collision detection until the optimal following target is selected, thereby achieving accurate and reliable multi-obstacle avoidance. Furthermore, compared to traditional speed planning methods based on convex optimization theory to find the global optimum, the method proposed in this embodiment has the following advantages: the method proposed in this embodiment is non-convex optimization, and each iteration only considers the local optimum. Speed planning is only re-performed when a collision is detected again. The solution calculation provided by this embodiment is simpler than the existing technology and uses less computing resources.

[0036] Specifically, in one embodiment, the above step S101 specifically includes the following steps:

[0037] Step 3: Predict the vehicle's heading at each position on its planned path, and predict the point where the current obstacle enters its planned path.

[0038] Specifically, the vehicle obtains the state parameters and other relevant information of each obstacle target from the sensing device, including the movement speed v of the obstacle target. obs , heading θ obs 、Position(x obs ,y obs ), and the current heading of the vehicle θ ego , the current position of the vehicle (x ego ,y ego ), vehicle speed information v ego Then, this embodiment establishes the obstacle target prediction trajectory equation based on the above data:

[0039] y = tanθ obs *x+(y obs -tanθ obs *x obs )

[0040] Afterwards, the vehicle's planned path information is obtained from the decision processing unit, including the position coordinates ((x0, y0), (x1, y1), ..., (x n ,y n )), and obtain the planned path expression of the vehicle through polynomial fitting:

[0041] y=a0+a1x+a2x z +…+a k xk

[0042] By combining the above obstacle target prediction trajectory equation and the vehicle planning path expression, the vehicle's heading θ1 at each point on the vehicle's planned path and the time t when the obstacle target enters the planned path can be calculated. in , and the entry point position (x in ,y in ), s in .

[0043] Step 4: Predict the vehicle's position and speed on its planned path when the current obstacle target cuts into it.

[0044] Specifically, using the entry point calculated in step three above, the corresponding driving position and driving speed can be calculated when the current obstacle target cuts into the planned path of the vehicle if the vehicle continues to drive on the planned path according to the current speed plan.

[0045] Step 5: Based on the heading of the vehicle at each position on the planned path of the vehicle, determine the entry speed of the current obstacle target at the entry point.

[0046] Specifically, by using the above step 3 to calculate the heading of the vehicle at each position on the planned path, the speed of the obstacle at the entry point can be calculated. The angle between the obstacle and the tangent direction of the path when it enters the planned path of the vehicle is α1 = θ obs -θ1, then calculate the obstacle at the entry point s in The cutting speed v in the direction (i.e. the tangent direction of the cutting point) obs_s for:

[0047] v obs_s =v obs *cosα1

[0048] Step 6: Use the collision detection distance, cut-in speed, driving speed, vehicle length, and current obstacle target length to determine whether the vehicle will collide with the current obstacle target. The collision detection distance is the distance between the cut-in point and the driving position.

[0049] Specifically, the safety distance is analyzed and determined using the obstacle's entry speed, the vehicle's speed, the vehicle's length, the current obstacle's length, and the distance between the obstacle and the vehicle's corresponding position when it reaches the entry point (collision detection distance), resulting in accurate collision detection. (Collision scenarios in this embodiment include, but are not limited to, the vehicle rear-ending the obstacle; the vehicle directly colliding with the obstacle during entry; and the vehicle overtaking the obstacle and then having the obstacle enter the vehicle's planned path and then rear-end the vehicle.) Compared to existing collision detection methods, this algorithm is simpler and saves computing resources.

[0050] Specifically, in this embodiment, step six specifically includes the following steps:

[0051] Step seven: When the cut-in speed is greater than or equal to the driving speed, determine whether the collision detection distance is not greater than the first safety distance. If the collision detection distance is not greater than the first safety distance, it is determined that the vehicle will collide with the current obstacle target;

[0052] Step eight: When the cut-in speed is less than the driving speed, determine whether the collision detection distance is not greater than the second safety distance. If the collision detection distance is not greater than the second safety distance, it is determined that the vehicle will collide with the current obstacle target; wherein, the first safety distance is calculated by summing the second deceleration distance when the obstacle target decelerates to the driving speed, the vehicle length, the current obstacle target length, and a preset safety threshold; the second safety distance is calculated by summing the vehicle length, the current obstacle target length, and a preset safety threshold.

[0053] Specifically, the collision detection and judgment processes of steps seven and eight above are described by the following specific formulas:

[0054] If v obs_s ≥v ego_predict , when , the vehicle will collide with the obstacle target or cause the obstacle target to rear-end the vehicle;

[0055] If v obs_s <v ego_predict , when s ego_predict -s in ≤k2*L EGO +k3*L obs +dis-safe, the vehicle will collide with the obstacle target or cause the obstacle target to rear-end the vehicle;

[0056] In the formula, k2, k3, and k4 are proportional parameters configured during the application of the algorithm, dis-safe is a preset safety threshold configured during the application of the algorithm, and can be adjusted according to actual situations. v obs_s is the cut-in speed of the obstacle target, L EGO is the vehicle length, L obs is the length of the obstacle target, v ego_predict is the driving speed corresponding to the vehicle when the obstacle target cuts into the planned path of the vehicle; s ego_predict is the driving position corresponding to the vehicle when the obstacle target cuts into the planned path of the vehicle; s in is the cut-in position of the obstacle target; s ego_predict -s in i.e., the distance between the obstacle target and the vehicle when the obstacle target cuts into the planned path of the vehicle (collision detection distance); It is used to characterize the second deceleration distance required when the obstacle target moves faster than the vehicle, the obstacle target decelerates to the vehicle speed, and the distance between the obstacle target and the vehicle is zero when it reaches the vehicle speed. It can be seen that the algorithm for judging whether the obstacle target collides with the vehicle in this embodiment only needs to simply compare the collision detection distance with two safety distances, which reduces the algorithm complexity and improves the overall algorithm calculation efficiency on the premise of ensuring the collision detection accuracy.

[0057] Specifically, in one embodiment, step S103 described above specifically includes the following steps:

[0058] Step Nine: Calculate the relative speed between the vehicle and the following vehicle target based on the current state parameters of the vehicle and the currently predicted state parameters of the following vehicle target.

[0059] Step Ten: Determine the desired vehicle speed of the vehicle based on the current state parameters of the vehicle and the speed limit information of the vehicle path;

[0060] Step Eleven: Perform linear programming calculation using the relative distance between the vehicle and the following vehicle target, the relative speed between the vehicle and the following vehicle target, the desired vehicle speed, and the current vehicle speed of the vehicle, and output the first acceleration control quantity;

[0061] Step Twelve: Calculate the predicted state parameters of the vehicle after the second preset time using the first acceleration control quantity;

[0062] Step Thirteen: Take the predicted state parameters of the vehicle as the current state parameters of the vehicle, take the predicted state parameters of the following vehicle target after the second preset time as the currently predicted state parameters of the following vehicle target, and return to Step Nine for iterative calculation;

[0063] Step Fourteen: Generate a speed control scheme based on the predicted state parameters of the vehicle obtained from each iteration.

[0064] Specifically, in this embodiment, after the vehicle determines the following vehicle target, the relative speed between the vehicle and the following vehicle target can be calculated according to the current state parameters of the vehicle obtained by the sensing device in real time and the state parameters of the following vehicle target predicted in advance. In addition, in order to meet the dual conditions of not colliding and not exceeding the speed limits of each section of the vehicle planning path in the embodiments of the present invention, so that the vehicle speed reaches the optimum, the desired vehicle speed of the vehicle is also determined based on the current state parameters of the vehicle and the speed limit information of the vehicle path. Then, perform linear programming calculation using the relative distance between the vehicle and the following vehicle target, the relative speed between the vehicle and the following vehicle target, the desired vehicle speed, and the current vehicle speed of the vehicle, and output the first acceleration control quantity after the second preset time (for example, the next moment), and then calculate the predicted state parameters of the vehicle after the second preset time using the first acceleration control quantity, that is, the vehicle speed, position, heading, etc. at the next moment. The specific formulas are as follows:

[0065]

[0066] In the formula, x and y represent the change in the direction position of the vehicle, s represents the distance from the position of the vehicle to the starting point of the planned path of the vehicle, and T represents the second preset time.

[0067] The linear programming calculation process involved in step eleven can be implemented through algorithms such as sliding mode, interpolation table, PID, etc. The output result uses the maximum acceleration constraint of the vehicle and the maximum deceleration constraint of the vehicle to limit the acceleration control amount. The specific mathematical operation process of the linear programming calculation is the prior art and will not be elaborated here. Then, the acceleration control amount and prediction state parameters at the next moment are used as input values and returned to step nine to iteratively calculate the acceleration control amount and prediction state parameters at the moment after the next moment, so as to generate a speed control scheme based on the results of repeated iterations (s 0， s1......s t ),(v 0， v1......v t ),(a 0， a1......a t ). The principle of the speed planning algorithm in this embodiment is simple, with a small amount of calculation and easy to implement.

[0068] It should be noted that in this embodiment, considering that due to the reasons of the vehicle's own hardware, there is usually a response delay. Before calculating the prediction state parameters of the vehicle, the first acceleration control amount is input into a first-order inertia link (inertia link - longitudinal kinematic model) to simulate the vehicle response delay, so as to obtain the acceleration of the actual vehicle response, and calculate the prediction state parameters of the vehicle according to the acceleration of the actual vehicle response, further improving the accuracy of the vehicle response speed planning result and the success rate of obstacle avoidance. The first-order inertia link model proposed in this embodiment is as follows:

[0069]

[0070] In the formula, u represents the first acceleration control amount (similarly, u can also be applied to the second acceleration control amount calculated when there is no obstacle), a t-1 represents the acceleration of the actual response at the previous moment, a t represents the acceleration of the actual response at the current moment, can be converted to in the form, and ts represents the inertia time constant.

[0071] Finally, select the reference acceleration at the k-th moment in the speed planning scheme, and convert it into brake and throttle control quantities through a pre-calibrated relationship table to achieve driverless speed planning and longitudinal control; the embodiment of the present invention is based on real-time control theory, and the reference acceleration at the first moment is the control quantity calculated based on the current real vehicle state; starting from the second moment, they are all control quantities iterated from the inertial link - longitudinal kinematic model. Generally, the real-time control quantity can be directly used (i.e., the first moment), and the k-th reference acceleration control can also be used during actual vehicle debugging as a means to compensate for the system response delay.

[0072] Specifically, in an embodiment, the adaptive speed planning method provided by the embodiment of the present invention further includes the following steps:

[0073] Step 15: If there is no collision between the obstacle target sensed by the vehicle itself and the vehicle itself on the planned path of the vehicle itself, perform speed planning for the vehicle itself based on the speed limit information on the planned path of the vehicle itself.

[0074] Specifically, in this embodiment, when there is no collision between the obstacle target sensed by the vehicle itself and the vehicle itself, the vehicle itself enters the constant speed planning process. The constant speed planning process does not consider the collision situation and only performs linear programming calculation based on the speed limit information on the planned path of the vehicle itself, so as to achieve the optimal vehicle speed without exceeding the speed limit value of each section. The specific planning steps are as follows:

[0075] 1. Determine the desired vehicle speed of the vehicle itself based on the current state parameters of the vehicle itself and the speed limit information of the vehicle path.

[0076] 2. Calculate the second acceleration control quantity of the vehicle itself based on the distance between the vehicle itself and the next speed limit point, the desired vehicle speed, and the current vehicle speed of the vehicle itself.

[0077] 3. Use the second acceleration control quantity to calculate the predicted state parameters of the vehicle itself after the second preset time.

[0078] 4. Take the predicted state parameters of the vehicle itself as the current state parameters of the vehicle itself, return to step 1, and perform iterative calculation.

[0079] 5. Generate a constant speed control scheme based on the predicted state parameters of the vehicle itself obtained from each iteration to control the vehicle itself to travel according to the constant speed control scheme.

[0080] Specifically, the specific steps of generating the constant speed control scheme are similar to the steps of generating the speed control scheme in the above adaptive speed planning process, except that the linear programming calculation process of the constant speed control scheme does not need to consider the obstacle target and does not need to introduce parameters such as relative speed. For the description of other steps, refer to the relevant descriptions in steps 9 to 14 above, and will not be elaborated here.

[0081] It should be noted that in the constant speed control scheme, the response delay caused by the vehicle's own hardware factors is also considered, and a first-order inertia link model is introduced to adjust the second acceleration control amount. In addition, during the driving process of the vehicle according to the constant speed control scheme, the vehicle also performs collision detection at a first preset time. If an obstacle target that can be collided is found, the above step S102 is returned to perform adaptive speed planning.

[0082] Specifically, in one embodiment, the above step ten and step 1 specifically include the following steps:

[0083] Step sixteen: Calculate the remaining distance to travel and the first deceleration distance based on the current state parameters of the vehicle and the speed limit information of the vehicle's path. The remaining distance to travel is the distance between the vehicle and the next speed limit point, and the first deceleration distance is the distance required for deceleration when the current vehicle speed is higher than the speed limit value of the next speed limit point;

[0084] Step seventeen: When the first condition or the second condition is met, the desired vehicle speed is equal to the speed limit value of the next speed limit point. When the third condition or the fourth condition is met, the desired vehicle speed is equal to the current speed limit value. Among them, the first condition is that the speed limit value of the next speed limit point is not less than the current vehicle speed of the vehicle, and the remaining distance to travel is not greater than zero; the second condition is that the speed limit value of the next speed limit point is less than the current vehicle speed of the vehicle, and the difference between the remaining distance to travel and the first deceleration distance is less than a preset threshold; the third condition is that the speed limit value of the next speed limit point is not less than the current vehicle speed of the vehicle, and the remaining distance to travel is greater than zero; the fourth condition is that the speed limit value of the next speed limit point is less than the current vehicle speed of the vehicle, and the difference between the remaining distance to travel and the first deceleration distance is not less than a preset threshold.

[0085] Specifically, in this embodiment, the summary of generating the desired vehicle speed based on the speed limit information of each section of the vehicle's planned path is as follows:

[0086] If the next speed limit point is slower than the current vehicle speed, it is necessary to calculate the first deceleration distance required to decelerate from the current vehicle speed to the speed limit value based on the current deceleration, and judge: If the remaining distance to travel of the vehicle to the next speed limit point is exactly equal to or less than the first deceleration distance, the desired vehicle speed of the constant speed cruise should be changed to the next speed limit value and start decelerating to ensure that the deceleration is in place when reaching the speed limit point; otherwise, when the speed of the next speed limit point is greater than the current speed, it is necessary to start accelerating when reaching the speed limit point to avoid accelerating in advance and exceeding the speed limit of the current section.

[0087] The expression for calculating the desired vehicle speed is as follows:

[0088] dis=s hext_limit -s now

[0089] dis need =k*(v next_limit2 -v ego 2 ) / (2*a min )

[0090] Where dis is the distance to be traveled, s next_limit is the location of the next speed limit point, s now is the current position of the vehicle, dis need is the first deceleration distance, v ego is the current speed of the vehicle, v next_limit is the speed limit value of the next speed limit point, a min is the maximum deceleration constraint.

[0091]

[0092] Where, v des is the expected speed, is the current speed limit value, and k1 is the speed limit position adjustment value, which is used to measure the difference between the distance to be traveled and the first deceleration distance. It is usually 0, but in actual application, the formula is not so accurate. Therefore, during debugging, adjust the k1 value according to the actual situation. For example, changing it to 1 means that deceleration starts 1 meter earlier based on the first deceleration distance.

[0093] Specifically, if Figure 2 and Figure 3 As shown, in an actual application scenario embodiment, the above steps are further explained as follows:

[0094] The vehicle's planned path, speed, heading, and location at each moment are all known. The vehicle can use its sensing devices to detect nearby obstacles, including but not limited to cars, pedestrians, and roadblocks. The vehicle's decision processing unit can predict the speed, heading, and location of obstacles at each moment.

[0095] Based on this, during the driving process of the vehicle, obstacle target detection is continuously performed at a first preset time. If no obstacle target is detected, the constant-speed planning process is executed. The vehicle calculates the expected vehicle speed at each moment according to the method in step 15 above, combined with the speed limit information of each section on the vehicle's path. Then, under the condition of not exceeding the expected vehicle speed, the speed planning of the vehicle is carried out. And in the calculation process of speed planning, a first-order inertia link model is introduced to avoid inaccurate output of the acceleration control amount due to the response delay of the vehicle's own hardware. If an obstacle target is sensed nearby, first, it is judged whether the vehicle will collide with each obstacle target. If it is detected that the vehicle will collide with one or more obstacle targets, the vehicle enters the adaptive speed planning process. The vehicle selects the obstacle target with the shortest relative distance from each obstacle target and uses this obstacle target as the following target to calculate the relative speed. Then, linear programming calculation is performed using the relative distance, relative speed, expected vehicle speed, and the state parameters of the vehicle and the obstacle target to obtain an optimal speed operation curve that neither collides nor exceeds the speed limit, that is, the speed planning scheme. During the driving process of the vehicle according to the speed planning scheme, the vehicle still continuously performs collision detection to prevent new collisions from occurring to the obstacle targets that did not collide originally during the deceleration process of the vehicle. Once the vehicle detects a new collision target, a following target is reselected from the new collision targets and speed planning is performed. In the above iterative process, an optimal speed control scheme is output.

[0096] In the embodiment of the present invention, the adaptive control method is applied to the field of driverless speed planning, and a set of speed planning method based on adaptive control is proposed, which can implement functions such as starting, following, stopping, and constant-speed cruising. At the same time, compared with the traditional speed planning method based on convex optimization theory, the technical solution provided by the embodiment of the present invention has no complex solution operations, low calculation resource occupation; based on real-time control theory, there is no problem of solution failure; the speed planning result does not require secondary smoothing processing; the planning control is integrated design, and the algorithm structure is streamlined. On the basis of introducing obstacle prediction information, path planning information and global speed limit information are also introduced as prior information for speed planning to further improve the practicality of speed planning. Considering the controller response constraint and vehicle command execution delay problems, the embodiment of the present invention establishes an inertia link - longitudinal kinematic model to simulate the vehicle's real-time response feedback. The control quantity output within the prediction period is fed back through the inertia link - longitudinal kinematic model to generate the state quantity of the next moment, and then the future state parameters of the vehicle are iteratively predicted, which can provide a more accurate basis for collision detection and avoid misresponse of the controller and affect the control effect; in this embodiment, the optimal following target is selected through iterative collision detection, considering the problem that after overtaking the obstacle target that is about to enter the driving trajectory, it may cause the obstacle target vehicle to rear-end the vehicle.

[0097] As Figure 4 shown, this embodiment also provides an adaptive speed planning device, which includes:

[0098] A collision detection unit 101 is configured to determine whether an obstacle target sensed by the vehicle will collide with the vehicle on the planned path of the vehicle. For detailed content, refer to the relevant description of step S101 in the above method embodiment, which will not be elaborated here.

[0099] A following target unit 102 is configured to select the obstacle target with the shortest relative distance as the following target from the obstacle targets that will collide. The relative distance is the distance from each obstacle target to the current position of the vehicle when the obstacle targets are on the planned path of the vehicle. For detailed content, refer to the relevant description of step S102 in the above method embodiment, which will not be elaborated here.

[0100] A speed planning unit 103 is configured to plan a speed control scheme based on the relative distance between the vehicle and the following target, the state parameters of the vehicle, and the state parameters of the following target, so as to control the vehicle to travel according to the speed control scheme. The state parameters include speed, heading, and position. For detailed content, refer to the relevant description of step S103 in the above method embodiment, which will not be elaborated here.

[0101] The adaptive speed planning device provided by the embodiment of the present invention is configured to execute the adaptive speed planning method provided by the above embodiment, and its implementation manner and principle are the same. For detailed content, refer to the relevant description of the above method embodiment, which will not be elaborated here.

[0102] Figure 5 A domain controller according to an embodiment of the present invention is shown. The domain controller at least includes a sensing and processing unit 901, a decision-making and processing unit 902, a control and processing unit 903, and a communication unit 904. The sensing and processing unit 901, the decision-making and processing unit 902, the control and processing unit 903, and the communication unit 904 can communicate with each other through a bus or other means. Figure 5 Taking the bus mode as an example.

[0103] In this embodiment, the sensing and processing unit 901 and the decision-making and processing unit 902 each include an independent processor. The sensing and processing unit 901 and the decision-making and processing unit 902 may each include an independent memory, or may use a shared memory.

[0104] In the embodiment of the present invention, the sensing and processing unit 901 is mainly applied to the construction machinery scenario, and its main function is to perform sensing and fusion processing on the image data obtained from the camera to obtain the environmental information of the current construction machinery, and then send the environmental information to the control and processing unit 903 or the decision-making and processing unit 902 according to the data type of the environmental information signal.

[0105] The decision-making processing unit 902 is mainly used to formulate driving or operation strategies after integrating information such as the surrounding environment, operation scenarios, and vehicle status, and finally issue control commands. The main function of the communication unit 904 is to communicate with the cloud server, upload device-related status and information to the cloud service, and request the cloud server to assist in computing and processing. The control processing unit 903 is mainly used for the conversion of communication protocols between signals.

[0106] Those skilled in the art can understand that to implement all or part of the processes in the above-described embodiment methods, it can be completed by instructing relevant hardware through a computer program. The implemented program can be stored in a computer-readable storage medium and executed by the decision-making processing unit 902. When the program is executed, it can include the processes of the above-described method embodiments. Among them, the storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium can also include a combination of the above types of memories.

[0107] Since the computing tasks of the decision-making processing unit 902 are relatively heavy, especially involving the processing of artificial intelligence, neural networks, and data training, in this embodiment, the decision-making process can also be completed with the help of cloud computing resources. In other words, the location of the storage medium storing the program instructions / modules corresponding to the methods in the above method embodiments includes the domain controller and the cloud server. The cloud server receives the relevant image data sent by the domain controller, and then realizes the method of extracting the scene image features in the above method embodiments by running the executable program stored in the storage medium of the cloud server. In this embodiment, the cloud service can be a central server or an edge server.

[0108] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations fall within the scope defined by the appended claims.

Claims

1. An adaptive speed planning method, characterized in that: The method comprises: Predict the heading of the vehicle at each position on the planned path of the vehicle, and predict the entry point of the current obstacle target into the planned path of the vehicle; predict the driving position and driving speed of the vehicle on the planned path of the vehicle when the current obstacle target cuts into the planned path of the vehicle; determine the entry speed of the current obstacle target at the entry point based on the heading of the vehicle at each position on the planned path of the vehicle; when the entry speed is greater than or equal to the driving speed, determine whether the collision detection distance is not greater than the first safety distance, and if the collision detection distance is not greater than the first safety distance, determine that the vehicle will collide with the current obstacle target; when the When the cut-in speed is less than the driving speed, determining whether the collision detection distance is not greater than a second safety distance; if the collision detection distance is not greater than the second safety distance, determining that the vehicle will collide with the current obstacle target; wherein the collision detection distance is the distance between the cut-in point and the driving position, the first safety distance is calculated by summing the second deceleration distance when the obstacle target decelerates to the driving speed, the length of the vehicle, the current obstacle target length, and a preset safety threshold; the second safety distance is calculated by summing the length of the vehicle, the current obstacle target length, and the preset safety threshold; Select the obstacle with the closest relative distance from the obstacle targets that may collide as the following target, where the relative distance is the distance between each obstacle target and the current position of the vehicle when it is located on the planned path of the vehicle; A speed control scheme is planned based on the relative distance between the vehicle and the following target, the state parameters of the vehicle and the state parameters of the following target to control the vehicle to travel according to the speed control scheme, wherein the state parameters include speed, heading and position.

2. The method according to claim 1, characterized in that The method further comprises: When the vehicle is traveling according to the speed control scheme, re-determining whether an obstacle target perceived by the vehicle will collide with the vehicle on the planned path of the vehicle within a first preset time; If there is an obstacle target that may collide, the process returns to the step of selecting the obstacle target with the closest relative distance from the obstacle targets that may collide as the following target.

3. The method according to claim 1, characterized in that The speed control scheme is planned based on the relative distance between the host vehicle and the following target, the state parameters of the host vehicle, and the state parameters of the following target, including: Calculating the relative speed between the host vehicle and the following target based on the current state parameters of the host vehicle and the currently predicted state parameters of the following target; Determine the desired speed of the vehicle based on the vehicle's current state parameters and the vehicle's path speed limit information; Performing a linear programming calculation using the relative distance between the host vehicle and the following target, the relative speed between the host vehicle and the following target, the desired vehicle speed, and the current vehicle speed of the host vehicle to output a first acceleration control variable; Calculating a predicted state parameter of the vehicle after a second preset time using the first acceleration control amount; Using the predicted state parameter of the host vehicle as the current state parameter of the host vehicle, using the state parameter predicted after the second preset time of the following target as the current predicted state parameter of the following target, returning to the step of calculating the relative speed between the host vehicle and the following target based on the current state parameter of the host vehicle and the current predicted state parameter of the following target, and performing iterative calculation; The speed control scheme is generated based on the predicted state parameters of the vehicle obtained in each iteration.

4. The method according to claim 1, wherein If the obstacle target sensed by the vehicle will not collide with the vehicle on the planned path of the vehicle, the method further includes: The speed of the vehicle is planned based on the speed limit information of the vehicle on the planned path.

5. The method according to claim 4, characterized in that The speed planning of the vehicle based on the vehicle path speed limit information on the planned path of the vehicle includes: Determine the desired speed of the vehicle based on the vehicle's current state parameters and the vehicle's path speed limit information; Calculating a second acceleration control amount of the vehicle based on the distance between the vehicle and the next speed limit point, the desired vehicle speed, and the current vehicle speed of the vehicle; Calculating a predicted state parameter of the vehicle after a second preset time using the second acceleration control amount; Using the predicted state parameter of the vehicle as the current state parameter of the vehicle, returning to the step of determining the expected speed of the vehicle based on the current state parameter of the vehicle and the speed limit information of the vehicle path, and performing iterative calculation; A constant speed control scheme is generated based on the predicted state parameters of the vehicle obtained in each iteration to control the vehicle to travel according to the constant speed control scheme.

6. The method according to claim 3 or 5, characterized in that The determining of the expected speed of the vehicle based on the current state parameters of the vehicle and the speed limit information of the vehicle path includes: Calculating a distance to be traveled and a first deceleration distance based on the current state parameters of the vehicle and the speed limit information of the vehicle's path. The distance to be traveled is the distance between the vehicle and the next speed limit point. The first deceleration distance is the distance required to decelerate when the vehicle's current speed exceeds the speed limit value of the next speed limit point. When the first condition or the second condition is met, the expected vehicle speed is equal to the speed limit value of the next speed limit point; when the third condition or the fourth condition is met, the expected vehicle speed is equal to the current speed limit value; Among them, the first situation is that the speed limit value of the next speed limit point is not less than the current speed of the vehicle, and the distance to be traveled is not greater than zero; the second situation is that the speed limit value of the next speed limit point is less than the current speed of the vehicle, and the difference between the distance to be traveled and the first deceleration distance is less than a preset threshold; the third situation is that the speed limit value of the next speed limit point is not less than the current speed of the vehicle, and the distance to be traveled is greater than zero; the fourth situation is that the speed limit value of the next speed limit point is less than the current speed of the vehicle, and the difference between the distance to be traveled and the first deceleration distance is not less than a preset threshold.

7. An adaptive speed planning device, characterized in that: The device comprises: a collision detection unit, configured to predict the heading of the vehicle at various locations on the planned path of the vehicle and predict the entry point of the current obstacle target into the planned path of the vehicle; predict the driving position and driving speed of the vehicle on the planned path of the vehicle when the current obstacle target cuts into the planned path of the vehicle; determine the entry speed of the current obstacle target at the entry point based on the heading of the vehicle at various locations on the planned path of the vehicle; when the entry speed is greater than or equal to the driving speed, determine whether the collision detection distance is not greater than a first safety distance; if the collision detection distance is not greater than the first safety distance, determine that the vehicle will collide with the current obstacle target; when the cut-in speed is less than the driving speed, determine whether the collision detection distance is not greater than a second safety distance; if the collision detection distance is not greater than the second safety distance, determine that the vehicle will collide with the current obstacle target; wherein the first safety distance is calculated by summing the second deceleration distance when the obstacle target decelerates to the driving speed, the length of the vehicle, the length of the current obstacle target, and a preset safety threshold; and the second safety distance is calculated by summing the length of the vehicle, the length of the current obstacle target, and the preset safety threshold; A following target unit is used to select the obstacle target with the closest relative distance from the obstacle targets that may collide as the following target. The relative distance is the distance between each obstacle target and the current position of the vehicle when it is located on the planned path of the vehicle; The speed planning unit is used to plan a speed control plan based on the relative distance between the vehicle and the following target, the state parameters of the vehicle and the state parameters of the following target, so as to control the vehicle to travel according to the speed control plan. The state parameters include speed, heading and position.

8. A domain controller, characterized in that: include: A perception processing unit, a decision processing unit, a control processing unit and a communication unit, wherein the perception processing unit, the decision processing unit, the control processing unit and the communication unit are communicatively connected to each other, the decision processing unit stores computer instructions, and the decision processing unit executes the method according to any one of claims 1 to 6 by executing the computer instructions.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the method according to any one of claims 1 to 6.

Images