Reconfigurable vehicle cooperative control method of a double prediction controller

Abstract

This invention relates to the field of vehicle automatic control and provides a cooperative control method for reconfigurable vehicles using dual predictive controllers. The method includes: designing a reconfigurable vehicle observer based on extended Kalman filtering, dynamically adjusting the state vector and system matrix dimensions to adapt to module additions or subtractions; establishing a distributed vehicle dynamics model and model library, dynamically configuring degrees of freedom according to the number of modules and implementing configuration mapping through a parameterized interface; and constructing a hierarchical coupled controller, where the upper-level model predicts and controls the output of the total longitudinal force and additional yaw moment, while the lower-level quadratic programming dynamically allocates the driving torque to each hub motor. This invention forms a closed loop of "state perception - model reconstruction - control optimization," achieving cooperative control with small state errors and multi-objective dynamic balance, suitable for reconfigurable vehicles in scenarios such as emergency rescue and polar exploration.

Description

Technical Field

[0001] This invention belongs to the field of vehicle automatic control, and particularly relates to a reconfigurable vehicle cooperative control method with dual predictive controllers. Background Technology

[0002] With the increasing demands for adaptability to complex environments in civilian scenarios such as emergency rescue, polar exploration, and logistics transportation, traditional fixed-configuration vehicles are struggling to meet the challenges of dynamic task switching and diverse terrain. Multi-configuration switchable intelligent mobility platforms, through dynamic adjustments to mechanical structures (such as modular reconfiguration, wheel / track switching, and wheelbase extension), achieve flexible adaptation of functions and performance, becoming a core technology direction for improving cross-domain operational capabilities. Among these, reconfigurable vehicles, due to their modular autonomous reconfiguration capabilities, can overcome the bottlenecks in adaptability to complex terrain and heterogeneous tasks, possessing broad application prospects and becoming a current research hotspot. Reconfigurable vehicles need to address the abrupt changes in mass distribution, inertia parameters, and driving characteristics caused by dynamic reconfiguration, requiring the design of adaptive state observation, dynamic modeling, and control strategies to achieve efficient collaborative control.

[0003] When applied to reconfigurable vehicles, existing technologies suffer from the following technical problems: In terms of state observation, fixed-dimensional observers cannot adapt to changes in the dimension of the state vector caused by the dynamic addition or removal of modules, resulting in insufficient adaptability to parameter mutations and poor real-time performance; in terms of dynamic modeling, the lack of a distributed parametric modeling mechanism makes it impossible to dynamically adjust the model's degrees of freedom based on the number of modules and connection states, leading to a mismatch between the model and the real-time topology; in terms of control strategies, existing hierarchical control schemes suffer from problems such as rigid static rule bases and missing hard constraints, making it impossible to dynamically coordinate multi-objective optimizations such as trajectory tracking, stability, and energy efficiency when the number of modules is adjusted, resulting in conflicting control objectives. Summary of the Invention

[0004] The purpose of this invention is to provide a reconfigurable vehicle cooperative control method with dual predictive controllers, which aims to solve the technical problems existing in the prior art as identified in the background art.

[0005] This invention is implemented as follows: a reconfigurable vehicle cooperative control method with dual predictive controllers, the method comprising:

[0006] Step 1:

[0007] 1. Construct a four-dimensional state vector containing dynamic state, driving state, energy state, and reconfiguration state:

[0008] 2. Simultaneously construct a nonlinear state-space model:

[0009] 3. Extended Kalman filtering is used for nonlinear state estimation;

[0010] 4. Dynamically adjust the state vector and system matrix when modules are added or removed. and Dimensions.

[0011] Step 2:

[0012] 1. Decompose the reconfigurable vehicle into independent mobility modules, connection modules, and mission compartments, and define the physical parameters of each module;

[0013] 2. Establish the dynamic equations for a single module and the coupled dynamic model between modules;

[0014] 3. Determine the system's degrees of freedom based on the number of modules;

[0015] 4. Input the number of modules n and the wheelbase through the parameterized interface. The connection type (symmetric / asymmetric) is selected, and the automatically generated topology is output. Constraints are assigned between modules, and the corresponding degrees of freedom are activated.

[0016] Step 3:

[0017] Constructing a hierarchical coupled controller:

[0018] 1. Upper-level controller based on model predictive control

[0019] 2. Lower-level allocator based on quadratic programming

[0020] The driving torque is dynamically distributed to each wheel hub motor by a lower-level distributor based on quadratic programming, and the output is the optimal torque sequence adapted to the module configuration and energy state, ensuring vehicle driving performance and system safety.

[0021] The beneficial effects of this invention are:

[0022] This invention proposes a cooperative control method for reconfigurable vehicles based on a dual predictive controller of state and model. By employing a scalable observer to perceive the module connection state in real time and dynamically adjusting the dimension of the observation matrix when modules are added or removed, it solves the problem that traditional fixed-dimensional observers cannot adapt to changes in the dimension of the state vector. Based on real-time topology state-driven distributed dynamics modeling, it automatically generates a degree-of-freedom configuration and mechanical constraints of the connection mechanism that match the number of modules, overcoming the bottleneck of fixed models being unable to be updated online to adapt to topology changes. A hierarchical controller is constructed, with the upper-level model predictive control embedding configuration consistency optimization objectives and hard mechanical constraints of the connection mechanism, and the lower-level quadratic programming dynamically reconstructing the torque distribution boundary based on the module connection state, resolving the problem of conflicting multiple control objectives. Ultimately, a closed loop of "state perception - model reconstruction - control optimization" is formed, overcoming the bottlenecks of response delay, model mismatch, and control failure when modules are added or removed in traditional schemes. It achieves cooperative control with small state errors and dynamic balance of multiple objectives, providing an effective method for adaptive cooperative control of reconfigurable vehicles. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0024] A reconfigurable vehicle cooperative control method with dual predictive controllers, the method comprising:

[0025] Step 1:

[0026] 1. Construct a four-dimensional state vector containing dynamic state, driving state, energy state, and reconfiguration state:

[0027] ;

[0028] in:

[0029] Dynamic state for:

[0030] ;

[0031] For longitudinal velocity, For lateral velocity, The yaw rate is angular velocity. For longitudinal acceleration, It is lateral acceleration;

[0032] Drive state for:

[0033] ;

[0034] For the rotational speed of each wheel hub motor, For torque;

[0035] Energy Status for:

[0036] ;

[0037] Battery state of charge, For voltage, For current, For range extender power;

[0038] Reconstruction state for:

[0039] ;

[0040] This indicates the module is connected. For type.

[0041] 2. Simultaneously construct a nonlinear state-space model:

[0042] ;

[0043] in For state vectors, For the input vector, It is a nonlinear state transition function. This is the perturbation vector.

[0044] 3. Nonlinear state estimation is performed using extended Kalman filtering, including:

[0045] Prediction phase:

[0046] ;

[0047] in This is process noise;

[0048] Jacobian matrix calculation:

[0049] , ;

[0050] By performing a first-order Taylor expansion of the nonlinear function using the Jacobian matrix, the nonlinear state transition function and measurement function are locally linearized and transformed into an approximate linear model, enabling EKF to perform state estimation using linear filtering theory.

[0051] Measurement update phase:

[0052] ;

[0053] For measuring noise;

[0054] Kalman gain and state update:

[0055] .

[0056] 4. And dynamically adjust the state vector and system matrix when modules are added or removed. and Dimensions:

[0057] Module access processing: Lock signal detected Initialize energy state, expand state vector and system matrix , The dimension;

[0058] Module removal process: Lock signal detected Remove the corresponding module from the state vector to shrink the system matrix. , Dimensions.

[0059] Step 2:

[0060] 1. Decompose the reconfigurable vehicle into independent mobility modules, connection modules, and mission compartments, and define the physical parameters of each module;

[0061] 2. Establish the dynamic equations for single modules and the coupled dynamic models between modules.

[0062] The specific steps for establishing the single-module dynamic equations are as follows:

[0063] For the Each module establishes the equations of translation and rotation dynamics:

[0064] Translation equation: ;

[0065] Equation of rotation: ;

[0066] The specific steps for establishing the inter-module coupling dynamics model are as follows:

[0067] Establish a force transmission model between modules:

[0068] Axial traction force transmission: ;

[0069] Lateral shear force constraint: .

[0070] 3. Determine the system degrees of freedom based on the number of modules, including:

[0071] Overall degrees of freedom: 3 translations (X / Y / Z) + 3 rotations (yaw / pitch / roll), for a total of 6 global degrees of freedom;

[0072] Module degrees of freedom: Each motor module contains 2 rotational degrees of freedom for the wheels (around the axle).

[0073] Connection degrees of freedom: The connection between modules restricts translation and pitch / tilt rotation in the X / Y / Z directions, and only retains a small elastic deformation in the yaw direction (1 degree of freedom);

[0074] Total degrees of freedom: 7 + 2n (n: number of modules).

[0075] 4. Input the number of modules n and the wheelbase through the parameterized interface. The connection type (symmetric / asymmetric) is selected, and the automatically generated topology is output. Constraints are assigned between modules, and the corresponding degrees of freedom are activated.

[0076] Step 3:

[0077] Constructing a hierarchical coupled controller:

[0078] 1. Upper-level controller based on model predictive control

[0079] Design the objective function: ;

[0080] in: For trajectory tracking error term, For driving stability, For energy optimization, For configurational consistency, 、 、 、 These are weighting coefficients, which are based on the reconstructed state. With energy status Real-time adjustment, achieving adaptive adjustment under all operating conditions through a dynamic weight matrix;

[0081] Design constraints:

[0082] The equality constraint is:

[0083] ;

[0084] State vector It includes dynamic state, driving state, energy state, and reconfiguration state; input vector , representing the total longitudinal force and the additional yaw moment; a constant vector.

[0085] The inequality constraints are:

[0086] ;

[0087] Inequality constraints include hardware physical constraints, inter-module force transmission constraints, configuration dynamic constraints, and energy and drive constraints. Hardware physical constraints limit the safe range of total longitudinal force, additional yaw moment, and vehicle motion states (vehicle speed, yaw rate, etc.), as well as in-wheel motor torque and battery power limits. Inter-module force transmission constraints ensure connection stability through axial traction force and lateral shear force constraints. Configuration dynamic constraints specify the dynamic matching of force loading rate and degrees of freedom when modules are added or removed. Energy and drive constraints limit battery output power and limit torque to the high-efficiency range based on the in-wheel motor efficiency MAP. These constraints are dynamically adjusted in real time according to the module connection status, road adhesion coefficient, etc., to ensure system safety and stability.

[0088] The above constraints are combined into a unified form:

[0089] ;

[0090] Upper MPC output total longitudinal force Additional yaw moment and torque distribution range It provides optimization objectives and constraint boundaries for the lower-level QP algorithm.

[0091] 2. Lower-level allocator based on quadratic programming

[0092] Design the objective function:

[0093] ;

[0094] in: This is used to balance the torque of each wheel hub motor to avoid overloading a single wheel hub motor. Used to optimize energy utilization based on hub motor efficiency MAP. Ensure torque does not exceed hardware limits. Dynamic weighting coefficient. Based on the module connection status and battery SOC, adjustments are made in real time to achieve synergistic optimization of drive performance and safety.

[0095] Design constraints:

[0096] The driving force balance constraint is:

[0097] ;

[0098] Ensure a balance between total driving force and resistance, and adapt to dynamic torque distribution when modules are added or removed;

[0099] The yaw moment balance constraint is:

[0100] ;

[0101] Based on the torque distribution of hub motor mounting coordinates, the asymmetric configuration introduces wheelbase weighting to correct deviations.

[0102] The torque limit constraint for the hub motor is:

[0103] ;

[0104] Safety is ensured by combining dynamic temperature derating of the in-wheel motor;

[0105] The torque difference constraint between modules is:

[0106] ;

[0107] To prevent overload of the connecting mechanism, the weight is increased during configuration switching;

[0108] Battery power constraints are:

[0109] .

[0110] Dynamically reduce the SOC and prioritize the high-efficiency hub motor.

[0111] Each constraint is dynamically adjusted through mechanisms such as weight correction when the module connection state changes and power derating driven by energy state, to achieve adaptability to different configurations and operating conditions of reconfigurable vehicles, ensuring the feasibility of torque distribution and system safety.

[0112] The driving torque is dynamically distributed to each wheel hub motor by a lower-level distributor based on quadratic programming, and the output is the optimal torque sequence adapted to the module configuration and energy state, ensuring vehicle driving performance and system safety.

[0113] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0114] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

[0115] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A reconfigurable vehicle cooperative control method with dual predictive controllers, characterized in that, The method includes: A four-dimensional state vector is constructed, comprising dynamic state, driving state, energy state, and reconfiguration state. A nonlinear state-space model is also constructed, and an extended Kalman filter is used for nonlinear state estimation. The state vector and system matrix are dynamically adjusted when modules are added or removed. and The dimension; The reconfiguration state refers to the state constituted by the connection state and type information of each module of the reconfigurable vehicle. The reconfigurable vehicle is decomposed into independent mobility modules, connection modules, and mission compartments. Single-module dynamic equations and inter-module coupled dynamic models are established. The system degrees of freedom are determined based on the number of modules. Through a parameterized interface, the number of modules, wheelbase, and connection type are input, and the automatically generated topology, inter-module constraints, and corresponding activated degrees of freedom are output. A hierarchical coupled controller is constructed: the upper layer adopts the objective function of model predictive control and sets equality constraints and inequality constraints to output the total longitudinal force and additional yaw moment; the lower layer adopts the objective function of quadratic programming and sets driving force balance constraints, yaw moment balance constraints, hub motor torque limiting constraints, inter-module torque difference constraints and battery power constraints, and dynamically distributes the driving torque used in the driving force balance constraints to each hub motor. in: The state vector and system matrix are dynamically adjusted when modules are added or removed. and The dimensions are specifically: Module access processing: Lock signal detected Initialize energy state, expand state vector and system matrix , The dimension; Module removal process: Lock signal detected Remove the corresponding module from the state vector to shrink the system matrix. , The dimension; The upper layer uses model predictive control to design the objective function: ； in: For trajectory tracking error term, For driving stability, For energy optimization, For configurational consistency, , , , These are the weighting coefficients; The lower layer uses a quadratic programming design objective function: ； in: This refers to the torque balancing term for the hub motor; This is an efficiency optimization item for in-wheel motors; This is a hardware security protection item; It is a dynamic weighting coefficient that is adjusted according to the module connection status and battery SOC.

2. The method according to claim 1, characterized in that, The constructed state vector comprises a four-dimensional state, including dynamic state, driving state, energy state, and reconfiguration state. ； in: Dynamic state for: ； For longitudinal velocity, For lateral velocity, The yaw rate is angular velocity. For longitudinal acceleration, It is lateral acceleration; Drive state for: ； For the rotational speed of each wheel hub motor, For torque; Energy Status for: ； Battery state of charge, For voltage, For current, For range extender power; Reconstruction state for: ； This indicates the module is connected. For type.

3. The method according to claim 1, characterized in that, The construction of the nonlinear state-space model: ； in For state vectors, For the input vector, It is a nonlinear state transition function. This is the perturbation vector.

4. The method according to claim 1, characterized in that, The nonlinear state estimation using extended Kalman filtering includes: Prediction phase: ； in For process noise, The control input vector represents the system's state at time step [0, 1]. Drive control commands; Jacobian matrix calculation: ， ； In the formula, Let be the Jacobian matrix of the state transition function; Let be the Jacobian matrix of the observation function; Measurement update phase: ； In the formula, For observation vectors; For observation functions; It is a state vector; To observe noise; Kalman gain and state update: ； In the formula, Kalman gain is used to balance the confidence levels of predicted and measured values. Let be the state covariance matrix, representing the uncertainty of state prediction; Let Jacobian matrix be the observation function, which maps the state space to the observation space; The noise covariance matrix is ​​measured to represent the sensor error; For time steps The predicted state value; For observation vectors; For observation functions; For time steps Status update value; This is the updated state covariance matrix; It is an identity matrix.

5. The method according to claim 1, characterized in that, The establishment of single-module dynamic equations and inter-module coupled dynamic models is described above, wherein: The specific steps for establishing the single-module dynamic equations are as follows: For the Each module establishes the equations of translation and rotation dynamics: Translation equation: ; In the formula, For the first The quality of each module; For the first The acceleration vector of each module; For module For modules The force, and For module indexing, , Total number of modules; For the first The driving force of each module; For the first The rolling resistance of each module; Equation of rotation: ; In the formula, For the first The moment of inertia tensor of each module; For the first The angular velocity vector of each module; For the first The angular acceleration vector of each module; For module For modules The torque; For the first Aerodynamic drag torque of each module; The specific steps for establishing the inter-module coupling dynamics model are as follows: Establish a force transmission model between modules: Axial traction force transmission: ; In the formula, For module and Axial traction force between them; This is the axial stiffness coefficient of the connecting mechanism; For module and axial displacement difference; The axial damping coefficient of the connecting mechanism; This is the derivative of the axial displacement difference; Preload for the connecting mechanism; Lateral shear force constraint: ; In the formula, This refers to the lateral shear force between modules. This refers to the vertical shear force between modules; denoted as the coefficient of friction of the connecting surfaces.

6. The method according to claim 1, characterized in that, The system degrees of freedom are determined based on the number of modules, where the total degrees of freedom is 7 + 2n, and n is the number of modules. Specifically, this includes: Six global degrees of freedom (3 translations and 3 rotations), two rotational degrees of freedom for each motor module's wheels, and one connection degree of freedom for elastic deformation in the yaw direction between modules.

7. The method according to claim 1, characterized in that, The equality constraint is: ； In the formula, The coefficient matrix of the state vector; for The state vector at time t, The coefficient matrix of the input vector; for The input vector at time step; It is a constant vector; The inequality constraint is: ； In the formula, Let be the inequality coefficient matrix of the state vector; The inequality coefficient matrix of the input vector; This is the time-varying constraint boundary vector.

8. The method according to claim 1, characterized in that, The driving force balance constraint is: ； In the formula, For the first The output torque of each hub motor; The effective radius of the wheel; This represents the total longitudinal force output by the upper-level MPC. For vehicle driving resistance; The yaw moment balance constraint is: ； In the formula, This indicates the coordinates of the hub motor in the vehicle coordinate system; This is the additional yaw moment output by the upper-level MPC; The torque limit constraint for the hub motor is: ； In the formula, This is the lower limit of the torque of the hub motor. This is the upper limit of the hub motor torque; The torque difference constraint between modules is: ； In the formula, The coefficient of friction of the connecting mechanism; Preload the connecting mechanism; Battery power constraints are: ； In the formula, For the first The rotational speed of each hub motor; This is the maximum allowable output power of the battery.

Images