Limit state of track vehicle state prediction method and system
By constructing a control model for tracked vehicles and optimizing the parameter set, the problem of state prediction for unmanned tracked vehicles under the limit state of handling stability was solved, achieving high-precision dynamic description and real-time prediction.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BEIJING INST OF TECH
- Filing Date
- 2024-01-15
- Publication Date
- 2026-07-24
AI Technical Summary
Existing motion planning and control algorithms for unmanned tracked vehicles are based on kinematic models, which cannot accurately describe the high dynamic characteristics under extreme handling stability conditions, thus limiting the possibility of high-speed maneuvering.
A control model for tracked vehicles is constructed. By combining the vertical mechanics model, ground contact model, and horizontal dynamics model of tracked vehicles, the parameter set of the control model is optimized. The parameters are then optimized using the multiplexed rolling time-domain estimation method to predict the state vector of the tracked vehicle.
It achieves high-precision prediction of the tracked vehicle state under extreme conditions, taking into account the dynamic performance and ground adhesion characteristics of the tracked vehicle, with low computational load and real-time advantages.
Smart Images
Figure CN117864156B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of tracked vehicle dynamics technology, and in particular to a method and system for predicting the state of tracked vehicles under extreme conditions. Background Technology
[0002] Unmanned tracked vehicles are important special-purpose vehicles capable of performing various engineering and military tasks. High-speed maneuverability allows them to fully utilize the performance of the chassis and road surface, contributing to better mission completion. However, high-speed travel on curved paths results in significant acceleration. The vehicle's state changes rapidly, and the relative motion between the tracks and the ground, as well as the longitudinal and lateral mechanical characteristics, exhibit strong nonlinear and coupled properties. When a large sideslip angle exists between the direction of the vehicle's speed and its orientation, the vehicle experiences an abnormal drifting state, a typical limit state for handling stability.
[0003] Currently, motion planning and control algorithms for unmanned tracked vehicles are almost all based on kinematic models. However, kinematic models rely on the assumption that certain motion quantities do not change over short periods, which cannot accurately describe the high dynamic characteristics of unmanned tracked vehicles under handling stability limits. This limits the possibility of high-speed maneuvering by motion planning and control algorithms based on kinematic models. Therefore, a model that can accurately predict the state of an unmanned tracked vehicle under handling stability limits is essential for the subsequent development of control algorithms based on continuous target states. Summary of the Invention
[0004] The purpose of this invention is to provide a method and system for predicting the state of tracked vehicles under extreme conditions, thereby achieving accurate prediction of the state of tracked vehicles under extreme conditions.
[0005] To achieve the above objectives, the present invention provides the following solution.
[0006] On the one hand, the present invention provides a method for predicting the state of a tracked vehicle under extreme conditions, including the following steps.
[0007] Based on the vertical mechanics model, ground contact model, and horizontal dynamics model of the tracked vehicle, a control model for the tracked vehicle is constructed. The tracked vehicle control model includes control vectors, state vectors, and a control model parameter set. With the control model parameter set fixed, the tracked vehicle control model predicts the state vector at the next moment based on the current state vector and the current control vector. The tracked vehicle vertical mechanics model characterizes the relationship between the vertical load of the tracked vehicle and its movement. The ground contact model characterizes the nonlinear relationship between track slip, vertical load, and horizontal shear force during tracked vehicle movement. The tracked vehicle horizontal dynamics model characterizes the relationship between the horizontal shear force and the horizontal state parameters of the tracked vehicle.
[0008] Obtain the parameter optimization dataset. The parameter optimization dataset includes the control vectors, corresponding state vectors, and corresponding target state vectors for a consecutive number of time steps prior to the current time step; for any given control vector, the corresponding target state vector is the state vector for the next time step.
[0009] Based on the parameter optimization dataset, the control model parameter set in the tracked vehicle control model is optimized to obtain the optimized tracked vehicle control model.
[0010] Based on the current state vector and the current control vector, the optimized tracked vehicle control model is used to predict the state vector at the next moment.
[0011] On the other hand, corresponding to the aforementioned method for predicting the state of a tracked vehicle under extreme conditions, the present invention also provides a system for predicting the state of a tracked vehicle under extreme conditions. When the system is run by a computer, the system for predicting the state of a tracked vehicle under extreme conditions executes the method for predicting the state of a tracked vehicle under extreme conditions as described above.
[0012] According to specific embodiments provided by the present invention, the following technical effects are disclosed.
[0013] This invention provides a method and system for predicting the state of a tracked vehicle under extreme conditions. The method includes: constructing a tracked vehicle control model based on the tracked vehicle's vertical mechanical model, ground contact model, and horizontal dynamic model; obtaining the control vectors, corresponding state vectors, and target state vectors for several consecutive moments prior to the current moment as a parameter optimization dataset; optimizing the control model parameter set in the constructed tracked vehicle control model to obtain a tracked vehicle control model with optimized control model parameter set; and after the control model parameter set is fixed, accurately predicting the tracked vehicle's state vector at the next moment based on the tracked vehicle's current state vector and control vector at the current moment. Compared to existing motion planning and control algorithms based on kinematic models, the method of this invention considers the variation of the vertical load of the tracked vehicle with the movement of the tracked vehicle, the nonlinear relationship among the horizontal shear force, track slip, and vertical load of the tracked vehicle, and the relationship between the horizontal shear force and the horizontal state parameters of the tracked vehicle. In the constructed tracked vehicle control model, apart from the control model parameter set to be optimized, other parameters are fixed and easily measurable dimensional and inertial parameters. It fully considers the dynamic performance of the tracked vehicle and the adhesion characteristics between the tracked vehicle and the ground, and can achieve a high-precision dynamic description of unmanned tracked vehicles. At the same time, the parameters of the tracked vehicle control model are optimized using data from several consecutive moments before the current moment, which makes the method computationally intensive and has a strong real-time advantage. Attached Figure Description
[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0015] Figure 1 The flowchart shows a method for predicting the state of a tracked vehicle under extreme conditions, as provided in Embodiment 1 of the present invention.
[0016] Figure 2 This is a flowchart of steps A1 to A3 in the tracked vehicle state prediction method provided in Embodiment 1 of the present invention.
[0017] Figure 3 This is a schematic diagram illustrating the force and motion analysis performed in the tracked vehicle state prediction method provided in Embodiment 1 of the present invention.
[0018] Figure 4 This is a schematic diagram of the nonlinear relationship represented by the simplified magic formula in the tracked vehicle state prediction method provided in Embodiment 1 of the present invention.
[0019] Figure 5 This is a schematic diagram illustrating the analysis of horizontal dynamic parameters in the tracked vehicle state prediction method provided in Embodiment 1 of the present invention.
[0020] Figure 6 This is a schematic diagram of the structure of a tracked vehicle state prediction system under extreme conditions provided in Embodiment 2 of the present invention. Detailed Implementation
[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0022] The purpose of this invention is to provide a method and system for predicting the state of tracked vehicles under extreme conditions. It fully considers the dynamic performance of tracked vehicles and the adhesion characteristics between tracked vehicles and the ground, and can realize high-precision dynamic description of unmanned tracked vehicles, and achieve accurate prediction of the state of tracked vehicles under extreme conditions.
[0023] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0024] Example 1.
[0025] This embodiment provides a method for predicting the state of a tracked vehicle under extreme conditions, as shown in the attached figure. Figure 1 The flowchart shown illustrates that the method in this embodiment includes the following steps S1 to S4.
[0026] S1. Based on the vertical mechanical model, ground contact model, and horizontal dynamics model of the tracked vehicle, a control model for the tracked vehicle is constructed. The tracked vehicle control model includes control vectors, state vectors, and a control model parameter set. When the control model parameter set is fixed, the tracked vehicle control model predicts the state vector at the next moment based on the current state vector and the current control vector. The tracked vehicle vertical mechanical model characterizes the relationship between the vertical load of the tracked vehicle and its movement. The ground contact model characterizes the nonlinear relationship between track slip, vertical load, and horizontal shear force during tracked vehicle movement. The tracked vehicle horizontal dynamics model characterizes the relationship between the horizontal shear force and the horizontal state parameters of the tracked vehicle.
[0027] S2. Obtain the parameter optimization dataset. The parameter optimization dataset includes the control vectors, corresponding state vectors, and corresponding target state vectors for a series of consecutive time steps prior to the current time step; for any given control vector, the corresponding target state vector is the state vector for the next time step.
[0028] S3. Based on the parameter optimization dataset, optimize the control model parameter set in the tracked vehicle control model to obtain the optimized tracked vehicle control model. In step S3, when optimizing the control model parameter set in the tracked vehicle control model based on the parameter optimization dataset, a traditional optimization algorithm can be used to perform overall optimization of all control model parameters in the control model parameter set using each parameter optimization data point in the parameter optimization dataset.
[0029] In this embodiment, a multiplexed rolling time-domain estimation method is used to obtain the optimized tracked vehicle control model. First, the control model parameter set is divided into several subsets. Then, each estimation uses only the most recent N frames of data to optimize a specific intermediate subset of control model parameters, until the objective function of the tracked vehicle control model obtained from the optimized control model parameter set converges. During the multiplexed rolling time-domain estimation process, the parameter optimization dataset is continuously updated like a sliding window, taking only data from the N time steps prior to the current time step. Specifically, step S3 includes...
[0030] S31. Divide the control model parameter set into several intermediate control model parameter subsets. Each intermediate control model parameter subset includes at least one control model parameter; there is no overlap between the intermediate control model parameter subsets.
[0031] S32. Take the first intermediate control model parameter subset of the control model parameter set as the current control model parameter subset, and take the parameter optimization dataset as the current parameter optimization dataset.
[0032] S33. Optimize the control model parameters in the current control model parameter subset based on the current parameters, and update the control model parameter set based on the optimized current control model parameter subset. Step S33, optimizing the control model parameters in the current control model parameter subset based on the current parameters, specifically includes:
[0033] S331. Take the first parameter optimization data in the current parameter optimization dataset as the current parameter optimization data; the control vector, the corresponding state vector, and the corresponding target state vector at any time in the current parameter optimization dataset form a parameter optimization data.
[0034] S332. Using the control vector and corresponding state vector of the current parameter optimization data as input to the tracked vehicle control model, and using the corresponding target state vector as the target output of the tracked vehicle control model, optimize the control model parameters in the current control model parameter subset.
[0035] S333: Take the next parameter optimization data in the current parameter optimization dataset as the current parameter optimization data, take the optimized current control model parameter subset as the current control model parameter subset, and jump to step S332: Use the control vector and the corresponding state vector of the current parameter optimization data as the input of the tracked vehicle control model, and use the corresponding target state vector as the target output of the tracked vehicle control model to optimize the control model parameters in the current control model parameter subset until the last parameter optimization data in the current parameter optimization dataset completes the optimization of the control model parameters in the current control model parameter subset.
[0036] S34. Determine whether the objective function of the tracked vehicle control model determined based on the updated control model parameter set has converged, and obtain the first judgment result.
[0037] S35. If the first judgment result is yes, then the tracked vehicle control model determined according to the updated control model parameter set will be determined as the optimized tracked vehicle control model.
[0038] S36. If the first judgment result is negative, then the next intermediate control model parameter subset of the current control model parameter subset of the control model parameter set is taken as the current control model parameter subset, the parameter optimization dataset is updated, and the updated parameter optimization dataset is taken as the current parameter optimization dataset. Then, proceed to step S33: optimize the control model parameters in the current control model parameter subset according to the current parameter optimization dataset, and update the control model parameter set according to the optimized current control model parameter subset. Then, determine whether the objective function of the tracked vehicle control model determined according to the updated control model parameter set has converged, and obtain the first judgment result. Wherein, if the current control model parameter subset is the last intermediate control model parameter subset and the first judgment result is negative, then the first intermediate control model parameter subset of the updated control model parameter set is taken as the current control model parameter subset. The updated parameter optimization dataset is the dataset obtained by updating the parameter optimization dataset with the control vectors, corresponding state vectors, and corresponding target state vectors of several consecutive times before the current time.
[0039] Specifically, the tracked vehicle control model constructed in this embodiment is shown in equation (1).
[0040]
[0041] In equation (1), Let represent the state vector differentiated with respect to time, represent the changed state vector, and x represent the state vector. R represents the yaw rate, β represents the sideslip angle of the tracked vehicle's center of gravity, V represents the horizontal speed, X represents the X-axis coordinate of the tracked vehicle, and Y represents the Y-axis coordinate of the tracked vehicle. The yaw angle, u, represents the yaw angle in the direction of the tracked vehicle body; the standard control vector is u = [ω]. l ω r ] T ω l ω represents the rotational speed of the left drive wheel. r This represents the rotational speed of the right-hand drive wheel, p represents the control model parameter set, p = [D x C x B x D y C y B y ] T D x D represents the first longitudinal track contact parameter. y C represents the first lateral ground contact parameter. x C represents the second longitudinal track contact parameter. y B represents the second lateral track contact parameter. x B represents the third longitudinal track contact parameter. yLet T represent the third lateral ground contact parameter, and T represent the transpose of the matrix.
[0042] The objective function of the tracked vehicle control model is shown in equation (2).
[0043]
[0044] In equation (2), Δp represents the change in the parameter set of the control model, and J represents the value of the objective function. Let Δp represent the minimum value of the objective function J, i represent the index of the parameter optimization data (i ∈ [kN, k-1]), N represent the size of the current parameter optimization dataset, k represent the current optimization round, n represent the number of intermediate control model parameter subsets (n is less than or equal to the number of control model parameters), and σ(k) represent the index of the intermediate control model parameter subset being optimized in round k. mdl (i) represents the actual output of the tracked vehicle control model when optimizing using the data from the i-th parameter, x real (i) represents the target output of the tracked vehicle control model when optimizing using the data from the i-th parameter, where Q and R are weight coefficient matrices. Represents the state vector x mdl Differentiating with respect to time, f(·) represents the control model of the tracked vehicle, x mdl u represents the state vector input into the tracked vehicle control model. real Let p(k) represent the control vector input to the tracked vehicle control model, p(k) represent the control model parameter set at optimization k rounds, RK4(·) represent the fourth-order Runkkuta numerical integration algorithm, t be the time interval, and u be the input vector. real (i) represents the control vector input into the tracked vehicle control model when optimizing using the data of the i-th parameter, and Δp(k) represents the change in the parameter set of the control model during the k-th round of optimization. min p represents the minimum value of the control model parameter set. max This represents the maximum value of the control model parameter set.
[0045] The objective function in equation (2) represents the asynchronous update rate used in this embodiment. That is, the parameters of the control model are not estimated simultaneously as commonly used, but are estimated sequentially. For example, among the six parameters, C x B x C y and B y In nested trigonometric functions, the nonlinearity is strong and the computational load is large; therefore, a four-step approach is used to update these four parameters separately. And D... x and D yOutside the trigonometric functions, only the proportionality coefficient is considered, resulting in a relatively small computational burden. Therefore, these two parameters are estimated simultaneously. Thus, assigning n a value of 5 indicates that 5 steps of parameter estimation constitute one loop, with each loop estimating C sequentially. x B x C y B y and (D) x D y That is, when n=1, C x It is a variable, B x C y B y and (D) x D y ) is a quantitative measure; when n=2, B x It is a variable, C x C y B y and (D) x D y ) is a quantitative value; when n=3, C y It is a variable, B x C x B y and (D) x D y ) is a quantitative measure; when n=4, B y It is a variable, C x B x C y and (D) x D y () is a quantitative measure; when n=5, (D) x D y ) is a variable, C x B x C y and B y It is a quantitative measure.
[0046] S4. Based on the current state vector and control vector, predict the state vector for the next moment using the optimized tracked vehicle control model. The optimized tracked vehicle control model can be used as a reference to analyze and predict the normal driving and drifting of unmanned tracked vehicles.
[0047] Before constructing the tracked vehicle control model in step S1 based on the tracked vehicle's vertical mechanical model, ground contact model, and horizontal dynamic model, as follows: Figure 2 As shown, the method in this embodiment also includes the following steps A1 to A3.
[0048] A1. Based on the analyzed force conditions of the tracked vehicle when stationary and in motion in the vertical direction, construct a vertical mechanical model of the tracked vehicle. The contact between the track and the ground is very complex, related to the track structure, track tension, road wheel structure, terrain, and ground surface. However, this distributed load is relatively concentrated under the road wheels. Ignoring other effects, we mainly consider the effects of gravity and concentrated loads at the road wheels on the tracked vehicle. The force and motion analysis is as follows: Figure 3 As shown, both forces and kinematic quantities are analyzed in the vehicle's body coordinate system, such as... Figure 3 As shown, when the tracked vehicle accelerates, the rear track is the tight side and the front track is the loose side. F x,d,j This refers to the longitudinal force on the tracks. During braking, the front track is the tight side, and the rear track is the loose side. The tracked vehicle experiences horizontal force, and of the four wheels in contact with the ground, from front to back, they are the 1st to 4th road wheels. x and a y These are longitudinal and lateral accelerations, respectively. G is the track center distance, and r is the longitudinal and lateral accelerations. d,j It is the radius vector from the center of gravity of the tracked vehicle to the track plate below the road wheel, x CM x1 is the longitudinal distance from the driving wheel to the center of mass, and x2, x3 and x4 are the longitudinal distances from the driving wheel to the first to fourth load wheels, respectively.
[0049] The force analysis of the tracked vehicle in the vertical plane is performed, and the concentrated vertical load is calculated as shown in equation (1). Vertical load F z,j The static load F originates from three aspects: the shape and mass distribution of the tracked vehicle. z,j,0 The load change caused by acceleration (including k caused by longitudinal acceleration) j a x and ΔF caused by lateral acceleration z ), and the load changes caused by track tension. The vertical mechanical model of the tracked vehicle can be constructed as shown in equation (3).
[0050]
[0051] In equation (3), F z,j F represents the vertical load of a tracked vehicle. z,j,0 k represents the static load of a tracked vehicle. j a represents the proportionality coefficient that causes the change in vertical load due to longitudinal acceleration. x k represents longitudinal acceleration. ten F represents the proportionality coefficient that causes the change in vertical load due to lateral acceleration. x,d,j This represents the longitudinal force on the j-th load-bearing wheel on the x-axis on side d, where d takes the value l or r, l represents the left side and r represents the right side, ΔF zThe expression represents the change in vertical load on the left and right tracks caused by lateral acceleration, where m represents the mass of the tracked vehicle, and a represents the change in vertical load on the left and right tracks. y Let f represent lateral acceleration, h represent the height of the tracked vehicle's center of gravity, G represent the track center moment, g represent gravitational acceleration, and F represent the acceleration due to gravity. z,l,j F represents the vertical load on the j-th road wheel on the left. z,r,j This represents the vertical load on the j-th road wheel on the right.
[0052] A2. Based on the analyzed and determined contact adhesion characteristics between the tracked vehicle and the ground, a track-ground contact model is constructed. In this embodiment, a simplified magic formula as shown in equation (4) is used to represent the nonlinear relationship between track slippage and track shear force. The nonlinear relationship represented by equation (4) is as follows: Figure 4 As shown, the horizontal axis represents slip, and the vertical axis represents shear force.
[0053] F = D sin[Carctan(Bs)] (4).
[0054] In equation (4), F is the shear force in the x-direction or y-direction (F x,d,j or F y,d,j D, C, and B are three model parameters that affect the nonlinear characteristics of the ground contact. D is the peak factor. x and D y These are two components of D, where C is the shape factor. x and C y C consists of two components, and B is the stiffness factor. x and B y These are the two components of B, where s is the relative slippage between the track and the ground in the x or y direction (s x,d,j or s y,d,j The important features of the curve represented by equation (4) are shown in equation (5).
[0055]
[0056] In equation (5), k is the slope (stiffness) at the origin, which is the product of D, C, and B, and F m It is the maximum force, equal to the peak factor, s m It is the corresponding slip, F ∞ It is the force limit value of infinite slip ratio, and the ground contact model can be constructed as shown in equation (6).
[0057]
[0058] In equation (6), s x,d,j s represents the longitudinal slippage of the track of the j-th road wheel on side d. y,d,jThis represents the lateral slippage of the j-th road wheel on side d, where d takes the value l or r, l represents the left side and r represents the right side. x,d,j v represents the longitudinal translational velocity of the j-th road wheel track on side d caused by the movement of the tracked vehicle. y,d,j ω represents the lateral translational velocity of the j-th road wheel track on side d caused by the movement of the tracked vehicle. d The speed of the driving wheel on side d is represented by r. d s represents the radius of the driving wheel on side d. i,j F represents the combined longitudinal and lateral slippage of the track. x0,d,j F represents the shear force under pure longitudinal slip condition on side d. y0,d,j D represents the shear force under pure lateral slip condition on side d, μ represents the road adhesion coefficient, and D x D represents the first longitudinal track contact parameter. y C represents the first lateral ground contact parameter. x C represents the second longitudinal track contact parameter. y B represents the second lateral track contact parameter. x B represents the third longitudinal track contact parameter. y F represents the third lateral track contact parameter. z,d,j F represents the vertical load on the j-th road wheel on side d. x,d,j F represents the shear force under the combined longitudinal slippage of the j-th load-bearing wheel on side d. y,d,j It represents the shear force under the combined working condition of lateral slippage of the j-th load wheel on side d.
[0059] A3. Based on the relationship between the horizontal shear force of the road wheels and the dynamic parameters of the tracked vehicle on the horizontal plane determined by the analysis, a dynamic model of the tracked vehicle on the horizontal plane is constructed. In this embodiment, the horizontal shear force F calculated in step A1 is used. x,d,j and F y,d,j The input can be used to perform horizontal dynamics calculations for tracked vehicles, as shown in the schematic diagram. Figure 3 and Figure 5 . Figure 5 In the diagram, X and Y are the vertical and horizontal coordinates in the geodetic coordinate system, respectively, with the positive direction of the X-axis being eastward; v x,d,j and v y,d,j These are the longitudinal and lateral translational velocities of the track plates caused by the vehicle's movement; a t and a n These are tangential acceleration and normal acceleration, respectively; V is the horizontal vehicle speed; φ and These are the heading angle and heading angular velocity in the direction of the tracked vehicle's speed, respectively. and These are the yaw angle and yaw rate in the direction of the tracked vehicle body; β and These are the sideslip angle and sideslip velocity, respectively. The dynamic model of the tracked vehicle on the horizontal plane is shown in equation (7).
[0060]
[0061] In equation (7), F represents the principal vector acting on the tracked vehicle, β represents the sideslip angle of the tracked vehicle's center of gravity, and F hor,d,j F represents the track shear force vector under the combined longitudinal and lateral slippage of the j-th road wheel on side d. x,d,j and F y,d,j For F hor,d,j The component d takes the value l or r, where l represents the left side and r represents the right side, j represents the subscript of the road wheel, and M z The principal moment acting on a tracked vehicle is represented by r. d,j a represents the radius vector from the center of gravity of the tracked vehicle to the track plate of the j-th road wheel on side d. t F represents tangential acceleration. t Let m represent the tangential force, and a represent the mass of the tracked vehicle. n F represents the normal acceleration. n F represents the normal force. t and F n Let F have two components, V representing the horizontal vehicle speed, and I... z φ represents the yaw moment of inertia of the tracked vehicle, and φ represents the heading angle in the direction of the vehicle's velocity. R represents the yaw angle in the direction of the tracked vehicle body, X represents the X-axis coordinate of the tracked vehicle, Y represents the Y-axis coordinate of the tracked vehicle, and R represents the yaw rate.
[0062] The method provided in this embodiment considers the variation of the vertical load of the tracked vehicle with the movement of the tracked vehicle, the nonlinear relationship among the horizontal shear force, track slip, and vertical load of the tracked vehicle, and the relationship between the horizontal shear force and the horizontal plane state parameters of the tracked vehicle. In the tracked vehicle control model constructed, apart from the control model parameter set to be optimized, other parameters are fixed and easily measurable dimensional and inertial parameters. It fully considers the dynamic performance of the tracked vehicle and the adhesion characteristics between the tracked vehicle and the ground, and can realize a high-precision dynamic description of unmanned tracked vehicles. At the same time, the multiplexing rolling time-domain estimation method is used to optimize the control model parameter set of the tracked vehicle control model. Each optimization only needs to use data from several consecutive time steps before the current time to optimize a batch of parameters of the tracked vehicle control model, which makes the method computationally small and has a strong real-time advantage.
[0063] Example 2.
[0064] Furthermore, the method of Embodiment 1 of the present invention can also be used by means of Figure 6 The architecture of the tracked vehicle state prediction system under the extreme conditions shown is implemented as follows. Figure 6As shown, the tracked vehicle state prediction system may include a tracked vehicle vertical mechanics model construction module M1, a track-ground contact model construction module M2, a tracked vehicle horizontal plane dynamics model construction module M3, a tracked vehicle control model construction module M4, a parameter optimization dataset acquisition module M5, a control model parameter set optimization module M6, and a tracked vehicle state vector prediction module M7; some of these modules may also have sub-units for implementing their functions. Of course, Figure 6 The architecture shown is merely exemplary; it can be omitted as needed when implementing different functionalities. Figure 6 One or at least two components of the system shown.
[0065] Specific examples are used in this article, but the above description is only to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. Those skilled in the art should understand that the various modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, and thus, they can be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any specific combination of hardware and software.
[0066] Furthermore, those skilled in the art will recognize that, based on the principles of this invention, there will be variations in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as limiting the invention.
Claims
1. A method for predicting the state of a tracked vehicle under extreme conditions, characterized in that, include: A tracked vehicle control model is constructed based on the tracked vehicle's vertical mechanics model, ground contact model, and horizontal dynamics model. The tracked vehicle control model includes a control vector, a state vector, and a control model parameter set. When the control model parameter set is fixed, the tracked vehicle control model predicts the state vector at the next moment based on the current state vector and the current control vector. The tracked vehicle vertical mechanics model characterizes the relationship between the vertical load of the tracked vehicle and its movement. The ground contact model characterizes the nonlinear relationship between track slip, vertical load, and horizontal shear force during tracked vehicle movement. The tracked vehicle horizontal dynamics model characterizes the relationship between the horizontal shear force and the horizontal state parameters of the tracked vehicle. Obtain a parameter optimization dataset; the parameter optimization dataset includes control vectors, corresponding state vectors, and corresponding target state vectors for a series of consecutive time steps prior to the current time step; for any control vector at any time step, the corresponding target state vector is the state vector for the next time step of that time step; Based on the parameter optimization dataset, the control model parameter set in the tracked vehicle control model is optimized to obtain the optimized tracked vehicle control model. Based on the current state vector and the current control vector, the optimized tracked vehicle control model is used to predict the state vector at the next moment. Based on the parameter optimization dataset, the control model parameter set in the tracked vehicle control model is optimized, specifically including: The control model parameter set is divided into several intermediate control model parameter subsets; each intermediate control model parameter subset includes at least one control model parameter; there is no overlap between the intermediate control model parameter subsets. The first intermediate subset of control model parameters in the control model parameter set is taken as the current subset of control model parameters, and the parameter optimization dataset is taken as the current parameter optimization dataset. Based on the current parameter optimization dataset, optimize the control model parameters in the current control model parameter subset, update the control model parameter set based on the optimized current control model parameter subset, and then determine whether the objective function of the tracked vehicle control model determined based on the updated control model parameter set has converged, and obtain the first judgment result. If the first judgment result is yes, then the tracked vehicle control model determined according to the updated control model parameter set will be determined as the optimized tracked vehicle control model. If the first judgment result is negative, then the next intermediate control model parameter subset of the current control model parameter subset of the control model parameter set is taken as the current control model parameter subset, the parameter optimization dataset is updated, and the updated parameter optimization dataset is taken as the current parameter optimization dataset. Then, the process jumps to the following steps: optimize the control model parameters in the current control model parameter subset according to the current parameter optimization dataset, update the control model parameter set according to the optimized current control model parameter subset, and then determine whether the objective function of the tracked vehicle control model determined according to the updated control model parameter set has converged, and obtain the first judgment result, until the objective function of the tracked vehicle control model determined according to the updated control model parameter set has converged. Wherein, if the current control model parameter subset is the last intermediate control model parameter subset and the first judgment result is negative, then the first intermediate control model parameter subset of the updated control model parameter set is taken as the current control model parameter subset; the updated parameter optimization dataset is the dataset obtained by updating the parameter optimization dataset with the control vector, the corresponding state vector and the corresponding target state vector of several consecutive times before the current time.
2. The method for predicting the state of a tracked vehicle under extreme conditions according to claim 1, characterized in that, Based on the current parameter optimization dataset, optimize the control model parameters in the current control model parameter subset, specifically including: The first parameter optimization data in the current parameter optimization dataset is taken as the current parameter optimization data; the control vector, the corresponding state vector, and the corresponding target state vector at any time in the current parameter optimization dataset constitute a parameter optimization data; The control model parameters in the current control model parameter subset are optimized by using the control vector and the corresponding state vector of the current parameter optimization data as inputs to the tracked vehicle control model and the corresponding target state vector as the target output of the tracked vehicle control model. The next parameter optimization data in the current parameter optimization dataset is taken as the current parameter optimization data, and the optimized subset of current control model parameters is taken as the current control model parameter subset. Then, the process jumps to the step: using the control vector and the corresponding state vector of the current parameter optimization data as the input of the tracked vehicle control model, and using the corresponding target state vector as the target output of the tracked vehicle control model, the control model parameters in the current control model parameter subset are optimized until the last parameter optimization data in the current parameter optimization dataset completes the optimization of the control model parameters in the current control model parameter subset.
3. The method for predicting the state of a tracked vehicle under extreme conditions according to claim 1, characterized in that, The objective function of the tracked vehicle control model is shown in the following equation: ； in, Δp This represents the amount of change in the parameter set of the control model. J This represents the value of the objective function. Represents the value of the objective function J Minimum Δp , i Indicates the label of the parameter optimization data. i The value is [ k - N , k -1], N This indicates the size of the current parameter optimization dataset. k Indicates the current optimization round. n This indicates the number of the intermediate control model parameter subsets. n The number of control model parameters is less than or equal to σ ( k )express k During round-based optimization, the labels of the subset of intermediate control model parameters being optimized are used. x mdl ( i ) indicates the use of the first i When optimizing the parameter optimization data, the actual output of the tracked vehicle control model is... x real ( i ) indicates the use of the first i When optimizing the parameter optimization data, the target output of the tracked vehicle control model is... Q and R This is the weight coefficient matrix. State vector x mdl Differentiate with respect to time, f (·) indicates the control model for tracked vehicles. x mdl This represents the state vector input into the tracked vehicle control model. u real This represents the control vector input into the tracked vehicle control model. p ( k )express k The parameter set of the control model during round optimization. RK 4(·) denotes the fourth-order Runkkuta numerical integration algorithm. t For time intervals, u real ( i ) indicates the use of the first i When optimizing the parameter optimization data, the control vector input into the tracked vehicle control model is used. Δp ( k )express k During each round of optimization, the amount of change in the model parameter set is controlled. p min This represents the minimum value of the control model parameter set. p max This represents the maximum value of the control model parameter set.
4. The method for predicting the state of a tracked vehicle under extreme conditions according to claim 1, characterized in that, Before constructing the tracked vehicle control model based on the tracked vehicle's vertical mechanical model, ground contact model, and horizontal dynamic model, the tracked vehicle state prediction method under extreme conditions further includes: Based on the analysis of the forces acting on the tracked vehicle when it is stationary and when it is in motion in the vertical direction, a vertical mechanical model of the tracked vehicle is constructed. Based on the analysis and determination of the contact and adhesion characteristics between the tracked vehicle and the ground, a track-ground contact model is constructed. Based on the relationship between the horizontal shear force of the road wheels and the dynamic parameters of the tracked vehicle on the horizontal plane determined by analysis, a dynamic model of the tracked vehicle on the horizontal plane is constructed.
5. The method for predicting the state of a tracked vehicle under extreme conditions according to claim 1, characterized in that, The vertical mechanical model of the tracked vehicle is shown in the following equation: ； in, F z,j This indicates the vertical load on a tracked vehicle. F z,j,0 This indicates the static load of a tracked vehicle. k j This represents the proportionality coefficient that causes the vertical load to change due to longitudinal acceleration. a x Indicates longitudinal acceleration. k ten This represents the proportionality coefficient that causes the change in vertical load due to lateral acceleration. F x,d,j express d Side j One load-bearing wheel x Longitudinal force on the shaft, d Values l or r , l Represents the left side. r Represents the right side. ΔF z This represents the change in vertical load on the left and right tracks caused by lateral acceleration. m Indicates the mass of tracked vehicles. a y Indicates lateral acceleration. h Indicates the height of the center of gravity of a tracked vehicle. G Indicates the track center distance, g Represents gravitational acceleration. F z,l,j Indicates the leftmost j Vertical load of each road wheel F z,r,j Indicates the right-hand side j Vertical load of each road wheel.
6. The method for predicting the state of a tracked vehicle under extreme conditions according to claim 1, characterized in that, The ground contact model is shown in the following formula: ； in, s x,d,j express d Side j The longitudinal slippage of the track of each road wheel s y,d,j express d Side j Lateral slippage of the track of each road wheel d Values l or r , l Represents the left side. r Represents the right side. v x,d,j Indicates the movement caused by tracked vehicles d Side j The longitudinal translational speed of each road wheel track v y,d,j Indicates the movement caused by tracked vehicles d Side j Lateral translation speed of each road wheel track ω d express d Side drive wheel speed, r d express d Side drive wheel radius, s i,j This represents the combined longitudinal and lateral slippage of the tracks. F x0,d,j express d Shear force under pure longitudinal slip condition, F y0,d,j express d Shear force under pure lateral slip condition μ Indicates the road surface adhesion coefficient. D x This represents the first longitudinal track contact parameter. D y This represents the first lateral track contact parameter. C x This indicates the second longitudinal track contact parameter. C y This indicates the second lateral track contact parameter. B x This indicates the third longitudinal track contact parameter. B y This indicates the third lateral track contact parameter. F z,d,j express d Side j Vertical load of each road wheel F x,d,j express d Side j Shear force under combined longitudinal slippage of a single load-bearing wheel F y,d,j express d Side j Shear force under combined lateral slippage of one load-bearing wheel.
7. The method for predicting the state of a tracked vehicle under extreme conditions according to claim 1, characterized in that, The horizontal dynamics model of the tracked vehicle is shown in the following equation: ； in, F This represents the principal vector acting on the tracked vehicle. β Indicates the sideslip angle of the center of gravity of a tracked vehicle. F hor,d,j express d Side j Track shear force vector under combined longitudinal and lateral slippage of the road wheels F x,d,j and F y,d,j for F hor,d,j The amount, d Values l or r , l Represents the left side. r Represents the right side. j Indicates the subscript of the road wheel. M z This indicates the principal moment acting on a tracked vehicle. r d,j Indicates the center of gravity of the tracked vehicle d Side j The radius of the track pad of each road wheel a t Indicates tangential acceleration. F t Indicates tangential force. m Indicates the mass of tracked vehicles. a n Indicates normal acceleration. F n Represents normal force, F t and F n for F The two components, V Indicates horizontal vehicle speed. I z This represents the yaw moment of inertia of a tracked vehicle. φ The heading angle, representing the direction of speed of a tracked vehicle. The yaw angle, indicating the lateral direction of the tracked vehicle. X Indicating tracked vehicles X Axis coordinates Y Indicating tracked vehicles Y Axis coordinates R This represents the yaw rate.
8. The method for predicting the state of a tracked vehicle under extreme conditions according to claim 1, characterized in that, The control model for the tracked vehicle is shown in the following equation: ； in, This represents the derivative of the state vector with respect to time. x Represents the state vector. x =[ R β VXY ] T , R Indicates yaw rate. β Indicates the sideslip angle of the center of gravity of a tracked vehicle. V Indicates horizontal vehicle speed. X Indicating tracked vehicles X Axis coordinates Y Indicating tracked vehicles Y Axis coordinates The yaw angle, indicating the lateral direction of the tracked vehicle. u Standard control vector, u =[ ω l ω r ] T , ω l Indicates the rotational speed of the left drive wheel. ω r Indicates the rotational speed of the right-hand drive wheel. p Represents the control model parameter set. p =[ D x C x B x D y C y B y ] T , D x This represents the first longitudinal track contact parameter. D y This represents the first lateral track contact parameter. C x This indicates the second longitudinal track contact parameter. C y This indicates the second lateral track contact parameter. B x This indicates the third longitudinal track contact parameter. B y This indicates the third lateral track contact parameter. T This represents the transpose of a matrix.
9. A tracked vehicle state prediction system under extreme conditions, characterized in that, When the tracked vehicle state prediction system under extreme conditions is run by a computer, it executes a tracked vehicle state prediction method under extreme conditions as described in any one of claims 1-8.