An intelligent control method and system for unmanned forklifts

Abstract

This invention discloses an intelligent control method and system for unmanned forklifts, relating to the field of unmanned forklift control technology. The method includes: constructing a perception layer by deploying multi-dimensional sensors to acquire data from the unmanned forklift; correcting the position based on the data using a preset position optimization algorithm to generate precise positioning information; establishing a dynamic model based on the data; defining an optimization problem based on the precise positioning information, the dynamic model, and path planning results; solving the optimal control sequence using a nonlinear optimization algorithm; extracting the control input at the current moment; and applying it to the drive and steering systems for intelligent control. By employing a preset position optimization algorithm to achieve high-precision positioning, the system enhances its adaptability to environmental changes and sensor noise. Furthermore, based on precise positioning information and the dynamic model, it performs efficient path planning and tracking control, ensuring accurate navigation of the unmanned forklift in dynamic and complex warehousing environments, thereby improving the efficiency and reliability of logistics operations.

Description

Technical Field

[0001] This invention relates to the field of unmanned forklift control technology, specifically to an intelligent control method and system for unmanned forklifts. Background Technology

[0002] With the rapid development of the modern logistics industry, automated equipment is increasingly widely used in warehousing and transportation. As an important component of automated logistics equipment, the research and application of intelligent control technology for unmanned forklifts is crucial for improving logistics efficiency and reducing labor costs. However, existing unmanned forklift control methods still face many challenges in terms of precise positioning, path planning, and dynamic control.

[0003] In existing technologies, to improve the positioning accuracy and control stability of unmanned forklifts, researchers have introduced multi-sensor data fusion technology. This technology comprehensively utilizes data from multiple sensors to construct a unified environmental model and accurate positioning information. However, effectively integrating the data characteristics of different sensors and handling data heterogeneity and noise interference during multi-sensor data fusion remain pressing technical challenges that need to be addressed.

[0004] Furthermore, existing control algorithms for unmanned forklifts mostly employ linear control methods, which struggle to handle the nonlinear dynamic characteristics and complex constraints encountered in practical applications. Further research is needed to design efficient optimization algorithms that combine high-precision multi-sensor positioning information with complex nonlinear dynamic models. Summary of the Invention

[0005] Based on the shortcomings of the prior art described above, the purpose of this invention is to provide an intelligent control method and system for unmanned forklifts to solve the aforementioned technical problems.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an intelligent control method for an unmanned forklift, comprising:

[0007] By deploying multi-dimensional sensors to build a perception layer, the status data, location data, and static environmental feature location data of the unmanned forklift are acquired.

[0008] Based on the location data of the unmanned forklift and the location data of static environmental features, the location data of the unmanned forklift is corrected using a preset location optimization algorithm to generate accurate positioning information;

[0009] A dynamic model of the unmanned forklift is established based on its state data. Based on the precise positioning information, combined with the dynamic model and path planning results, an optimization problem is defined, and the optimal control sequence is solved using a nonlinear optimization algorithm.

[0010] The control input at the current moment is extracted based on the optimal control sequence and applied to the drive and steering system of the unmanned forklift for intelligent control.

[0011] The present invention is further configured such that the multi-dimensional sensor includes a lidar, a camera, an ultrasonic sensor, an inertial measurement unit, and an ultra-wideband positioning base station, and the status data includes linear velocity, driving torque, friction torque, angular velocity, and steering angle.

[0012] The present invention is further configured to correct the position data of the unmanned forklift using a preset position optimization algorithm, including:

[0013] Starting from the initial moment, the location data (x1(t), x2(t)) of the unmanned forklift and the location data (l1(t), l2(t)) of the static environment feature are acquired in the time series according to time step t, and the measurement distance s(t) between the unmanned forklift and the static environment feature is calculated.

[0014] The geometric deviation measure of time step t is calculated based on the location data (x1(t), x2(t)) of the unmanned forklift, the location data of static environmental features (l1(t), l2(t)) and the measurement distance s(t);

[0015] Based on the dynamic residual enhancement of time step t, the total cost function is obtained by summing the dynamic residual enhancements of all time steps.

[0016] Precise location information is obtained by minimizing the total cost function to solve for the values ​​of (x1(t), x2(t)).

[0017] The present invention further specifies that the calculation logic for the geometric deviation metric at time step t is: D(t)=((x1(t)-l1(t)) 2 +(x2(t)-l2(t)) 2 ) α -(s(t)) 2α , where D(t) is the geometric deviation measure of time step t, and α is the order of the polynomial transformation that controls the geometric difference;

[0018] The calculation logic for dynamic residual enhancement at time step t is: R(t)=(D(t)) 2 ·exp(δ·sin(β·s(t))), where R(t) is the dynamic residual enhancement at time step t, δ is the frequency parameter of the modulated sine function, and β is the scaling parameter of the control exponential factor;

[0019] The calculation logic for the total cost function is as follows: Where C is the total cost function and M is the number of time steps.

[0020] The present invention is further configured to obtain precise location information by minimizing the total cost function and solving for the values ​​of (x1(t), x2(t)), including:

[0021] Calculate the partial derivatives of the total cost function C with respect to the location variables x1(t) and x2(t). and

[0022] Set initial estimates for location variables. and And the number of iterations k = 0;

[0023] In each iteration k, the geometric deviation metric D is calculated. (k) (t), Dynamic residual enhancement R (k) (t) and partial derivatives and

[0024] According to partial derivatives and Update the position variable. Where η is the learning rate parameter, used to control the update step size;

[0025] Calculate the cost function value for the current iteration step. Determine if the convergence condition is met | C (k+1) -C (k) |<∈, where ∈ is the preset convergence threshold;

[0026] When the change in the cost function between two consecutive iterations is less than a preset threshold ∈, the iteration process terminates and the precise location information is output.

[0027] The present invention is further configured to establish a dynamic model of the unmanned forklift based on its state data, define an optimization problem based on the precise positioning information, the dynamic model, and the path planning results, and solve the optimal control sequence using a nonlinear optimization algorithm, including:

[0028] Establish the linear and angular motion equations of the unmanned forklift;

[0029] Discretize the continuous-time model to obtain the discretized linear and angular motion equations;

[0030] The dynamic model is represented in matrix form, and the state update equation is constructed.

[0031] At each control time, an optimization problem with a finite time step N is defined. The optimal control sequence is obtained by minimizing the cost function in the next N steps using a nonlinear optimization algorithm.

[0032] The present invention is further configured such that the linear motion equation of the unmanned forklift is: Where m is the total mass of the unmanned forklift, v(t) is the linear velocity of the unmanned forklift at time t, and F... d (t) represents the driving torque F acting on the unmanned forklift at time t. f (t) represents the frictional torque acting on the forklift at time t;

[0033] The equation of motion for the unmanned forklift is: Where I is the moment of inertia of the unmanned forklift about its center of mass, ω(t) is the angular velocity of the unmanned forklift at time t, L is the wheelbase of the unmanned forklift, θ(t) is the turning angle at time t, and C... f This is the coefficient of friction between the tire and the ground.

[0034] The discretized linear equation of motion is: After rearranging, we get: Δt is the time step of discretization;

[0035] The discretized angular motion equations are: After rearranging, we get:

[0036] The matrix form of the dynamic model is as follows: Where X(t) is the state vector at time t, and u(t) is the control input vector at time t;

[0037] The state update equation is: X(t+1)=A·X(t)+B·u(t)+N(X(t),u(t)), where A is the state transition matrix. B is the control input matrix. N(X(t),u(t)) is a nonlinear disturbance term.

[0038] The present invention is further configured such that, at each control moment, an optimization problem with a finite time step N is defined, and the optimal control sequence is obtained by minimizing the cost function over the next N steps using a nonlinear optimization algorithm, including:

[0039] Define the objective function and constraints, and set the initial control sequence U0(t);

[0040] At the current state point, a Taylor expansion is performed on the nonlinear dynamic model, retaining the first-order terms to obtain the linearized model;

[0041] Based on the linearized model, a quadratic approximation objective function and constraints are constructed. A sequential quadratic programming algorithm is used to iteratively solve the quadratic approximation problem and update the control increment until the objective function converges. The control input at the current moment is extracted from the optimal control sequence and applied to the dynamic model. Based on the dynamic model and the control input, the state of the forklift is updated.

[0042] The present invention is further configured such that the objective function is: Where J(U(t)) is the objective function of the control sequence U(t), x ref (t+k) is the reference state vector, the path planning result, Q is the state error weight matrix, and R is the control input weight matrix;

[0043] The constraints are: x(t+k+1)=A·x(t+k)+B·u(t+k)+N(x(t+k),u(t+k)),k=0,1,…,N-1, and the physical constraints of the control input: u min ≤u(t+k)≤u max k = 0, 1, ..., N-1, where u min and u max These are the minimum and maximum value vectors for controlling the input, respectively;

[0044] A Taylor expansion of the nonlinear dynamic model is performed, retaining the first-order terms, to obtain the linearized model: x(t+k+1)≈A·x(t+k)+B·u(t+k)+J(t+k)·Δx(t+k), where J is the Jacobian matrix, representing the partial derivatives of the nonlinear terms with respect to the state;

[0045] Based on the linear model, construct the objective function and constraints for a quadratic approximation: Among them, J approx (ΔU(t)) is the objective function of the second approximation, ΔU(t) is the control input increment vector, and H is the Hessian matrix, H = R + B. T QB; g is the gradient vector, g = B T Q(x(t+k)-x ref (t+k)).

[0046] The present invention also provides an intelligent control system for unmanned forklifts, the system comprising:

[0047] Perception module: A perception layer is built by deploying multi-dimensional sensors to acquire the status data, position data, and static environmental feature position data of the unmanned forklift;

[0048] Positioning module: Based on the location data of the unmanned forklift and the location data of static environmental features, the module uses a preset position optimization algorithm to correct the location data of the unmanned forklift and generate accurate positioning information.

[0049] Solution module: Based on the state data of the unmanned forklift, a dynamic model of the unmanned forklift is established. Based on the precise positioning information, combined with the dynamic model and path planning results, an optimization problem is defined, and the optimal control sequence is solved using a nonlinear optimization algorithm.

[0050] Control module: Extracts the control input at the current moment based on the optimal control sequence and applies it to the drive and steering system of the unmanned forklift for intelligent control.

[0051] This invention provides an intelligent control method and system for unmanned forklifts. The method constructs a perception layer by deploying multi-dimensional sensors to acquire the unmanned forklift's state data, position data, and static environmental feature position data. Based on the unmanned forklift's position data and static environmental feature position data, a preset position optimization algorithm is used to correct the unmanned forklift's position data, generating precise positioning information. A dynamic model of the unmanned forklift is established based on its state data. Based on the precise positioning information, combined with the dynamic model and path planning results, an optimization problem is defined, and a nonlinear optimization algorithm is used to solve for the optimal control sequence. The control input at the current moment is extracted from the optimal control sequence and applied to the unmanned forklift's drive and steering system for intelligent control. The resulting beneficial effects include:

[0052] 1. High-precision positioning: The system uses a preset position optimization algorithm to correct the position data. By calculating geometric deviation measurement and dynamic residual enhancement, a total cost function is formed and minimized. This enhances the system's adaptability to environmental changes and sensor noise, improves the positioning accuracy of the forklift, and ensures that it can achieve accurate navigation and positioning in dynamic and complex warehouse environments.

[0053] 2. Efficient Path Planning and Tracking Control: Based on precise positioning information and an established dynamic model, combined with path planning results, an optimization problem is defined and a nonlinear optimization algorithm is used to solve for the optimal control sequence. This method can plan future control inputs in advance within the predicted step size, ensuring that the forklift can travel efficiently and smoothly along the predetermined path, reducing path tracking errors and improving the efficiency and reliability of logistics operations.

[0054] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of 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. In the drawings:

[0056] Figure 1 A flowchart illustrating an intelligent control method for an unmanned forklift, as shown in an exemplary embodiment of the present invention;

[0057] Figure 2 This is a schematic diagram illustrating the structure of an intelligent control system for an unmanned forklift, as shown in an exemplary embodiment of the present invention. Detailed Implementation

[0058] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0059] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0060] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0061] Example 1

[0062] An intelligent control method for unmanned forklifts, such as Figure 1 As shown, including:

[0063] By deploying multi-dimensional sensors to build a perception layer, the status data, location data, and static environmental feature location data of the unmanned forklift are acquired.

[0064] Based on the location data of the unmanned forklift and the location data of static environmental features, the location data of the unmanned forklift is corrected using a preset location optimization algorithm to generate accurate positioning information;

[0065] A dynamic model of the unmanned forklift is established based on its state data. Based on the precise positioning information, combined with the dynamic model and path planning results, an optimization problem is defined, and the optimal control sequence is solved using a nonlinear optimization algorithm.

[0066] The control input at the current moment is extracted based on the optimal control sequence and applied to the drive and steering system of the unmanned forklift for intelligent control.

[0067] The present invention further specifies that the multi-dimensional sensor includes a lidar, a camera, an ultrasonic sensor, an inertial measurement unit (IMU), and an ultra-wideband (UWB) positioning base station, and the state data includes linear velocity, driving torque, friction torque, angular velocity, and steering angle. Specifically, in the intelligent control method for unmanned forklifts, accurate perception of the environment and the forklift's own state is the foundation for achieving efficient and precise control. Therefore, the present invention further specifies the composition of the multi-dimensional sensor, including a lidar, a camera, an ultrasonic sensor, an inertial measurement unit (IMU), and an ultra-wideband (UWB) positioning base station. These sensors work together to comprehensively collect status data, position data, and static environmental feature position data of the unmanned forklift. The status data includes linear velocity, which represents the speed of the forklift in a straight line and reflects its movement state; driving torque, which represents the torque acting on the forklift's drive system and is used to control the forklift's acceleration and deceleration; frictional torque, which represents the frictional torque between the forklift's tires and the ground, affecting the forklift's motion stability and handling performance; angular velocity, which represents the forklift's rotational speed and reflects its steering state and rotational dynamics; and steering angle, which represents the steering angle of the forklift's front wheels and is used to control the forklift's steering behavior, affecting the forklift's travel path.

[0068] The present invention is further configured to correct the position data of the unmanned forklift using a preset position optimization algorithm, including:

[0069] Starting from the initial moment, the location data (x1(t), x2(t)) of the unmanned forklift and the location data (l1(t), l2(t)) of the static environment feature are acquired in the time series according to time step t. The measurement distance s(t) between the unmanned forklift and the static environment feature is calculated. Specifically, the static environment feature represents the coordinate position of a fixed landmark feature point in the warehouse environment. The coordinate position is known. Since the position is represented by a two-dimensional plane, the calculation logic of the measurement distance between the unmanned forklift and the static environment feature is the Euclidean distance.

[0070] The geometric deviation metric for time step t is calculated based on the location data (x1(t), x2(t)) of the unmanned forklift, the location data of static environmental features (l1(t), l2(t)) and the measurement distance s(t). The invention further specifies that the calculation logic for the geometric deviation metric for time step t is: D(t) = ((x1(t) - l1(t)) 2 +(x2(t)-l2(t)) 2 ) α -(s(t)) 2α Here, D(t) is the geometric deviation metric at time step t, and α is the order of the polynomial transformation controlling the geometric difference. Specifically, the error between the current position of the unmanned forklift and the static environmental feature point is evaluated by calculating the geometric deviation metric D(t) at time step t. The core idea is to perform a nonlinear transformation on the Euclidean distance between the actual position of the forklift and the position of the feature point, and compare it with the nonlinear transformation result of the distance s(t) measured by the sensor, thereby obtaining a measure reflecting the degree of geometric deviation. By adjusting the order of the polynomial transformation α, the sensitivity of this deviation to different error magnitudes can be controlled. α is usually a positive real number, greater than 0. When α>1, large errors are amplified; when α<1, large errors are weakened, and more attention is paid to the distinction between small and medium-range errors.

[0071] The dynamic residual enhancement at time step t is calculated, and the total cost function is obtained by summing the dynamic residual enhancements at all time steps. The calculation logic for the dynamic residual enhancement at time step t is: R(t)=(D(t)). 2 ·exp(δ·sin(β·s(t))), where R(t) is the dynamic residual enhancement at time step t, δ is the frequency parameter of the modulation sine function, and β is the scaling parameter of the control exponential factor; the calculation logic of the total cost function is as follows: Where C is the total cost function and M is the number of time steps. Specifically, the dynamic residual enhancement R(t) is calculated based on the geometric deviation metric D(t). By introducing an exponential weighting term, the sensitivity to error is enhanced, improving the positioning optimization effect. The dynamic residual enhancements R(t) of all time steps are accumulated and summed to form the total cost function C. By minimizing the total cost function, accurate positioning information is obtained, realizing high-precision positioning and stable control of the unmanned forklift. The frequency parameter δ of the modulated sine function is selected according to the system's response to dynamic changes to control the influence intensity of the exponential weighting. The value range is positive real numbers. The larger δ is, the stronger the influence of the exponential weighting, making the residual more significant under specific conditions. The scaling parameter β of the exponential factor adjusts the frequency of the sine function to control the periodic change of the exponential weighting term. The value range is positive real numbers. The larger β is, the more the frequency of the sine function is increased, making the exponential weighting term fluctuate more frequently with changes in the measured distance.

[0072] Precise positioning information is obtained by minimizing the total cost function to solve for the values ​​of (x1(t), x2(t)). The invention is further configured such that obtaining precise positioning information by minimizing the total cost function to solve for the values ​​of (x1(t), x2(t)) includes:

[0073] Calculate the partial derivatives of the total cost function C with respect to the location variables x1(t) and x2(t). and Specifically, Gradient information is obtained by taking the partial derivatives of the total cost function C with respect to the position variables x1(t) and x2(t), which is used to guide the update direction of the position variables.

[0074] Set initial estimates for location variables. and And the number of iterations k = 0;

[0075] In each iteration k, the geometric deviation metric D is calculated. (k) (t), Dynamic residual enhancement R (k) (t) and partial derivatives and

[0076] According to partial derivatives and Update the position variable. Where η is the learning rate parameter, used to control the update step size; specifically, in each iteration, the position variables x1(t) and x2(t) are adjusted according to the current partial derivative information to gradually reduce the value of the total cost function C. The learning rate parameter η is used to control the step size of the position variable update, affecting the convergence speed and stability of the optimization algorithm, and its value range is a positive real number less than 1.

[0077] Calculate the cost function value for the current iteration step. Determine if the convergence condition is met | C (k+1) -C (k) |<∈, where ∈ is the preset convergence threshold;

[0078] The iteration process terminates and outputs precise location information when the change in the cost function between two consecutive iterations is less than a preset threshold ∈. By calculating partial derivatives and employing an iterative update strategy, the system is ensured to quickly and stably approximate the optimal solution, thus enhancing its robustness and adaptability.

[0079] The present invention is further configured to establish a dynamic model of the unmanned forklift based on its state data, define an optimization problem based on the precise positioning information, the dynamic model, and the path planning results, and solve the optimal control sequence using a nonlinear optimization algorithm, including:

[0080] The linear and angular motion equations of the unmanned forklift are established; the present invention further specifies that the linear motion equation of the unmanned forklift is: Where m is the total mass of the unmanned forklift, v(t) is the linear velocity of the unmanned forklift at time t, and F... d (t) represents the driving torque F acting on the unmanned forklift at time t. f (t) represents the frictional torque acting on the forklift at time t; the angular motion equation of the unmanned forklift is: Where I is the moment of inertia of the unmanned forklift about its center of mass, ω(t) is the angular velocity of the unmanned forklift at time t, L is the wheelbase of the unmanned forklift, θ(t) is the turning angle at time t, and C... f The coefficient of friction between the tires and the ground is used. Specifically, by establishing the linear and angular motion equations of the unmanned forklift, the motion characteristics of the forklift are modeled from a dynamic perspective, providing a precise mathematical foundation for subsequent intelligent control algorithms. By separating and modeling linear and angular motion, the dynamic behavior of the forklift in the forward and turning directions can be described separately. In the linear motion equation, the linear acceleration of the unmanned forklift is determined by the difference between the driving force and the friction force. This is achieved by controlling F... d (t) allows for precise control of the forklift speed; This indicates the steering torque generated by the combined action of the steering angle and driving force. This item explains that when the steering angle is not zero and a driving force is present, the forklift will generate a turning torque, causing it to rotate. This represents the nonlinear frictional torque term related to angular velocity. Its cubic relationship increases the sensitivity to friction at high angular velocities. As the rotational speed increases, the effect of the frictional torque becomes more significant, suppressing excessively high angular velocities and ensuring steering stability. This is achieved by adjusting the steering angle θ(t) and controlling the driving force F. d (t) enables precise control of the forklift's steering behavior;

[0081] Discretize the continuous-time model to obtain the discretized linear and angular motion equations; the discretized linear motion equations are: After rearranging, we get: Δt is the discretized time step; the discretized angular motion equations are: After rearranging, we get: Specifically, by discretizing the continuous-time dynamics model of the unmanned forklift, discretized linear and angular motion equations suitable for digital control systems are obtained. The core purpose of the discretization process is to transform the continuous-time differential equations into discrete-time difference equations, so as to achieve real-time control and computation in the digital controller.

[0082] The dynamic model is represented in matrix form, and the state update equations are constructed. The matrix form of the dynamic model is as follows: Where X(t) is the state vector at time t, and u(t) is the control input vector at time t; the state update equation is: X(t+1)=A·X(t)+B·u(t)+N(X(t),u(t)), where A is the state transition matrix. B is the control input matrix. N(X(t),u(t)) is a nonlinear disturbance term. Specifically, by establishing a state-space model of the unmanned forklift, its dynamic behavior is described to achieve precise motion control and positioning optimization. The state-space model consists of a state vector, a control input vector, a state transition matrix, a control input matrix, and nonlinear disturbance terms. Through this model, the motion state of the unmanned forklift at each time step t can be systematically described, and the linear velocity v(t) and angular velocity ω(t) of the forklift can be adjusted using the control input u(t), thereby achieving precise navigation and control.

[0083] At each control time, an optimization problem with a finite time step N is defined. The optimal control sequence is obtained by minimizing the cost function in the next N steps using a nonlinear optimization algorithm.

[0084] The present invention is further configured such that, at each control moment, an optimization problem with a finite time step N is defined, and the optimal control sequence is obtained by minimizing the cost function over the next N steps using a nonlinear optimization algorithm, including:

[0085] Define the objective function and constraints, and set the initial control sequence U0(t); the present invention further sets the objective function as: Where J(U(t)) is the objective function of the control sequence U(t), x ref (t+k) is the reference state vector, representing the path planning result. Q is the state error weight matrix, and R is the control input weight matrix. The constraints are: x(t+k+1)=A·x(t+k)+B·u(t+k)+N(x(t+k),u(t+k)), k=0,1,…,N-1, and the physical constraints of the control input are: u min ≤u(t+k)≤u maxk = 0, 1, ..., N-1, where u min and u max These are the minimum and maximum vectors of the control input, respectively. Specifically, the objective function J(U(t)) aims to minimize the state error and energy consumption of the control input of the unmanned forklift. This is achieved by weighted summation of the state error and control input over N future time steps, forming an optimization objective. Constraints include: state transition constraints: the next state of the unmanned forklift is determined by the current state, control input, and nonlinear disturbance terms; control input constraints: the control input must be within physical limits to ensure the safe and stable operation of the system; and an initial control sequence U0(t) is used as the starting point of the optimization algorithm. By minimizing the objective function, the optimal control sequence is obtained.

[0086] At the current state point, a Taylor expansion is performed on the nonlinear dynamic model, retaining the first-order terms to obtain a linearized model: x(t+k+1)≈A·x(t+k)+B·u(t+k)+J(t+k)·Δx(t+k), where J is the Jacobian matrix, representing the partial derivatives of the nonlinear terms with respect to the state. Specifically, by linearizing the nonlinear dynamic model of the unmanned forklift, the design and implementation of control and optimization algorithms are simplified. The specific steps include performing a Taylor expansion on the nonlinear model at the current state point, retaining the first-order terms to obtain the linearized model. The linearized model facilitates the application of linear control theory, thereby achieving high-precision positioning and intelligent control of the unmanned forklift. The core idea of ​​this process is to simplify complex nonlinear dynamic behavior using linear approximation, enabling the control algorithm to maintain high accuracy while possessing higher computational efficiency and feasibility. The Jacobian matrix, which describes the linear approximation of a nonlinear function at a certain point, is existing technology and will not be elaborated upon here. By performing a Taylor expansion on the nonlinear dynamic model of the unmanned forklift and retaining the first-order terms, a linearized model is obtained. This invention successfully simplifies the description of complex nonlinear systems, making the motion control and optimization of the unmanned forklift more efficient and accurate. The linearized model is not only applicable to advanced linear control strategies but also significantly improves the system's computational efficiency and real-time response capability.

[0087] Based on the linearized model, a quadratic approximation objective function and constraints are constructed. A sequential quadratic programming algorithm is used to iteratively solve the quadratic approximation problem, updating the control increment until the objective function converges. The control input at the current moment is extracted from the optimal control sequence and applied to the dynamic model. Based on the dynamic model and the control input, the forklift's state is updated. The quadratic approximation objective function and constraints are constructed based on the linearized model. Among them, J approx (ΔU(t)) is the objective function for the second approximation, ΔU(t) is the control input increment vector, and H is the Hessian matrix. g is the gradient vector. Specifically, by constructing a quadratic approximation objective function and constraints based on a linearized model, the Sequential Quadratic Programming (SQP) algorithm is used to iteratively solve the quadratic approximation problem, updating the control increment until the objective function converges. Finally, the control input at the current moment is extracted from the optimal control sequence and applied to the dynamic model. Based on the dynamic model and the control input, the forklift's state is updated. The entire process aims to achieve high-precision positioning and intelligent control of unmanned forklifts. By constructing a quadratic approximation objective function and constraints based on a linearized model and employing the Sequential Quadratic Programming (SQP) algorithm, high-precision positioning and intelligent control of unmanned forklifts in complex dynamic warehousing environments are achieved. Specifically, Taylor expansion is used to linearize the nonlinear dynamic model, simplifying the control and optimization problem; a quadratic approximation objective function J is constructed. approx The constraints (ΔU(t)) make the optimization problem suitable for efficient quadratic programming methods. The SQP algorithm iteratively solves the quadratic approximation problem, gradually updating the control increment until the objective function converges, thereby obtaining the optimal control input. Through a systematic optimization control method, it significantly improves the operational performance, positioning accuracy, and safety of unmanned forklifts in automated logistics systems, demonstrating broad application prospects and significant technical advantages.

[0088] Example 2

[0089] See also Figure 2 The exemplary unmanned forklift intelligent control system includes:

[0090] Perception module: A perception layer is built by deploying multi-dimensional sensors to acquire the status data, position data, and static environmental feature position data of the unmanned forklift;

[0091] Positioning module: Based on the location data of the unmanned forklift and the location data of static environmental features, the module uses a preset position optimization algorithm to correct the location data of the unmanned forklift and generate accurate positioning information.

[0092] Solution module: Based on the state data of the unmanned forklift, a dynamic model of the unmanned forklift is established. Based on the precise positioning information, combined with the dynamic model and path planning results, an optimization problem is defined, and the optimal control sequence is solved using a nonlinear optimization algorithm.

[0093] Control module: Extracts the control input at the current moment based on the optimal control sequence and applies it to the drive and steering system of the unmanned forklift for intelligent control.

[0094] It should be noted that the unmanned forklift intelligent control system provided in the above embodiments and the unmanned forklift intelligent control method provided in the above embodiments belong to the same concept. The specific ways in which each module and unit performs operations have been described in detail in the method embodiments, and will not be repeated here. In practical applications, the unmanned forklift intelligent control system provided in the above embodiments can be assigned to different functional modules as needed, that is, the internal structure of the system can be divided into different functional modules to complete all or part of the functions described above, and this is not a limitation here.

[0095] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0096] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0097] In this application, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or multiple items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0098] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0099] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0100] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0101] In the several embodiments provided in this application, it should be understood that the disclosed system can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0102] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0103] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0104] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0105] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for intelligent control of an unmanned forklift, characterized in that, include: By deploying multi-dimensional sensors to build a perception layer, the status data, location data, and static environmental feature location data of the unmanned forklift are acquired. Based on the location data of the unmanned forklift and the location data of static environmental features, the location data of the unmanned forklift is corrected using a preset location optimization algorithm to generate accurate positioning information; A dynamic model of the unmanned forklift is established based on its state data. Based on the precise positioning information, combined with the dynamic model and path planning results, an optimization problem is defined, and the optimal control sequence is solved using a nonlinear optimization algorithm. This includes establishing the linear and angular motion equations for the unmanned forklift; Discretize the continuous-time model to obtain the discretized linear and angular motion equations; The dynamic model is represented in matrix form, and the state update equation is constructed. Define a finite time step at each control moment. The optimization problem is solved by minimizing the future. The cost function within the step is solved using a nonlinear optimization algorithm to obtain the optimal control sequence; Define the objective function and constraints, and set the initial control sequence. ; At the current state point, a Taylor expansion is performed on the nonlinear dynamic model, retaining the first-order terms to obtain the linearized model; Based on the linear model, a quadratic approximation objective function and constraints are constructed. A sequential quadratic programming algorithm is used to iteratively solve the quadratic approximation problem and update the control increment until the objective function converges. The control input at the current moment is extracted from the optimal control sequence and applied to the dynamic model. The state of the forklift is updated according to the dynamic model and the control input. The objective function is: ,in, For control sequence The objective function, This is the reference state vector corresponding to the path planning result. Here is the state error weight matrix. To control the input weight matrix; The constraints are: And the physical constraints controlling the input: in and Let A be the minimum and maximum value vectors of the control input, respectively; let B be the state transition matrix and B be the control input matrix. for The state vector at time K. for K-time control input vector; Performing a Taylor expansion on the nonlinear dynamic model and retaining the first-order terms yields the linearized model: ,in, Let be the Jacobian matrix, representing the partial derivatives of the nonlinear terms with respect to the state; Based on the linear model, construct the objective function and constraints for a quadratic approximation: ,in, The objective function is a quadratic approximation. To control the input increment vector, It is a Hessian matrix. ; The gradient vector, ; The control input at the current moment is extracted based on the optimal control sequence and applied to the drive and steering system of the unmanned forklift for intelligent control.

2. The intelligent control method for an unmanned forklift according to claim 1, characterized in that, The multi-dimensional sensors include lidar, cameras, ultrasonic sensors, inertial measurement units, and ultra-wideband positioning base stations. The status data includes linear velocity, driving torque, friction torque, angular velocity, and steering angle.

3. The intelligent control method for an unmanned forklift according to claim 2, characterized in that, The position data of the unmanned forklift is corrected using a preset position optimization algorithm, including: Starting from the initial moment, according to the time step Obtain the location data of unmanned forklifts in time series. and static environment feature location data Calculate the measurement distance of unmanned forklifts and static environmental features. ; Based on the location data of the unmanned forklift Static environmental feature location data and measuring distance Calculate time steps Geometric deviation measure; According to time step computation time step The dynamic residual enhancements of all time steps are accumulated and summed to obtain the total cost function; Solving by minimizing the total cost function The value yields precise location information; Time step The calculation logic for the geometric deviation metric is as follows: ,in, For time step Geometric deviation measure To control the order of the polynomial transformation for geometric differences; Time step The calculation logic for dynamic residual enhancement is as follows: ,in, For time step Enhancement of dynamic residuals, For the frequency parameters of the modulation sine function, To control the scaling parameters of the exponential factor; The calculation logic for the total cost function is as follows: ,in, Let be the total cost function. This represents the number of time steps.

4. The intelligent control method for an unmanned forklift according to claim 1, characterized in that, Solving by minimizing the total cost function The value yields precise location information, including: Calculate the minimum total cost function For position variables and partial derivatives and ; Set initial estimates for location variables. and and the number of iterations ; In each iteration In the calculation of geometric deviation measure Dynamic residual enhancement and partial derivatives and ; According to partial derivatives and Update the position variable. ; ,in, The learning rate parameter controls the update step size. Calculate the cost function value for the current iteration step. Determine whether the convergence condition is met. ,in, This is the preset convergence threshold; When the change in the cost function between two consecutive iterations is less than a preset threshold When the iteration process terminates, the precise positioning information is output.

5. The intelligent control method for an unmanned forklift according to claim 1, characterized in that, The linear motion equation of the unmanned forklift is: ,in, The total mass of the unmanned forklift. For unmanned forklifts in time linear velocity, In time The driving torque acting on the unmanned forklift In time The frictional torque acting on the forklift; The equation of motion for the unmanned forklift is: ,in, Let the moment of inertia of the unmanned forklift about its center of mass be denoted as . For unmanned forklifts in time angular velocity, This refers to the wheelbase of the unmanned forklift. For time The steering angle, This is the coefficient of friction between the tire and the ground. The discretized linear equation of motion is: After rearranging, we get: , This represents the time step for discretization; The discretized angular motion equations are: After rearranging, we get: ; The matrix form of the dynamic model is as follows: , ,in, For time The state vector, For time The control input vector; The state update equation is: ,in, Here is the state transition matrix. , To control the input matrix, , It is a nonlinear disturbance term. .

6. An intelligent control system for an unmanned forklift, used to implement the intelligent control method for an unmanned forklift as described in any one of claims 1-5, characterized in that, include: Perception module: A perception layer is built by deploying multi-dimensional sensors to acquire the status data, position data, and static environmental feature position data of the unmanned forklift; Positioning module: Based on the location data of the unmanned forklift and the location data of static environmental features, the module uses a preset position optimization algorithm to correct the location data of the unmanned forklift and generate accurate positioning information. Solution module: Based on the state data of the unmanned forklift, a dynamic model of the unmanned forklift is established. Based on the precise positioning information, combined with the dynamic model and path planning results, an optimization problem is defined, and the optimal control sequence is solved using a nonlinear optimization algorithm. Control module: Extracts the control input at the current moment based on the optimal control sequence and applies it to the drive and steering system of the unmanned forklift for intelligent control.

Images