Navigation path planning trolley attitude control method based on MPC algorithm, electronic equipment and storage medium
By building a linear car state space model in an embedded system and solving the MPC optimization problem using the fast gradient method, the problem of computational complexity and power consumption limitation in an embedded system is solved, and efficient processing and real-time control of the highly nonlinear Ackerman car model is achieved.
Patent Information
- Application Number
- CN202411583668.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-07
- Publication Date
- 2025-02-28
AI Technical Summary
In embedded systems, the high computational complexity and power consumption limitations of MPC algorithms are difficult to apply directly, especially when dealing with highly nonlinear Ackerman cart motion models.
By building a linearized car Akerman car state space model and using the fast gradient method (FGM) to solve MPC optimization problems, reduce the computational complexity and memory usage, and achieve real-time control.
On a hardware platform with limited computing resources, efficient processing of highly nonlinear Ackerman car model under low resources is achieved to meet real-time control requirements.
Smart Images
Figure CN119536257A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of navigation path planning, and in particular to a navigation path planning vehicle posture control method, electronic equipment and storage medium based on an MPC algorithm. Background Art
[0002] The hardware-friendly MPC (Model Predictive Control) implementation aims to enable the MPC algorithm to run efficiently in embedded systems or hardware platforms with limited computing resources (such as smart cars, mobile robots, etc.). The goal of this implementation is to reduce the computational complexity and power consumption of the algorithm and ensure that the system can run in scenarios with high real-time requirements, especially in low-power processors or embedded devices.
[0003] MPC is a powerful control algorithm that generates a series of future control instructions by solving an optimization problem that includes a prediction model, control objectives, and constraints online. Although MPC performs well in complex control scenarios, its high computational complexity makes it difficult to apply directly in embedded systems, especially when the nonlinear motion model of the Ackerman car is added. The design of hardware-friendly MPC needs to address the following challenges: (1) Real-time requirements: It is necessary to ensure that MPC can quickly complete the optimization calculation in each control cycle and update the control input in real time. (2) Computational resource limitations: Embedded systems usually have limited computing power, requiring the algorithm to have low computational complexity. (3) Power consumption limitations: Embedded hardware is often used in low-power scenarios, requiring the algorithm to execute efficiently and avoid excessive power consumption caused by frequent calculations. (4) Nonlinear model processing: The Ackerman car model is highly nonlinear. How to efficiently process this model under low resource conditions is one of the core difficulties in implementation. Summary of the invention
[0004] The purpose of the present invention is to overcome the deficiencies of the prior art and provide a navigation path planning vehicle posture control method, electronic equipment and storage medium based on the MPC algorithm.
[0005] The objective of the present invention is achieved through the following technical solutions:
[0006] In a first aspect of the present invention, a method for controlling a navigation path planning vehicle posture based on an MPC algorithm is provided, wherein a controller for controlling the posture of a vehicle conforming to an Ackerman vehicle motion model is provided, and for each control cycle, the method comprises the following steps:
[0007] Build the Ackerman car state space model of the car after linearization. The output value of the space model is the next state, and the input value of the space model includes the current state and the current control input.
[0008] Get the current state of the car through sensor measurement or state observer;
[0009] Construct an MPC optimization problem, that is, predict the control input for the next N steps based on the current state, and set constraints on the current state and / or the current control input;
[0010] Using FGM to solve the optimization problem and obtain the optimal control input;
[0011] Output the control signal of the preferred control input.
[0012] Furthermore, the construction of the Ackerman car state space model of the car after linearization processing includes:
[0013] Build a nonlinear Ackerman car state space model;
[0014] The nonlinear Ackerman cart state space model is linearized around the current state through Taylor expansion to obtain a linear approximate model, namely the Ackerman cart state space model of the linearized cart.
[0015] Furthermore, constructing the MPC optimization problem includes:
[0016] establishing an objective function that aims to minimize state error and cost of control input;
[0017] The constraint conditions of the objective function are set, wherein the constraint conditions include the constraint of the current state and / or the current control input.
[0018] Further, the using FGM to solve the optimization problem includes:
[0019] Convert the optimization problem into a standard quadratic programming problem;
[0020] The standard quadratic programming problem is solved using FGM.
[0021] Furthermore, the use of FGM to solve the standard quadratic programming problem includes
[0022] Initialize control input and set step size;
[0023] Based on the step size and the standard quadratic programming problem, the control input is iteratively updated using gradient descent, and if there are constraints on the control input, the next control state is projected onto the constraint set using a projection operation;
[0024] The iteration stops when the preset termination condition is met.
[0025] Furthermore, the initialized control input is set to a zero vector or a solution obtained from the last optimization; the preset termination conditions include the gradient update amplitude being less than a certain threshold, or reaching a maximum number of iterations.
[0026] Further, the outputting of the control signal of the preferred control input includes:
[0027] The control output of the first time step of the control signal of the solved preferred control input generates a corresponding control signal.
[0028] Furthermore, the control input is a full-state control input or a key control input; the key control input is a key parameter control input or an application scenario control input.
[0029] According to a second aspect of the present invention, there is provided an electronic device, comprising a storage unit and a processing unit, wherein the storage unit stores computer instructions executable on the processing unit, and the processing unit executes the steps of the method described in the first aspect when executing the computer instructions.
[0030] According to a third aspect of the present invention, there is provided a storage medium having computer instructions stored thereon, wherein the computer instructions, when executed, execute the steps of the method described in the first aspect.
[0031] The beneficial effects of the present invention are:
[0032] In an exemplary embodiment of the present invention, by introducing the fast gradient method, the hardware-friendly MPC can achieve real-time control in an embedded system with lower computational cost and smaller memory usage, and adapt to hardware platforms with limited computing resources, thereby achieving efficient processing of highly nonlinear Ackerman car models under low resource conditions. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1 The present invention is a flowchart of a method for controlling a navigation path planning vehicle posture based on an MPC algorithm provided in an exemplary embodiment of the present invention. DETAILED DESCRIPTION
[0034] The technical solution of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0035] In the description of the present invention, it should be noted that the directions or positional relationships indicated by "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc. are directions or positional relationships based on the drawings, which are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.
[0036] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, "installation", "connection" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0037] In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.
[0038] See also Figure 1 , Figure 1 A navigation path planning vehicle attitude control method based on an MPC algorithm provided in an exemplary embodiment of the present invention is shown, and a controller for attitude control of a vehicle conforming to the Ackerman vehicle motion model is shown, and for each control cycle, the following steps are included:
[0039] Build the Ackerman car state space model of the car after linearization. The output value of the space model is the next state, and the input value of the space model includes the current state and the current control input.
[0040] Get the current state of the car through sensor measurement or state observer;
[0041] Construct an MPC optimization problem, that is, predict the control input for the next N steps based on the current state, and set constraints on the current state and / or the current control input;
[0042] Using FGM to solve the optimization problem and obtain the optimal control input;
[0043] Output the control signal of the preferred control input.
[0044] Specifically, in this exemplary embodiment, the Ackerman car state space model of the car with linearization processing is first constructed: the motion model of the Ackerman car is nonlinear, but the nonlinear model can be approximated as a linear model by Taylor expansion or other linearization methods in a short time. Through linearization, the constructed MPC optimization problem can be simplified to a quadratic programming (QP) problem, which greatly reduces the computational complexity. Although linearization will bring a certain loss of accuracy, this error is acceptable within a reasonable control cycle.
[0045] In addition, traditional MPC algorithms often rely on quadratic programming (QP) solvers, but this has low execution efficiency on hardware with limited resources. Hardware-friendly MPC can use the Fast Gradient Method to solve the quadratic programming problem. The computational complexity of this method is O(1) and it has a small memory footprint, which is very suitable for embedded systems. The advantages of the fast gradient method in embedded systems are: (1) Low computational complexity: The computational complexity of FGM is O(1), which means that the amount of calculation for each iteration is constant and does not depend on the size of the problem, which is suitable for running in resource-constrained systems. (2) Small memory footprint: FGM only needs to store the control input vector and the gradient, so the memory footprint is extremely small, which is very suitable for embedded devices. (3) Fast convergence: Since FGM is a gradient-based optimization algorithm, it has a fast convergence speed and usually only requires a small number of iterations to obtain an approximate optimal solution. Therefore, in this exemplary embodiment, based on the hardware-friendly MPC, FGM replaces the traditional quadratic programming solver to significantly improve the computational efficiency.
[0046] The specific application steps are as follows: Linearization of state space model: First, linearize the nonlinear model of the Ackerman car and convert it into a quadratic programming problem. Solve the control input: Use FGM to iteratively solve the control input and quickly approach the optimal solution. Real-time update: Due to the fast convergence speed of FGM, the control input can be updated in real time within each control cycle, meeting the real-time requirements of MPC.
[0047] By introducing the fast gradient method, the hardware-friendly MPC can achieve real-time control in embedded systems with low computational cost and small memory usage, and adapt to hardware platforms with limited computing resources, thereby achieving efficient processing of highly nonlinear Ackerman car models under low resource conditions.
[0048] The following content will describe the preferred exemplary embodiments:
[0049] More preferably, in an exemplary embodiment, the construction of the Ackerman car state space model of the car after linearization processing includes:
[0050] Build a nonlinear Ackerman car state space model;
[0051] The nonlinear Ackerman cart state space model is linearized around the current state through Taylor expansion to obtain a linear approximate model, namely the Ackerman cart state space model of the linearized cart.
[0052] Specifically, in this exemplary embodiment, the construction of the Ackerman car state space model of the car subjected to linearization processing first constructs a nonlinear Ackerman car state space model, and then performs linearization processing. More specifically:
[0053] The equation of the nonlinear Ackerman cart state space model is:
[0054] x k+1 =f(x k ,u k )
[0055] where x k is the current state, u k is the current control input, x k+1 is the next state. Through Taylor expansion, it is in the current state x k Linearize near to get the linear approximate model
[0056] x k+1 ≈A k x k +B k u k +C k
[0057] in: is the partial derivative matrix (Jacobian matrix) of the state with respect to the state variables, is the partial derivative matrix of the state with respect to the control input, C k is a constant term during linearization.
[0058] The linearized model can be simplified into a quadratic programming (QP) problem, which is easy to solve efficiently on resource-limited embedded systems.
[0059] Reduce the prediction step size: Adjusting the prediction time domain length of MPC and reducing the number of prediction steps can reduce the amount of calculation. Conventional MPC usually uses a longer prediction step size, but in embedded systems, reasonably shortening the prediction step size can reduce the computational burden without significantly reducing performance.
[0060] More preferably, in an exemplary embodiment, the current state x of the vehicle obtained by sensor measurement or state observer is obtained. k , providing optimized initial conditions for MPC.
[0061] More preferably, in an exemplary embodiment, constructing the MPC optimization problem includes:
[0062] establishing an objective function that aims to minimize state error and cost of control input;
[0063] The constraint conditions of the objective function are set, wherein the constraint conditions include the constraint of the current state and / or the current control input.
[0064] Specifically, in this exemplary embodiment, an MPC optimization problem is constructed with the goal of minimizing the cost of state deviation and control input within the next N steps (prediction time domain) while satisfying physical or safety constraints.
[0065] For the objective function: The objective function of MPC is quadratic, aiming to minimize the cost of state error and control input:
[0066]
[0067] Where: x ref is the reference state, the state that the system expects to reach; Q and R are the weight matrices of state error and control input respectively; N is the length of the prediction time domain; J represents the gradient.
[0068] For constraints: Constrain the state and control input of the system, for example:
[0069] Control input limit: u min ≤u k ≤u max .
[0070] More preferably, in an exemplary embodiment, the using FGM to solve the optimization problem includes:
[0071] Convert the optimization problem into a standard quadratic programming problem;
[0072] The standard quadratic programming problem is solved using FGM.
[0073] Specifically, in this exemplary embodiment, when it is necessary to use FGM to solve the optimization problem, it is necessary to first perform a quadratic programming transformation. That is, the objective function is constructed into a standard quadratic programming problem form:
[0074]
[0075] u is the control input vector to be optimized. H is the symmetric positive definite matrix in the quadratic optimization problem. g is the gradient vector of the linear term. This can be achieved through the corresponding transformation.
[0076] More preferably, in an exemplary embodiment, the use of FGM to solve the standard quadratic programming problem includes:
[0077] Initialize control input and set step size;
[0078] Based on the step size and the standard quadratic programming problem, the control input is iteratively updated using gradient descent, and if there are constraints on the control input, the next control state is projected onto the constraint set using a projection operation;
[0079] The iteration stops when the preset termination condition is met.
[0080] Specifically, in this exemplary embodiment, a specific implementation of FGM solving a standard quadratic programming problem is disclosed:
[0081] (1) Initialization: The initial control input u0 can be set to the zero vector or the solution obtained from the last optimization, and the step size can be set. Where L is the largest eigenvalue of the matrix H (Lipschitz constant), which is related to the largest eigenvalue of the matrix H and is used to determine the optimal step size.
[0082] (2) Iteratively update the control input: Use gradient descent to update the control input u k ,
[0083]
[0084] The gradient is calculated as:
[0085]
[0086] This iterative process converges quickly, usually approaching the optimal solution within a few iterations.
[0087] (3) Constraint processing: If there are constraints on the control input, use the projection operation to convert u k+1 Projection onto the constraint set:
[0088]
[0089] If the control input u is subject to constraints, such as vehicle speed and steering angle range limits, a projection operation can be applied after each iteration to convert the calculated u k+1 Projected into the space of allowed constraints. Where C is the set of constraints on the control input, such as maximum speed or maximum steering angle.
[0090] (4) Termination condition: When the iteration meets the preset termination condition (such as the gradient update amplitude is less than a certain threshold, or the maximum number of iterations is reached), it stops:
[0091]
[0092] Since the iterative steps of FGM converge quickly, a nearly optimal control input can usually be obtained after a small number of iterations.
[0093] More preferably, in an exemplary embodiment, the outputting the control signal of the preferred control input includes:
[0094] The control output of the first time step of the control signal of the solved preferred control input generates a corresponding control signal.
[0095] Specifically, in this exemplary embodiment, the optimal control input obtained by FGM is Only the control input of the first time step is used To update the system state. That is, this exemplary embodiment adopts a rolling horizon optimization: MPC reconstructs the optimization problem in each control cycle through a rolling horizon optimization strategy and re-solves it using FGM. That is, in each control cycle, MPC obtains a new system state and performs prediction and optimization again; only the first control input obtained by the current optimization is applied, and the input of the remaining prediction period will be updated in the next control cycle.
[0096] It should be noted that the control input of the first time step is used to update the system state instead of directly applying all the control inputs of the entire optimization prediction period, mainly because of the following reasons:
[0097] 1. Receding Horizon Control
[0098] MPC uses a rolling horizon optimization strategy, that is, in each control cycle, MPC will reconstruct the optimization problem and solve it. This is because the system state may deviate from the predicted trajectory in actual execution due to model errors, external interference, and unmodeled dynamic behavior. By re-obtaining the actual state of the system and optimizing it in each control cycle, the control input can be adjusted in real time, making the system more flexible and robust.
[0099] 2. Feedback Correction (Closed-Loop Feedback)
[0100] MPC is essentially a closed-loop control strategy that relies on real-time feedback. In each control cycle, MPC recalculates the control input based on the current measured or estimated state. Using only the control input of the first time step ensures that the system can respond to state changes and uncertainties in a timely manner, thereby improving the robustness and performance of the control.
[0101] 3. Reduce the impact of model errors
[0102] The system model is usually an approximate model and may not fully describe the dynamic behavior of the actual system. Therefore, the prediction of the next N steps may not be completely accurate. By applying the control input of the first time step in each control cycle and re-optimizing in the next cycle, the prediction error can be gradually corrected and the impact of model inaccuracy can be reduced.
[0103] 4. Dealing with external disruptions and a changing environment
[0104] External disturbances (such as environmental changes or unmodeled dynamic factors) may affect the system state. If the control inputs for the entire forecast period are directly applied, the system will lose its ability to adapt to external changes. Each control cycle is re-optimized based on the latest state information, which can better cope with these external disturbances and uncertainties.
[0105] 5. Real-time requirements
[0106] MPC is often used in real-time control systems, such as autonomous vehicles or industrial process control. Only the control input of the first time step is applied in each control cycle, so that the controller can complete the optimization and update the system state in a limited time to meet the real-time requirements.
[0107] Therefore, in summary, by applying only the control input of the first time step, MPC can dynamically adapt to the state changes and environmental uncertainties of the actual system while continuously updating predictions and optimizations. This strategy improves the robustness, flexibility, and real-time responsiveness of the system, thereby achieving more stable and efficient control effects.
[0108] More preferably, in an exemplary embodiment, the control input is a full-state control input or a key control input; the key control input is a key parameter control input or an application scenario control input.
[0109] Specifically, in this exemplary embodiment, a dimension reduction processing step is also included: by analyzing the control requirements of the Ackerman car, the dimension of the optimization problem can be reduced.
[0110] For example, the full-state optimization problem can be transformed into an optimization problem that only considers key control inputs (such as steering angle and vehicle speed), reducing the dimension of the optimization variables and thus reducing the amount of computation. Dimensionality reduction can further simplify the control problem according to the specific application scenarios of the vehicle (such as low-speed driving, turning in a specific environment).
[0111] The complete nonlinear model of the Ackerman car describes the motion of the vehicle through the following state equation:
[0112]
[0113] can be simplified to the following lower dimensional model, focusing on the key variables:
[0114]
[0115] Where x and y are the positions, θ is the heading angle, v is the speed, δ is the steering angle, and L is the wheelbase.
[0116] Having the same inventive concept as the above exemplary embodiment, another exemplary embodiment of the present invention provides an electronic device, including a storage unit and a processing unit, wherein the storage unit stores computer instructions that can be executed on the processing unit, and the processing unit executes the steps of the method when running the computer instructions.
[0117] The electronic device is presented in the form of a general-purpose computing device. The components of the electronic device may include, but are not limited to: the at least one processing unit, the at least one storage unit, and a bus connecting different system components (including the storage unit and the processing unit).
[0118] The storage unit stores program codes, which can be executed by the processing unit, so that the processing unit performs the steps according to various exemplary embodiments of the present invention described in the above “Exemplary Method” section of this specification. For example, the processing unit can perform the following steps: Figure 1 The method shown in .
[0119] The storage unit may include a readable medium in the form of a volatile memory unit, such as a random access memory unit (RAM) and / or a cache memory unit, and may further include a read-only memory unit (ROM).
[0120] The storage unit may also include a program / utility having a set (at least one) of program modules, such program modules including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.
[0121] The bus may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.
[0122] The electronic device may also communicate with one or more external devices (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may communicate with one or more devices that enable a user to interact with the electronic device, and / or may communicate with any device that enables the electronic device to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed through an input / output (I / O) interface. Furthermore, the electronic device may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) through a network adapter. The network adapter communicates with other modules of the electronic device through a bus. It should be understood that other hardware and / or software modules may be used in conjunction with the electronic device, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.
[0123] Through the above description, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to this exemplary embodiment can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to this exemplary embodiment.
[0124] With the same inventive concept as the above exemplary embodiment, another exemplary embodiment of the present invention provides a storage medium having computer instructions stored thereon, and the steps of the method are executed when the computer instructions are executed.
[0125] Based on this understanding, the technical solution of this embodiment can essentially or partly contribute to the prior art or part of the technical solution can be embodied in the form of a software product (program product). The computer software product is stored in a storage medium and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention.
[0126] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.
[0127] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, in which readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.
[0128] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.
[0129] Program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages such as Java, C++, etc., and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).
[0130] Obviously, the above embodiments are merely examples for clear explanation, and are not intended to limit the implementation methods. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived from them are still within the protection scope of the invention.
Claims
1. A navigation path planning car attitude control method based on the MPC algorithm, a controller for attitude control of a car that conforms to the Ackerman car motion model, characterized by: For each control cycle, the following steps are included: Build the Ackerman car state space model of the car after linearization. The output value of the space model is the next state, and the input value of the space model includes the current state and the current control input. Get the current state of the car through sensor measurement or state observer; Construct an MPC optimization problem, that is, predict the control input for the next N steps based on the current state, and set constraints on the current state and / or the current control input; Using FGM to solve the optimization problem and obtain the optimal control input; Output the control signal of the preferred control input.
2. The navigation path planning vehicle posture control method based on the MPC algorithm according to claim 1 is characterized in that: The Ackerman car state space model of the car after linearization is constructed, including: Build a nonlinear Ackerman car state space model; The nonlinear Ackerman cart state space model is linearized around the current state through Taylor expansion to obtain a linear approximate model, namely the Ackerman cart state space model of the linearized cart.
3. The navigation path planning vehicle posture control method based on the MPC algorithm according to claim 1 is characterized in that: The construction of the MPC optimization problem includes: establishing an objective function that aims to minimize state error and cost of control input; The constraint conditions of the objective function are set, wherein the constraint conditions include the constraint of the current state and / or the current control input.
4. The navigation path planning vehicle posture control method based on the MPC algorithm according to claim 1 is characterized in that: The method of using FGM to solve the optimization problem includes: Convert the optimization problem into a standard quadratic programming problem; The standard quadratic programming problem is solved using FGM.
5. The navigation path planning vehicle posture control method based on the MPC algorithm according to claim 4 is characterized in that: The use of FGM to solve the standard quadratic programming problem includes Initialize control input and set step size; Based on the step size and the standard quadratic programming problem, the control input is iteratively updated using gradient descent, and if there are constraints on the control input, the next control state is projected onto the constraint set using a projection operation; The iteration stops when the preset termination condition is met.
6. The navigation path planning vehicle posture control method based on the MPC algorithm according to claim 4 is characterized in that: The initialized control input is set to the zero vector or the solution obtained from the last optimization; the preset termination conditions include the gradient update amplitude being less than a certain threshold or reaching the maximum number of iterations.
7. The navigation path planning vehicle posture control method based on the MPC algorithm according to claim 1 is characterized in that: The outputting the control signal of the preferred control input comprises: The control output of the first time step of the control signal of the solved preferred control input generates a corresponding control signal.
8. The navigation path planning vehicle posture control method based on the MPC algorithm according to claim 1 is characterized in that: The control input is a full-state control input or a key control input; the key control input is a key parameter control input or an application scenario control input.
9. An electronic device, comprising a storage unit and a processing unit, wherein the storage unit stores computer instructions that can be executed on the processing unit, wherein: When the processing unit runs the computer instructions, the steps of the method according to any one of claims 1 to 8 are performed.
10. A storage medium having computer instructions stored thereon, characterized in that: When the computer instructions are executed, the steps of the method according to any one of claims 1 to 8 are executed.