Efficient driving interaction method and system based on backward reachable set safety assurance

Abstract

The present invention discloses an efficient driving interaction method and system based on backward reachable set safety assurance, wherein the method includes the following steps: S1. Dividing the sampled values ​​of the vehicle state vector according to a certain grid size, offline calculating and storing the level set function values ​​by solving the level set equation, constructing a mapping relationship between the vehicle state vector and the level set function value to form a discrete table; when the level set function values ​​are all zero, it corresponds to a zero level set, which corresponds to the safe distance boundary of the interacting vehicle relative to the ego vehicle; S2. Obtaining the ego vehicle's state vector online and searching the corresponding level set function value according to the discrete table; S3. Comparing the found level set function value with a set threshold and entering different control constraints based on the comparison result; S4. Generating control instructions based on the control constraints to control the ego vehicle's driving. The present invention improves vehicle traffic efficiency in interactive scenarios.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to an efficient driving interaction method and system based on backward reachable set safety assurance. Background Art

[0002] With the continuous advancement of key technologies for intelligent connected vehicles, single-vehicle autonomous driving technology is gradually maturing. To further operate efficiently and safely in dynamic and uncertain interactive environments, intelligent vehicles must consider the risks associated with their actions and the feedback interactions with other risk-sensitive agents, avoiding overly conservative control strategies. Ignoring either of these elements makes it difficult to design safe and efficient interaction strategies.

[0003] To this end, rigorous mathematical proof and reasoning are required to give a set of dangerous states, and then conduct interactive trials outside this set to strive for priority access, meet the expected goals of decision-making and planning, and improve vehicle interaction efficiency. When approaching the edge of this set, additional control constraints must be imposed. Through a smooth control sequence, the vehicle can stay away from dangerous states while ensuring passenger comfort, avoiding the use of safety failure control strategies that maximize safety but sacrifice comfort. Summary of the Invention

[0004] The present invention addresses the problem that existing intelligent driving vehicle control systems in uncertain interactive environments have difficulty in generating safe and reliable control sequences while balancing efficiency and comfort. This invention provides an efficient control method and system based on backward reachable set safety assurance in interactive scenarios.

[0005] The technical solution adopted in the present invention is:

[0006] The present invention provides an efficient driving interaction method based on backward reachable set security assurance, comprising the following steps:

[0007] S1. Divide the sampled values ​​of the vehicle state vector according to a certain grid size, calculate and store the level set function value offline by solving the level set equation, and construct a mapping relationship between the vehicle state vector and the level set function value to form a discrete table; when the level set function value is all zero, it corresponds to a zero level set, which corresponds to the safe distance boundary of the interacting vehicle relative to the self-vehicle;

[0008] S2, obtain the state vector of the vehicle online and find the corresponding level set function value according to the discrete table;

[0009] S3. Compare the found level set function value with the set threshold and enter different control constraints based on the comparison result. If it is less than the first threshold, it enters the control hard constraint, and does not allow the ego vehicle to select a value other than the control value. If it is greater than the second threshold, it enters the control soft constraint, and allows the ego vehicle to select a control value within a certain range.

[0010] S4. Generate control instructions based on the control constraints to control the vehicle's driving.

[0011] Following the above technical solution, the level set equation is constructed based on the positional relationship between the ego vehicle and the surrounding interactive vehicles and the overall system dynamics equation composed of the ego vehicle and the interactive vehicles.

[0012] Following the above technical solution, the state vector includes position, velocity, heading angle, yaw rate and the position, velocity and heading angle of the interacting vehicle.

[0013] Following the above technical solution, after the vehicle state vector is obtained in step S2, it is filtered through an extended Kalman filter.

[0014] Following the above technical solution, if the level set function value is between the first threshold and the second threshold, no additional control constraints are introduced.

[0015] Following the above technical solution, in step S2, the level set function values ​​are interpolated between the grid points to find the level set function value in any given state.

[0016] Following the above technical solution, in step S4, the control constraints are specifically substituted into the MPC solver to solve for safe and smooth control instructions.

[0017] The present invention also provides an efficient driving interaction system based on backward reachable set security assurance, comprising:

[0018] The discrete table construction module is used to divide the sampled values ​​of the vehicle state vector according to a certain grid size, calculate and store the level set function value offline by solving the level set equation, and construct a mapping relationship between the vehicle state vector and the level set function value to form a discrete table. When the level set function value is all zero, it corresponds to a zero level set, which corresponds to the safe distance boundary of the interacting vehicle relative to the ego vehicle.

[0019] The table lookup module is used to obtain the state vector of the vehicle online and find the corresponding level set function value according to the discrete table;

[0020] The control constraint selection module is used to compare the found level set function value with the set threshold and enter different control constraints according to the comparison result. If it is less than the first threshold, it enters the control hard constraint, and the ego vehicle is not allowed to select values ​​other than the control amount. If it is greater than the second threshold, it enters the control soft constraint, and the ego vehicle is allowed to select the control amount within a certain range.

[0021] The control instruction generation module is used to generate control instructions according to the control constraints to control the driving of the vehicle.

[0022] Following the above technical solution, the control constraint selection module is further configured to not introduce additional control constraints when the level set function value is between the first threshold and the second threshold.

[0023] The present invention also provides a computer storage medium storing a computer program executable by a processor, wherein the computer program executes the efficient driving interaction method based on backward reachable set safety assurance described in the above technical solution.

[0024] The beneficial effects of the present invention are as follows: the present invention mainly utilizes the level set function value table obtained by offline solving the backward reachable set to explicitly express the risk level of the ego vehicle when the interactive vehicle interacts with the ego vehicle in different states; the level set function value in the current state can be found more accurately by online table search based on the collected vehicle state information; the corresponding control constraints are obtained based on the comparison between the level set function value and the threshold, and then the corresponding control instructions are generated to control the driving of the ego vehicle; the present invention avoids the impact of the method of generating emergency safety control with only safety as the guide, on the traffic efficiency of the ego vehicle, the comfort of passengers and the normal driving of other traffic participants.

[0025] Furthermore, by filtering the vehicle status information collected by the sensor using an extended Kalman filter, interference information can be better deleted.

[0026] Furthermore, by comparing with a pre-set threshold, the safe range of the level set function value in the current state is found, and then different level set function gradient value constraints are given to narrow the set that can be adopted by the current control sequence.

[0027] Furthermore, in the process of online searching for the level set function value, the present invention performs interpolation processing between grid points to evaluate the level set function value in any given state, thereby meeting the real-time requirement.

[0028] Furthermore, the control constraint is introduced into the MPC solver to ensure that the solved control sequence will meet the evolution requirements of the level set function value.

[0029] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0031] Figure 1This is a flow chart of an efficient driving interaction method based on backward reachable set safety assurance according to an embodiment of the present invention;

[0032] Figure 2 1 is a diagram showing several example slices of the level set function values ​​according to an embodiment of the present invention, where (a) is state 1 and (b) is state 2.

[0033] Figure 3 This is a flow chart for determining the safe range of the level set function value in the current state of an embodiment of the present invention. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0035] It should be noted that the illustrations provided in the embodiments of the present invention are only schematic illustrations of the basic concept of the present invention. Therefore, the drawings only show components related to the present invention and are not drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component can be changed at will, and the component layout type may also be more complicated.

[0036] In the present invention, it should also be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer" and the like are used to indicate positions or locations based on those shown in the accompanying drawings. These terms are intended solely to facilitate the description of the present application and to simplify the description. They are not intended to indicate or imply that the devices or components referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present application. Furthermore, the terms "first" and "second" are used solely for descriptive and distinguishing purposes and should not be construed as indicating or implying relative importance.

[0037] The present invention takes into account the impact of the vehicle's behavior on the behavior of other vehicles in interactive scenarios, and improves the traffic efficiency in interactive scenarios by attempting to seize the right of way within a safe and permitted set range; in more dangerous situations, a smoother approach is used to generate a safe and reliable control sequence.

[0038] Example 1

[0039] like Figure 1 As shown, the efficient driving interaction method based on backward reachable set security assurance in an embodiment of the present invention includes the following steps:

[0040] S1. Divide the sampled values ​​of the vehicle state vector according to a certain grid size, calculate and store the level set function value offline by solving the level set equation, and construct a mapping relationship between the vehicle state vector and the level set function value to form a discrete table; when the level set function value is all zero, it corresponds to a zero level set, which corresponds to the safe distance boundary of the interacting vehicle relative to the self-vehicle;

[0041] S2, obtain the state vector of the vehicle online and find the corresponding level set function value according to the discrete table;

[0042] S3. Compare the found level set function value with the set threshold and enter different control constraints based on the comparison result. If it is less than the first threshold, it enters the control hard constraint, and the vehicle is not allowed to select values ​​other than the control value. If it is greater than the second threshold, it enters the control soft constraint, and the vehicle is allowed to select the control value within a certain range.

[0043] S4. Generate control instructions based on the control constraints to control the vehicle's driving.

[0044] The level set method is a numerical technique that can express complex topological changes in the boundary of a backward reachable set through simple set operations. It can handle complex surface parameterization problems from a numerical computation perspective and ultimately provide a safe distance boundary for interacting vehicles relative to the ego vehicle. The level set equation used is as follows:

[0045]

[0046] in is the level set function that satisfies the Lipschitz continuity condition, f is the system dynamics equation, is the gradient of the level set function.

[0047] According to the distance boundary where the interacting vehicle may collide with the self-vehicle, a dangerous target set representing the final collision is predefined. The boundary of the target set is the level set function at the final value:

[0048]

[0049] Among them, x r and y r The horizontal and vertical coordinates of the interacting vehicle in a Cartesian coordinate system with the ego vehicle as the origin, i.e., their relative positions, are represented respectively; R represents the minimum safe distance. A larger value of the level set function indicates a safer state.

[0050] The reason why the present invention solves the level set equation offline is that it is extremely difficult to solve the level set equation under the higher-dimensional system dynamics equation. Considering that the vehicle is a platform with limited computing power, by gridding the relative state vectors of the vehicle and the interacting vehicle and calculating the level set function value under the corresponding state offline, the computing power demand is separated from the vehicle platform. However, as a price, the vehicle will sacrifice some storage space to record the discrete solution obtained offline, such as Figure 2 shown.

[0051] The relationship between the level set function value and the safety distance is as follows Figure 2 As shown in the two states (a) and (b) in the figure, any position in the entire image has its own level set function value. When this value is 0, it is the zero level set, which is also the safe distance boundary. If this boundary is crossed, that is, the level set function value is less than 0, the interacting vehicle will inevitably collide with the ego vehicle. Therefore, it is required to impose a hard constraint when the level set function value is close to 0, so that the ego vehicle stays away from this boundary. Figure 2 The different states refer to the states of the self-vehicle and the interactive vehicle other than the fixed horizontal and vertical relative positions. For example, in this embodiment, the vehicle state vector is a grid composed of 7-dimensional states, and the 7-dimensional state cannot be directly displayed. Therefore, two slices are taken, that is, the 5-dimensional state is fixed as a constant, and the remaining 2-dimensional state is a variable. Through this operation, the mapping relationship between the 7-dimensional state and the 1-dimensional result can be extracted into a 2-dimensional state and a 1-dimensional result, and the mapping relationship can be directly displayed. Therefore, Figure 2 The 5 dimensions in the input are fixed, and the level set function values ​​at different relative positions are displayed in a planar manner.

[0052] Furthermore, the design of the system dynamics equation f takes into account the relative motion characteristics between the ego vehicle and the interacting vehicle, and is the overall system dynamics equation consisting of the ego vehicle and the interacting vehicle.

[0053] The designed state vector x is as follows:

[0054]

[0055] Among them, x r and y r They represent the horizontal and vertical coordinates of the interacting vehicle in the Cartesian coordinate system with the ego vehicle as the origin, i.e., the relative position; Expressed as relative heading angle; v rx 、v ey and v o They are respectively represented as the longitudinal and lateral velocities of the vehicle and the velocity of the interacting vehicle; ω e It is expressed as the yaw angular velocity of the vehicle.

[0056] Furthermore, the state vector gridding granularity is 27×20×9×20×10×20×5, which is uniformly distributed in the 7-dimensional space of state vector x∈[-6,20]×[-10,10]×[0,2π]×[1,20]×[-1,1]×[1,20]×[-1,1].

[0057] In step S2, the acquired vehicle state vectors may be filtered by an extended Kalman filter.

[0058] Furthermore, during the online level set function search, the present invention interpolates between grid points to estimate the level set function value for any given state, meeting real-time requirements. Furthermore, by replacing the most dangerous (for the ego vehicle) control state with the interacting vehicle's current control state, the ego vehicle avoids adopting an overly conservative control strategy while ensuring a safe response.

[0059] Specifically, the method for determining the level set function value in step S3 is as follows: if the current level set function value is less than the threshold p1, a hard control constraint is introduced to prevent the control sequence selected by the controller from causing the ego vehicle to fall into an inevitable collision risk, thereby ensuring interactive safety; if the current level set function value is greater than the threshold p2, a soft control constraint is introduced. When the controller selects an overly conservative control sequence, it will bear a larger objective function value as a penalty, prompting the ego vehicle to adopt a control sequence that is conducive to improving traffic efficiency; if the current level set function value is greater than the threshold p1 and less than the threshold p2, no additional control constraint is introduced.

[0060] Among them, the control hard constraint set and the control soft constraint set are respectively expressed as:

[0061]

[0062] Among them, U e (x) represents the control constraint set under state x, u e is the control quantity, satisfying the physical constraints of the vehicle execution unit:

[0063] u min ≤u e ≤u max

[0064] Step S4 can substitute the obtained control constraints into the MPC solver to solve for safe and smooth control instructions. The objective function and constraints of the MPC used are:

[0065]

[0066] stx(k+1)=A k x(k)+B k u(k)

[0067] y(k)=C k x(k)

[0068] u(k)∈U e [x(k)]+S

[0069] Δu(k)=u(k)-u(k-1)

[0070] y min -ε≤y(k)≤y max +ε

[0071] u min ≤u(k)≤u max

[0072] Δu min ≤Δu(k)≤Δu max

[0073] Where Q, R, ρ, and γ are the output deviation y(k+i)-y ref (k+i), control increment Δu(k+i), relaxation factor ε and weight matrix S; P and C are state time domain and control time domain respectively; x(k), u(k), y(k) are state quantity, control quantity and output quantity at time k respectively; y max 、y min 、u max 、u min , Δu max , Δu min They are the upper and lower limits of output quantity, control quantity and control increment respectively.

[0074] When the control constraint set is a hard constraint, S is zero, that is, the control amount is not allowed to take the set U e [x(k)]; when the control constraint set is a soft constraint, S is any positive value, that is, the control amount is allowed to take the set U e [x(k)], but the larger the value of S is, the larger the value of the objective function will be, which will limit the degree of relaxation during soft constraints.

[0075] Based on the current state of the level set function, this method introduces different control constraint strategies by comparing it with two predefined thresholds, p1 and p2. Unlike conventional methods that simply set the zero level set as the control trigger threshold and trigger the control instruction as a safety-maximizing action, this method incorporates control constraints generated by increasing or decreasing the level set function value into the MPC solution process, ensuring both vehicle efficiency and passenger comfort.

[0076] Example 2

[0077] The main steps of this embodiment are the same as those of embodiment 1, except for the method of solving the level set equation. In this embodiment, the level set function value is obtained by solving the viscosity solution of the Hamilton-Jacobi (HJ) partial differential equation.

[0078] The specific level set equation used is as follows:

[0079]

[0080] in is the level set function that satisfies the Lipschitz continuity condition, f is the system dynamics equation, is the gradient of the level set function.

[0081] The above level set equation is encoded into the terminal value HJ partial differential equation:

[0082]

[0083] in, is the Hamilton function, expressed as:

[0084]

[0085] Always less than or equal to 0, this is to ensure Always greater than or equal to 0, avoiding the level set function The defined reachable set shrinks over time.

[0086] Recorded as It is the level set function at the final value. The final value partial differential equation here is relative to the initial value partial differential equation. The time in the final value partial differential equation is reversed from the current moment. Defined as:

[0087]

[0088] Among them, x r and y r The horizontal and vertical coordinates of the interacting vehicle in a Cartesian coordinate system with the ego vehicle as the origin, i.e., their relative positions, are represented respectively; R represents the minimum safe distance. A larger value of the level set function indicates a safer state.

[0089] The designed state vector is as follows:

[0090]

[0091] Among them, x r and y rThey represent the horizontal and vertical coordinates of the interacting vehicle in the Cartesian coordinate system with the ego vehicle as the origin, i.e., the relative position; Expressed as relative heading angle; v ex 、v ey and v o They are respectively represented as the longitudinal and lateral velocities of the vehicle and the velocity of the interacting vehicle; ω e It is expressed as the yaw angular velocity of the vehicle.

[0092] The state vector gridding granularity is 27×20×9×20×10×20×5, which is uniformly distributed in the 7-dimensional space of state vector x∈[-6,20]×[-10,10]×[0,2π]×[1,20]×[-1,1]×[1,20]×[-1,1].

[0093] Let the control quantity of the interactive vehicle u0=[ω o a o ] T , the control quantity u of the vehicle e =[δF x ] T , combined with the state vector x, we can obtain the overall system dynamics equation f consisting of the vehicle and the interacting vehicle:

[0094]

[0095] Where δ is the front wheel turning angle of the vehicle; F x =F xf +F xr is the longitudinal driving force of the vehicle; I zz is the moment of inertia of the vehicle around the z axis; d f d r are the distances from the vehicle's center of mass to the center of the front and rear axles respectively;

[0096] Substitute the above formula into the HJ partial differential equation to calculate and store the level set function value.

[0097] The vehicle status data such as the position, speed, heading angle, yaw rate of the self-vehicle and the position, speed, heading angle of the interactive vehicle are obtained online and filtered through the extended Kalman filter.

[0098] Interpolation is performed between the grid points to find the value of the level set function at any given state. The interpolation method used is the radial basis function interpolation method, which has the general form:

[0099]

[0100] in is the basis function; κ iis the weight coefficient of the i-th basis function. Considering the relatively complex mapping relationship between the relative state and the level set function and its gradient, the present invention adopts the Gaussian function as the basis function, that is:

[0101]

[0102] Among them ‖xx i ‖ is the Euclidean distance from the i-th point to the center point; σ is the variance of the Gaussian function, which controls the width of the basis function.

[0103] The found level set function value is compared with the set threshold, and different control constraints are obtained according to the state.

[0104] like Figure 3 As shown in the figure, the state determination method is as follows: if the current level set function value is less than the threshold p1, a hard control constraint is introduced to prevent the control sequence selected by the controller from causing the ego vehicle to fall into an inevitable collision risk, thereby ensuring interactive safety; if the current level set function value is greater than the threshold p2, a soft control constraint is introduced. When the controller selects an overly conservative control sequence, it will bear a larger objective function value as a penalty, prompting the ego vehicle to adopt a control sequence that is conducive to improving traffic efficiency; if the current level set function value is greater than the threshold p1 and less than the threshold p2, no additional control constraint is introduced.

[0105] Among them, the control hard constraint set and the control soft constraint set are respectively expressed as:

[0106]

[0107] Among them, U e (x) represents the control constraint set under state x, u e is the control quantity, satisfying the physical constraints of the vehicle execution unit:

[0108] u min ≤u e ≤u max

[0109] Bringing the control constraint set into the MPC solution process, the target problem is to minimize the following cost function under the premise of satisfying the following constraints:

[0110]

[0111] stx(k+1)=A k x(k)+B k u(k)

[0112] y(k)=C k x(k)

[0113] u(k)∈Ue [x(k)]+S

[0114] Δu(k)=u(k)-u(k-1)

[0115] y min -ε≤y(k)≤y max +ε

[0116] u min ≤u(k)≤u max

[0117] Δu min ≤Δu(k)≤Δu max

[0118] Where Q, R, ρ, and γ are the output deviation y(k+i)-y ref (k+i), control increment Δu(k+i), relaxation factor ε and weight matrix S; P and C are state time domain and control time domain respectively; x(k), u(k), y(k) are state quantity, control quantity and output quantity at time k respectively; y max 、y min 、u max 、u min , Δu max , Δu min They are the upper and lower limits of output quantity, control quantity and control increment respectively.

[0119] When the control constraint set is a hard constraint, S is zero, that is, the control amount is not allowed to take the set U e [x(k)]; when the control constraint set is a soft constraint, S is any positive value, that is, the control amount is allowed to take the set U e [x(k)], but the larger the value of S is, the larger the value of the objective function will be, which will limit the degree of relaxation during soft constraints.

[0120] Example 3

[0121] The efficient driving interaction system based on backward reachable set security assurance in the embodiment of the present invention is mainly used to implement the above method embodiment, which mainly includes:

[0122] The discrete table construction module is used to divide the sampled values ​​of the vehicle state vector according to a certain grid size, calculate and store the level set function value offline by solving the level set equation, and construct a mapping relationship between the vehicle state vector and the level set function value to form a discrete table. When the level set function value is all zero, it corresponds to a zero level set, which corresponds to the safe distance boundary of the interacting vehicle relative to the ego vehicle.

[0123] The table lookup module is used to obtain the state vector of the vehicle online and find the corresponding level set function value according to the discrete table;

[0124] The control constraint selection module is used to compare the found level set function value with the set threshold and enter different control constraints based on the comparison result. If it is less than the first threshold, it enters the control hard constraint and does not allow the vehicle to select values ​​other than the control value. If it is greater than the second threshold, it enters the control soft constraint and allows the vehicle to select the control value within a certain range.

[0125] The control instruction generation module is used to generate control instructions according to the control constraints to control the driving of the vehicle.

[0126] Furthermore, the control constraint selection module is further configured to not introduce additional control constraints when the level set function value is between a first threshold and a second threshold.

[0127] Furthermore, the discrete table building module is also specifically used to convert the state quantity Sampling is performed according to the specified upper and lower limits and grid size, and then applied to the level set equation. The resulting level set function values ​​are stored in memory, forming an offline data table. During driving, an extended Kalman filter filters the vehicle state information collected by the sensors. This more accurate state information is used to search the offline table for the corresponding level set function value online. This module trades a small amount of storage space for faster computing speed.

[0128] The control constraint selection module is also used to compare the current level set function value with the set threshold, find its safe range, give different level set function gradient value constraints, and narrow the set that can be adopted by the current control sequence.

[0129] The control constraint is then introduced into the MPC solver through the control instruction generation module, thereby ensuring that the solved control sequence will meet the evolution requirements of the level set function value.

[0130] The functions of each module mainly correspond to the method embodiment and will not be described in detail here.

[0131] The present application also provides a computer-readable storage medium, such as a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., an SD or DX memory), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a disk, an optical disk, a server, an App store, etc., on which a computer program is stored, and when the program is executed by a processor, a corresponding function is implemented. When the computer-readable storage medium of this embodiment is executed by a processor, an efficient driving interaction method based on backward reachable set safety assurance of the method embodiment is implemented.

[0132] It should be pointed out that, according to the needs of implementation, the various steps / components described in this application can be split into more steps / components, or two or more steps / components or partial operations of steps / components can be combined into new steps / components to achieve the purpose of the present invention.

[0133] The size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0134] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention.

Claims

1. An efficient driving interaction method based on backward reachable set security guarantee, characterized by: The following steps are involved: S1. Divide the sampled values ​​of the vehicle state vector according to a certain grid size, calculate and store the level set function value offline by solving the level set equation, and construct a mapping relationship between the vehicle state vector and the level set function value to form a discrete table; when the level set function value is all zero, it corresponds to a zero level set, which corresponds to the safe distance boundary of the interacting vehicle relative to the self-vehicle; S2, obtain the state vector of the vehicle online and find the corresponding level set function value according to the discrete table; S3, compare the found level set function value with the set threshold, and enter different control constraints according to the comparison result; If it is less than the first threshold, it enters the control hard constraint, and the vehicle is not allowed to select a value other than the control amount; if it is greater than the second threshold, it enters the control soft constraint, and the vehicle is allowed to select the control amount within a certain range; S4. Generate control instructions based on the control constraints to control the vehicle's driving.

2. The efficient driving interaction method based on backward reachable set security guarantee according to claim 1 is characterized in that: The level set equation is constructed based on the positional relationship between the ego vehicle and the surrounding interactive vehicles and the overall system dynamics equation composed of the ego vehicle and the interactive vehicles.

3. The efficient driving interaction method based on backward reachable set security guarantee according to claim 1 is characterized in that: The state vector includes the horizontal and vertical coordinates of the interactive vehicle in the Cartesian coordinate system with the ego vehicle as the origin, the relative heading angle, the longitudinal and lateral velocities of the ego vehicle and the velocity of the interactive vehicle, as well as the yaw angular velocity of the ego vehicle.

4. The efficient driving interaction method based on backward reachable set security guarantee according to claim 1 is characterized in that: After the vehicle state vector is obtained in step S2, it is filtered through an extended Kalman filter.

5. The efficient driving interaction method based on backward reachable set security guarantee according to claim 1 is characterized in that: If the level set function value is between the first threshold and the second threshold, no additional control constraints are introduced.

6. The efficient driving interaction method based on backward reachable set security guarantee according to claim 1 is characterized in that: In step S2, the level set function values ​​are interpolated between the grid points to find the level set function value at any given state.

7. The efficient driving interaction method based on backward reachable set security guarantee according to claim 1 is characterized in that: In step S4, the control constraints are specifically substituted into the MPC solver to calculate safe and smooth control instructions.

8. An efficient driving interaction system based on backward reachable set security assurance, characterized by: include: The discrete table construction module is used to divide the sampled values ​​of the vehicle state vector according to a certain grid size, calculate and store the level set function value offline by solving the level set equation, and construct a mapping relationship between the vehicle state vector and the level set function value to form a discrete table. When the level set function value is all zero, it corresponds to a zero level set, which corresponds to the safe distance boundary of the interacting vehicle relative to the ego vehicle. The table lookup module is used to obtain the state vector of the vehicle online and find the corresponding level set function value according to the discrete table; The control constraint selection module is used to compare the found level set function value with the set threshold and enter different control constraints according to the comparison result; If it is less than the first threshold, it enters the control hard constraint, and the vehicle is not allowed to select a value other than the control amount; if it is greater than the second threshold, it enters the control soft constraint, and the vehicle is allowed to select the control amount within a certain range; The control instruction generation module is used to generate control instructions according to the control constraints to control the driving of the vehicle.

9. The efficient driving interaction system based on backward reachable set security guarantee according to claim 8 is characterized in that: The control constraint selection module is further configured to not introduce additional control constraints when the level set function value is between a first threshold and a second threshold.

10. A computer storage medium, characterized in that A computer program executable by a processor is stored therein, and the computer program executes the efficient driving interaction method based on backward reachable set safety assurance according to any one of claims 1 to 7.

Images