Concrete transportation unmanned vehicle control method and system facing cooling tower cylinder wall

Abstract

The invention provides a concrete transportation unmanned vehicle control method and system for a cooling tower cylinder wall, and relates to the technical field of building construction.The method comprises the steps that vehicle operation parameters and track characteristic parameters are obtained in real time; calculating the total required torque, and judging whether the vehicle is in a driving working condition, a braking working condition, a steering working condition or a slip working condition; if the vehicle is in the driving working condition, distributing the torque to each wheel by adopting an adaptive particle swarm algorithm based on simulated annealing to reduce the energy consumption of the motor as a target; if the vehicle is in the braking working condition, a multi-target genetic algorithm based on a variable weight coefficient is adopted, energy recovery maximization and braking stability are dynamically balanced, and torque is distributed to all wheels; if the vehicle is in the steering working condition, self-adaptive model predictive control based on a dynamic model is adopted, and torque is distributed to all wheels; and if the wheel is in the slip working condition, an improved sliding mode variable structure control algorithm is adopted, the preset slip rate is taken as a target, and the torque of each wheel is independently controlled by the torque.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of building construction, and in particular to a concrete transportation unmanned vehicle control method and system for cooling tower cylinder wall. BACKGROUND

[0002] With the development of building construction automation and intelligence, unmanned driving technology is gradually applied to special construction scenes with high risk and high intensity. In the construction process of cooling tower cylinder wall, concrete is often transported along the high-altitude ring track to the designated pouring position. Due to the high-rise structure of the cooling tower cylinder wall and the limited construction space, the concrete transportation task is often accompanied by long-time continuous operation, complex track conditions and load mutation. The traditional transportation method relying on manual or conventional mechanical equipment not only has high labor intensity, but also has safety hazards such as falling and overturning. If the intelligent unmanned vehicle can run stably in such scenes, it will significantly improve the construction safety and transportation efficiency.

[0003] However, the existing unmanned vehicle transportation control method is mainly based on a single control algorithm or a fixed torque distribution strategy, which has deficiencies in trajectory tracking, dynamics adaptation, energy recovery and safe operation, and is difficult to meet the actual needs of the complex working conditions of the cooling tower ring track, mainly in the following aspects: Firstly, the cooling tower cylinder wall track is a high-altitude ring structure, which contains long radius curvature sections and slope alternation, and the existing system lacks real-time perception and correction of curvature, slope and other geometric parameters, which easily leads to vehicle body shaking, adhesion reduction or even skidding at curves or slopes.

[0004] Secondly, the concrete load during transportation is dynamic, and as the pouring process gradually decreases, the mass distribution and dynamics of the vehicle change significantly, while the traditional method does not consider the influence of load on vehicle driving and braking performance, which easily causes uneven power distribution and braking failure.

[0005] Thirdly, the construction site has high requirements for energy utilization, and the existing system often ignores the balance between energy recovery and stability during braking, resulting in insufficient vehicle endurance and frequent charging affecting construction efficiency. SUMMARY

[0006] The purpose of the present application is to provide a concrete transportation unmanned vehicle control method and system for cooling tower cylinder wall, to solve the problem of the existing unmanned vehicle transportation control method mainly based on a single control algorithm or a fixed torque distribution strategy, which has deficiencies in trajectory tracking, dynamics adaptation, energy recovery and safe operation.

[0007] To achieve the above object, the present application provides the following technical scheme: a concrete transportation unmanned vehicle control method for cooling tower cylinder wall, steps include: real-time acquisition of vehicle operating parameters and track characteristic parameters; based on the vehicle operating parameters and the track characteristic parameters, the total demand torque is calculated, and it is determined whether the vehicle is currently in driving condition, braking condition, steering condition or slip condition; if it is determined that the vehicle is in driving condition, an adaptive particle swarm algorithm based on simulated annealing is used to distribute the total demand torque to each wheel with the goal of reducing motor energy consumption; if it is determined that the vehicle is in braking condition, a multi-objective genetic algorithm based on variable weight coefficient is used to dynamically balance the maximization of energy recovery and braking stability, and the total demand torque is distributed to each wheel; if it is determined that the vehicle is in steering condition, an adaptive model predictive control based on a dynamics model is used to solve the optimal additional yaw torque and calculate the torque of each wheel through the bottom torque distribution controller; if it is determined that the vehicle is in slip condition, an improved sliding mode variable structure control algorithm is used to independently control the wheel torque with the goal of preset slip rate.

[0008] Optionally, the step of calculating the total demand torque based on the vehicle operating parameters and the track characteristic parameters specifically includes: the vehicle operating parameters include vehicle speed, wheel speed and wheel load; the track characteristic parameters include track curvature and track slope; a longitudinal dynamics model is constructed based on the track slope to calculate feedforward torque; based on the error between the vehicle speed and the expected vehicle speed, parameter adaptive PID control adjustment is performed according to the wheel load and the track slope to obtain feedback torque; the total demand torque is calculated by comprehensively considering the feedforward torque and the feedback torque.

[0009] Optionally, the step of determining whether the vehicle is currently in driving condition, braking condition, steering condition or slip condition specifically includes: when the track curvature is lower than a preset curvature threshold and the total demand torque is positive, it is determined as driving condition; when the track curvature is lower than a preset curvature threshold and the total demand torque is negative, it is determined as braking condition; when the track curvature is higher than a preset curvature threshold, it is determined as steering condition; the wheel slip rate is calculated based on the vehicle speed and the wheel speed, and when the wheel slip rate exceeds a preset slip rate threshold and the duration exceeds a preset time length, it is determined as slip condition.

[0010] Optionally, the adaptive particle swarm algorithm based on simulated annealing includes: the torque distribution coefficient is initialized as a particle, the particle swarm size, particle speed boundary, upper and lower limits of distribution coefficient, inertia weight, learning factor and simulated annealing initial temperature and decay coefficient are set, and a search space of [0, 1] is constructed; an adaptive function is constructed with the minimization of energy consumption as the optimization goal; the front and rear axle distribution torques are calculated by inputting the distribution coefficient, and the total loss power is obtained, wherein the simplified form of the adaptive function is: ;in, The fitness function value; This represents the total drive torque of the front axle; This represents the total drive torque of the rear axle; The front axle single-wheel drive torque is distributed evenly. The rear axle single-wheel drive torque is distributed evenly. For the efficiency of the front axle motor; For the efficiency of the rear axle motor; This refers to the front axle motor speed; The rear axle motor speed is determined; the torque distribution coefficient is updated based on the individual historical best and the group historical best solutions, and an annealing mechanism is introduced to linearly reduce the inertia weight according to the number of iterations in the early stage of iteration to guide global exploration, and to adaptively adjust the learning factor using the current temperature value and accept the probability of inferior solutions for local optimization in the later stage of iteration; when the maximum number of iterations or the convergence condition is met, the iteration is terminated, the optimal torque distribution coefficient is output, and the total required torque is distributed to each wheel according to the optimal torque distribution coefficient.

[0011] Optionally, the multi-objective genetic algorithm based on variable weight coefficients includes: constructing a multi-objective optimization function based on energy recovery rate and stability index, wherein the energy recovery rate is calculated by the ratio of actual recovered energy to ideal recovered energy during vehicle braking, and the stability index is calculated by the matching degree between front and rear axle normal load distribution and braking torque distribution; setting the population size, crossover probability, mutation probability, and number of iterations to construct a multi-objective genetic algorithm model, introducing a variable weight mechanism of fuzzy control, and adaptively adjusting the weight coefficients of the energy recovery rate and the braking stability according to the vehicle operating state; based on a preset object... The system employs a constraint and preset safety conditions, and introduces an elite retention strategy. The individual with the highest fitness in the current population is considered an elite and retained for the next generation, while the individual with the lowest fitness in the next generation is eliminated. This iteratively optimizes the front and rear axle torque distribution coefficients to generate a target solution set that satisfies the preset constraints. The target solution set is then processed offline to generate a query mapping table indexed by vehicle speed and braking intensity, storing the optimal front and rear axle torque distribution coefficients. During vehicle braking, the current vehicle speed and braking intensity are acquired in real time, and the corresponding optimal distribution coefficient is retrieved by querying the mapping table, thereby achieving comprehensive optimization control of braking energy recovery and vehicle stability.

[0012] Optionally, the variable weighting mechanism of the fuzzy control specifically includes: using vehicle speed V and braking intensity Z as input parameters of the fuzzy controller, and outputting the weighting level K1 for energy recovery and the weighting level K2 for braking stability; wherein the universe of discourse for vehicle speed V is set to [0, 30], and the fuzzy subset is {VL, VM, VH}; the universe of discourse for braking intensity Z is set to [0, 1], and the fuzzy subset is {ZL, ZM, ZH}; and the universe of discourse for the weighting coefficients is set to [0, 1]. [1] The fuzzy subset is {LL, L, M, H, HH}, where the weight levels of LL, L, M, H, and HH increase progressively. When the vehicle is at low speed VL, the weight level K1 of braking stability gradually increases from LL to M as the braking intensity Z increases. When the vehicle is at medium speed VM, energy recovery is prioritized under low braking intensity ZL, energy recovery is maximized under medium braking intensity ZM, and braking stability is prioritized under high braking intensity ZH. When the vehicle is at high speed VH, the weight level of braking stability increases to M under low braking intensity ZL, increases to H under medium braking intensity ZM, and increases to HH under high braking intensity, while the weight level of energy recovery decreases to LL.

[0013] Optionally, the steps of the adaptive model predictive control based on the dynamic model specifically include: constructing a discrete state-space model of vehicle dynamics based on vehicle operating parameters and track characteristic parameters, using the vehicle's longitudinal speed, lateral speed, yaw rate, yaw angle, longitudinal position and lateral position in the global coordinate system as state variables, and using the additional yaw torque as the control input; predicting the state variables in the future time domain based on the discrete state-space model of vehicle dynamics, and constructing a cost function to comprehensively evaluate trajectory tracking error, yaw stability and control smoothness; adaptively adjusting the lateral error weight and yaw error weight in the cost function, increasing the lateral error weight and yaw error weight when the track curvature increases to improve curve tracking accuracy and stability, and decreasing the lateral error weight when the track curvature decreases to optimize control smoothness; in each control cycle, performing secondary programming based on the adaptively adjusted cost function and introducing control input constraints and control increment constraints, solving online to obtain the optimal additional yaw torque, and calculating the torque of each wheel through the underlying torque distribution controller.

[0014] Optionally, the adaptive adjustment function is: ; ;in, For curvature; As the reference curvature; This is the saturation value; Baseline horizontal weights; Weighting of baseline yaw rate error; As the baseline yaw angle error weight; This is the horizontal weighting amplification factor; For horizontal error weights; As the weight for the yaw rate error; This is the weight for the yaw angle error; This represents the weighting of the lateral accuracy error.

[0015] Optionally, the step of independently controlling the torque of each wheel using the improved sliding mode variable structure control algorithm with a preset slip ratio as the target specifically includes: calculating the slip ratio based on the wheel speed and vehicle speed, constructing a sliding mode surface based on the slip ratio and the target slip ratio, and designing a sliding mode control law to generate a control signal. The calculation formula for the sliding mode control law is as follows: ;in, The rate of change of the sliding surface over time; For sliding surface functions; This is the gain of the isotropic approaching term; The gain is the exponentially approaching term; This represents the actual slip ratio of the wheel; The target slip ratio, i.e., the preset slip ratio, is used. Based on the control signal, vehicle motion state parameters, and wheel dynamic parameters, the target control torque applied to the hub motor is calculated. The formula for calculating the target control torque is as follows: ;in, To control the torque to the target; The moment of inertia of the wheel; The longitudinal velocity at the center of the wheel; For sliding surface functions; This is the gain of the isotropic approaching term; The gain is the exponentially approaching term; This represents the actual slip ratio of the wheel; The radius of the wheel; It is a saturation function; The longitudinal reaction force of the contact surface to the wheel; The longitudinal acceleration at the wheel center; the target control torque An application is made to the corresponding hub motor so that the actual slip ratio of the drive wheel converges to the preset slip ratio.

[0016] On the other hand, the present invention also provides a control system for an unmanned concrete transport vehicle facing the wall of a cooling tower, comprising: an acquisition module for acquiring vehicle operating parameters and track characteristic parameters in real time; a working condition determination module for calculating the total required torque based on the vehicle operating parameters and the track characteristic parameters, and determining whether the vehicle is currently in a driving condition, braking condition, steering condition, or slip condition; a driving condition module for, if the vehicle is determined to be in a driving condition, distributing the total required torque to each wheel using an adaptive particle swarm optimization algorithm based on simulated annealing, with the goal of reducing motor energy consumption; and a braking condition module. The system includes three modules: a braking module and a steering module. The steering module is used to dynamically balance maximizing energy recovery with braking stability using a multi-objective genetic algorithm based on variable weight coefficients, and distribute the total required torque to each wheel. The steering module is used to solve for the optimal additional yaw torque and calculate the torque of each wheel through an underlying torque distribution controller if the vehicle is determined to be in a steering condition. The slippage module is used to independently control the torque of each wheel using an improved sliding mode variable structure control algorithm with a preset slip ratio as the target if the vehicle is determined to be in a slippage condition.

[0017] Compared with the prior art, the beneficial effects of the present invention are: This application achieves efficient and stable operation of an intelligent unmanned vehicle transporting cooling tower wall concrete in complex track environments by introducing multi-dimensional perception and multi-algorithm collaborative optimization control. It acquires vehicle operating parameters and track characteristic parameters in real time, identifies conventional driving, conventional braking, steering, and slippage conditions, and matches corresponding control algorithms. An adaptive feedforward PID method is used to calculate and correct the required torque, effectively ensuring rapid response and steady-state accuracy of longitudinal drive under load changes and gradient fluctuations. In driving conditions, an adaptive particle swarm optimization algorithm based on simulated annealing is used to optimize the allocation of total required torque, achieving low-energy drive. In braking conditions, multi-objective optimization coordinates braking energy recovery and smooth braking. In steering conditions, adaptive model predictive control based on a dynamic model is used for trajectory tracking, improving the safety and stability of the unmanned vehicle on high-altitude circular tracks. In slippage conditions, improved sliding mode variable structure control suppresses wheel slippage and maintains optimal adhesion. This application enables the unmanned vehicle to achieve global optimized control under complex working conditions while ensuring operational safety, significantly improving driving stability, safety and endurance, and is particularly suitable for high-altitude curved track scenarios such as transporting concrete for cooling tower walls. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the method steps of the present invention.

[0019] Figure 2This is a flowchart of the adaptive particle swarm optimization algorithm based on simulated annealing of the present invention.

[0020] Figure 3 This is a flowchart of the multi-objective genetic algorithm based on variable weight coefficients of the present invention.

[0021] Figure 4 This is a schematic diagram of the system structure of the present invention.

[0022] In the diagram: 10 - Acquisition module, 20 - Working condition determination module, 30 - Driving working condition module, 40 - Braking working condition module, 50 - Steering working condition module, 60 - Slippage working condition module. Detailed Implementation

[0023] The present invention will now be clearly and completely described in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0024] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be used interchangeably where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0025] Those skilled in the art will understand that, unless explicitly stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in the specification of this application means the presence of features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or wireless coupling. The term “and / or” as used herein includes all or any units and all combinations of one or more associated listed items.

[0026] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0027] It should be understood that the sequence number and size of each step in this embodiment do not imply the order of execution. The execution order of each process is determined by its function and internal logic, and should not constitute any limitation on the implementation process of this application embodiment.

[0028] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0029] Please refer to Figures 1-3 This invention discloses a control method for an unmanned concrete transport vehicle oriented towards the wall of a cooling tower. The method includes the following steps: real-time acquisition of vehicle operating parameters and track characteristic parameters; calculation of the total required torque based on the vehicle operating parameters and track characteristic parameters, and determination of whether the vehicle is currently in a driving, braking, steering, or slipping condition; if the vehicle is in a driving condition, an adaptive particle swarm optimization algorithm based on simulated annealing is used to distribute the total required torque to each wheel with the goal of reducing motor energy consumption; if the vehicle is in a braking condition, a multi-objective genetic algorithm based on variable weight coefficients is used to dynamically balance maximizing energy recovery and braking stability, and distribute the total required torque to each wheel; if the vehicle is in a steering condition, an adaptive model predictive control based on a dynamic model is used to solve for the optimal additional yaw torque and calculate the torque of each wheel through a bottom-level torque distribution controller; if the vehicle is in a slipping condition, an improved sliding mode variable structure control algorithm is used to independently control the torque of each wheel with a preset slip ratio as the target.

[0030] This application achieves efficient and stable operation of an intelligent unmanned vehicle transporting cooling tower wall concrete in complex track environments by introducing multi-dimensional perception and multi-algorithm collaborative optimization control. It acquires vehicle operating parameters and track characteristic parameters in real time, identifies conventional driving, conventional braking, steering, and slippage conditions, and matches corresponding control algorithms. An adaptive feedforward PID method is used to calculate and correct the required torque, effectively ensuring rapid response and steady-state accuracy of longitudinal drive under load changes and gradient fluctuations. In driving conditions, an adaptive particle swarm optimization algorithm based on simulated annealing is used to optimize the allocation of total required torque, achieving low-energy drive. In braking conditions, multi-objective optimization coordinates braking energy recovery and smooth braking. In steering conditions, adaptive model predictive control based on a dynamic model is used for trajectory tracking, improving the safety and stability of the unmanned vehicle on high-altitude circular tracks. In slippage conditions, improved sliding mode variable structure control suppresses wheel slippage and maintains optimal adhesion. This application enables the unmanned vehicle to achieve global optimized control under complex working conditions while ensuring operational safety, significantly improving driving stability, safety and endurance, and is particularly suitable for high-altitude curved track scenarios such as transporting concrete for cooling tower walls.

[0031] In some embodiments, the step of calculating the total required torque based on the vehicle operating parameters and the track characteristic parameters specifically includes: the vehicle operating parameters including vehicle speed, wheel speed, and wheel load; the track characteristic parameters including track curvature and track slope; constructing a longitudinal dynamic model based on the track slope to calculate the feedforward torque; based on the error between the vehicle speed and the expected vehicle speed, performing parameter adaptive PID control adjustment according to the wheel load and the track slope to obtain the feedback torque; and combining the feedforward torque and the feedback torque to calculate the total required torque.

[0032] Specifically, the acquisition module is installed on the intelligent unmanned vehicle body, including lidar, inertial measurement unit, wheel speed sensor, GPS and gravity sensor, etc., for multi-source perception of vehicle operating parameters and track characteristic parameters.

[0033] The vehicle's operating parameters and track characteristic parameters are collected using point clouds via LiDAR. Noise is removed through statistical filtering and contact surface segmentation, retaining only track-related points. Track edge points are extracted based on geometric features and clustering algorithms, and the track centerline is generated using polynomial fitting. Track curvature is calculated in real time to distinguish between straight and curved sections. Using an IMU, after calibration and filtering, the vehicle's pitch angle is obtained and approximated as the track slope. Abnormal bump signals are eliminated to ensure stable and reliable slope detection. The wheel speed of the unmanned vehicle is acquired in real time using Hall effect wheel speed sensors. In environments with enhanced positioning signals, Kalman filtering fusion of GPS and IMU data is introduced to ensure dynamic response and long-term accuracy of speed measurement. Strain gauge weight sensors are placed at the bottom of the hopper, and the concrete load is calculated in real time using signal amplification, filtering, and calibration curves. All sensor data are aligned based on a unified timestamp to form comprehensive status information including vehicle speed, load, track curvature, and slope, ensuring the safe operation of the unmanned vehicle.

[0034] Specifically, based on the total mass of the autonomous vehicle With track slope A longitudinal dynamic model is constructed, and the theoretical driving torque is calculated as feedforward compensation: ; in, For feedforward compensation torque; The total mass of the driverless vehicle; It is the acceleration due to gravity; For the desired acceleration; For rolling resistance; For air resistance; The radius is the wheel radius.

[0035] Meanwhile, based on vehicle speed error Feedback correction is performed using a parameter adaptive PID controller. ; in, For feedback correction torque; The proportional gain of the PID controller; The integral coefficient of the PID controller; The derivative coefficients of the PID controller; This represents the rate of change of vehicle speed error.

[0036] The PID parameters are dynamically adjusted based on vehicle speed, load, and gradient to determine the total torque demand. ; in, This represents the total required torque. For feedforward compensation torque; This is for feedback correction torque.

[0037] This application combines feedforward torque and parameter-adaptive PID feedback correction to improve the accuracy and response speed of torque control. By dynamically adapting to changes in vehicle load and track gradient, it reduces lag and errors in longitudinal drive, ensuring smooth operation of the unmanned vehicle on slopes or under fluctuating load conditions, thereby enhancing the reliability and efficiency of the transportation process.

[0038] In some embodiments, the step of determining whether the vehicle is currently in a driving condition, braking condition, steering condition, or slip condition specifically includes: when the track curvature is lower than a preset curvature threshold and the total required torque is positive, it is determined to be a driving condition; when the track curvature is lower than a preset curvature threshold and the total required torque is negative, it is determined to be a braking condition; when the track curvature is higher than a preset curvature threshold, it is determined to be a steering condition; the wheel slip ratio is calculated based on the vehicle speed and the wheel rotation speed, and when the wheel slip ratio exceeds a preset slip ratio threshold and the duration exceeds a preset duration, it is determined to be a slip condition.

[0039] Specifically, when the orbital curvature Less than the set threshold At that time, it is determined to be a straight-moving working condition: In drive mode, an adaptive particle swarm optimization algorithm based on simulated annealing is used to determine the total required torque. The goal of optimizing the energy allocation is to reduce motor energy consumption. In braking mode, a multi-objective genetic algorithm based on variable weight coefficients is used to dynamically balance maximizing energy recovery and maintaining braking stability, thereby achieving optimal distribution of braking torque across the four wheels.

[0040] When the orbital curvature Greater than the set threshold At this time, it is determined to be a steering condition: An adaptive model predictive control based on a dynamic model is adopted to solve for the optimal additional yaw torque and to calculate the torque of each wheel through a low-level torque distribution controller.

[0041] When the wheel slip ratio exceeds a preset slip ratio threshold and the duration exceeds a preset duration, it is determined to be a slipping condition. The slip ratio is calculated by detecting the rotational speed of each wheel using wheel speed sensors and combining this with the overall vehicle speed. ; in, Wheel slip ratio; The wheel's angular velocity; The radius of the wheel's rolling radius; The speed at the center of the wheel.

[0042] Based on the target slip ratio, a sliding surface is constructed, and a sliding control law is designed to adjust the braking torque of the wheel, thereby suppressing slippage and ensuring that the wheel always works in the optimal adhesion range.

[0043] When multiple operating conditions occur simultaneously, the system executes control according to safety priority, that is, it first handles the slipping condition, then the steering condition, and finally the normal driving or braking condition, thereby ensuring the safety and stability of the unmanned vehicle in the complex track environment of the cooling tower wall.

[0044] This application, through preset thresholds and duration determination, can quickly distinguish between driving, braking, steering, and slipping conditions. In the driving condition, an adaptive particle swarm optimization algorithm based on simulated annealing is used to optimize the allocation of total required torque, achieving low-energy driving. In the braking condition, multi-objective optimization is used to coordinate braking energy recovery and smooth braking. In the steering condition, adaptive model predictive control based on a dynamic model is used for trajectory tracking, improving the operational safety and stability of the unmanned vehicle on a high-altitude circular track. In the slipping condition, improved sliding mode variable structure control is used to suppress wheel slippage and maintain optimal adhesion, significantly improving the unmanned vehicle's adaptability and safety redundancy on complex tracks.

[0045] In some embodiments, the adaptive particle swarm optimization algorithm based on simulated annealing includes: initializing the torque distribution coefficient as particles, setting the particle swarm size, particle velocity boundary, upper and lower limits of the distribution coefficient, inertia weight, learning factor, and initial simulated annealing temperature and decay coefficient, and constructing a search space of [0,1]; constructing a fitness function with energy minimization as the optimization objective; calculating the front and rear axle torque distribution and obtaining the total power loss by inputting the distribution coefficient, wherein the simplified form of the fitness function is: ;in, The fitness function value; This represents the total drive torque of the front axle; This represents the total drive torque of the rear axle; The front axle single-wheel drive torque is distributed evenly. The rear axle single-wheel drive torque is distributed evenly. For the efficiency of the front axle motor; For the efficiency of the rear axle motor; This refers to the front axle motor speed; The rear axle motor speed is determined; the torque distribution coefficient is updated based on the individual historical best and the group historical best solutions, and an annealing mechanism is introduced to linearly reduce the inertia weight according to the number of iterations in the early stage of iteration to guide global exploration, and to adaptively adjust the learning factor using the current temperature value and accept the probability of inferior solutions for local optimization in the later stage of iteration; when the maximum number of iterations or the convergence condition is met, the iteration is terminated, the optimal torque distribution coefficient is output, and the total required torque is distributed to each wheel according to the optimal torque distribution coefficient.

[0046] Specifically, in the autonomous vehicle drive control process, firstly, parameter initialization is completed based on the constraints of the vehicle drive system to construct an optimization search space, with the front and rear axle drive torque distribution coefficients ranging from 0 to 1. Subsequently, based on the motor drive system efficiency matrix obtained from bench tests, a fitness function is constructed with the goal of minimizing energy consumption. The front and rear axle distributed torques and total power loss are calculated according to the distribution coefficients. In the iterative optimization phase, a simulated annealing adaptive particle swarm optimization algorithm is used to update the distribution coefficients. The particle swarm optimization uses individual and group optima to guide the global search, while the simulated annealing mechanism improves the accuracy of local optimization by dynamically adjusting the inertia weights and learning factors, avoiding getting trapped in local optima. When the fitness function converges or reaches the set number of iterations, the optimal distribution coefficients are output, and the front and rear axle target torque matrix is ​​generated. This result is input to the drive actuator in real time to achieve efficient drive control of the autonomous vehicle.

[0047] First, the particle swarm is initialized, and its fitness function is calculated. In the early stage of iteration, only better solutions are accepted. In the later stage of iteration, the Metropolis criterion is used to select whether to accept worse solutions. Individual and global optimum values ​​are updated, and the inertia weight coefficient and self and swarm learning factors are adjusted. Next, the velocity and position of the particles are updated, and boundary conditions are handled to reduce the temperature in the simulated annealing. Finally, it is determined whether the conditions are met. If they are met, the run ends. If not, the fitness function is recalculated, and the above process is repeated.

[0048] Under the current driving torque and motor speed, while keeping the total output power of the motor drive system constant, the overall energy consumption can be optimized by minimizing the total input power of the motor drive system. The total input power of the motor drive system can be expressed as: ; in, , The rotational speed of a single drive motor on the front and rear axles; , The efficiency of the front and rear axle single motor drive systems under the current torque and speed; The sum of the target torques distributed to the two drive motors on the front axle; This represents the sum of the target torques allocated to the two drive motors on the rear axle. To improve the algorithm's efficiency, the fitness function can be simplified to: ; in, The fitness function value; This represents the total drive torque of the front axle; This represents the total drive torque of the rear axle; The front axle single-wheel drive torque is distributed evenly. The rear axle single-wheel drive torque is distributed evenly. For the efficiency of the front axle motor; For the efficiency of the rear axle motor; This refers to the front axle motor speed; This refers to the speed of the rear axle motor.

[0049] In particle swarm optimization, each particle represents an allocation coefficient k, and its velocity and position are updated during the iteration process. ; in, Inertial weights; For the individual's optimal; It is the global optimum; , A random number between 0 and 1; , For self and social factors.

[0050] The simulated annealing mechanism allows for the acceptance of "worse solutions" with a probability in the later stages of iteration, and the acceptance probability follows the Metropolis criterion: ; Temperature updates use exponential decay: ; in, For temperature; This is the cooling coefficient.

[0051] Simultaneously adjust the learning factor dynamically: ; in, , Initial values ​​for self and social factors; The initial temperature; It is a random value between 0 and 1.

[0052] The inertia weight is adjusted as follows: ; in, This represents the current iteration number; This represents the maximum number of iterations. This represents the inertia weight for the current iteration; This represents the maximum value of the inertia weight; This represents the minimum value of the inertia weight.

[0053] This application employs an adaptive particle swarm optimization algorithm based on simulated annealing for driving conditions, achieving optimization that minimizes energy consumption. By simplifying the fitness function, the algorithm's operating efficiency is improved, and by optimizing the torque distribution coefficient, the total loss of the motor system is reduced, extending the autonomous vehicle's range and enhancing its economic efficiency.

[0054] In some embodiments, the multi-objective genetic algorithm based on variable weight coefficients includes: constructing a multi-objective optimization function based on energy recovery rate and stability index, wherein the energy recovery rate is calculated by the ratio of actual recovered energy to ideal recovered energy during vehicle braking, and the stability index is calculated by the matching degree between front and rear axle normal load distribution and braking torque distribution; setting the population size, crossover probability, mutation probability, and number of iterations to construct a multi-objective genetic algorithm model; introducing a variable weight mechanism of fuzzy control; and adaptively adjusting the weight coefficients of the energy recovery rate and the braking stability according to the vehicle operating state; based on pre- Physical constraints and preset safety conditions are set, and an elite retention strategy is introduced. The individual with the highest fitness in the current population is regarded as an elite individual and retained in the next generation population, while the individual with the lowest fitness in the next generation population is eliminated. The front and rear axle torque distribution coefficients are iteratively optimized to generate a target solution set that meets the preset constraints. The target solution set is processed offline to generate a query mapping table indexed by vehicle speed and braking intensity, which stores the optimal front and rear axle torque distribution coefficients. When the vehicle brakes, the current vehicle speed and braking intensity are obtained in real time, and the corresponding optimal distribution coefficient is retrieved by querying the mapping table to achieve comprehensive optimization control of braking energy recovery and vehicle stability.

[0055] Specifically, regarding braking stability and braking energy recovery rate Let be the objective function. As shown below: ; in, It is a multi-objective optimization function; This is the weighting coefficient for braking stability; This is the weighting coefficient for braking energy recovery rate; For braking stability indicators; This refers to the braking energy recovery rate.

[0056] The closer the coefficient of adhesion is to the braking intensity, the more fully the adhesion conditions are utilized and the more rational the braking force distribution. Ideally, the coefficient of adhesion should always equal the braking intensity; therefore, the degree to which the braking intensity approaches the coefficient of adhesion is used as a basis for stability. The formula is as follows: ; in, This refers to the coefficient of friction of the vehicle's front wheels. This is the coefficient of friction for the rear wheels of the vehicle.

[0057] The ratio of the actual recovered energy to the recoverable energy of the power battery is used as the objective function for regenerative braking, as shown in the following formula: ; in, This refers to the actual energy recovered. To recover energy for the ideal purpose.

[0058] This application employs a multi-objective genetic algorithm based on variable weight coefficients to collaboratively optimize energy recovery and braking stability under braking conditions. By dynamically balancing the recovery rate and adhesion conditions, energy regeneration is maximized while ensuring balanced vehicle braking force, reducing battery charging frequency, thereby improving construction efficiency and lowering operating costs.

[0059] In some embodiments, the variable weighting mechanism of the fuzzy control specifically includes: using vehicle speed V and braking intensity Z as input parameters of the fuzzy controller, and outputting the weighting level K1 for energy recovery and the weighting level K2 for braking stability; wherein, the universe of discourse of vehicle speed V is set to [0,30], and the fuzzy subset is {VL, VM, VH}; the universe of discourse of braking intensity Z is set to [0,1], and the fuzzy subset is {ZL, ZM, ZH}; the universe of discourse of the weighting coefficients is set to [0,1], and the fuzzy subset is {LL, L, M, H, HH}, wherein the weighting levels of LL, L, M, H, and HH are... The weighting is progressively increased. At low vehicle speed (VL), the weighting level of braking stability (K1) gradually increases from LL to M as the braking intensity (Z) increases. At medium vehicle speed (VM), energy recovery is prioritized at low braking intensity (ZL), energy recovery is maximized at medium braking intensity (ZM), and braking stability is prioritized at high braking intensity (ZH). At high vehicle speed (VH), the weighting level of braking stability increases to M at low braking intensity (ZL), increases to H at medium braking intensity (ZM), and increases to HH at high braking intensity, while the weighting level of energy recovery decreases to LL.

[0060] Specifically, based on the variable weight coefficient mechanism of fuzzy control, vehicle speed and braking intensity are selected as input parameters of the fuzzy controller, and the output is the proportional coefficient of the two objective functions, so as to realize the dynamic allocation of braking stability and braking energy recovery weights under different driving conditions.

[0061] The fuzzy membership function is designed as follows: The universe of discourse for vehicle speed V is set to [0, 30], and the fuzzy subset is {VL, VM, VH}; the universe of discourse for braking intensity Z is set to [0, 1], and the fuzzy subset is {ZL, ZM, ZH}; where L, M, and H represent low, medium, and high membership levels, respectively. The universe of discourse for the proportional coefficients of stability weight level K1 and braking energy recovery weight level K2 are both [0, 1], and the fuzzy subsets are both {LL, L, M, H, HH}, where LL, L, M, H, and HH represent very low, low, medium, high, and very high membership levels, respectively.

[0062] At low speeds, due to limited kinetic energy, the weight of energy recovery is generally low, while the weight of stability gradually increases from very small to medium as braking intensity increases. At medium speeds, energy recovery is prioritized at low braking intensity, energy recovery is maximized at medium braking intensity, and stability is prioritized at high braking intensity. At high speeds, even at low braking intensity, a medium stability weight is required to maintain vehicle posture. The stability weight increases to a large value at medium braking intensity, and the stability weight is increased to a maximum at high braking intensity, while the weight of energy recovery decreases to a minimum.

[0063] Based on the above principles and with full consideration for driver safety, nine fuzzy rules were established, as shown in Table 1 below: Table 1: Fuzzy rule table.

[0064] SEQ ID NO V Z K1 K2 1 VL ZL LL L 2 VL ZM L L 3 VL ZH M M 4 VM ZL L M 5 VM ZM M HH 6 VM ZH H M 7 VH ZL M H 8 VH ZM H M 9 VH ZH HH LL This application enhances the flexibility and scenario adaptability of the control system by introducing a variable weight mechanism of fuzzy control, which adaptively adjusts the weights of the objective function based on vehicle speed and braking intensity. By dynamically allocating economic and safety priorities through fuzzy rules, the autonomous vehicle can cope with different driving states, such as low-speed light braking or high-speed emergency braking, thus improving the response accuracy to complex environments.

[0065] In some embodiments, the steps of the adaptive model predictive control based on the dynamic model specifically include: constructing a discrete state-space model of vehicle dynamics based on vehicle operating parameters and track characteristic parameters, using the vehicle's longitudinal speed, lateral speed, yaw rate, yaw angle, longitudinal position and lateral position in the global coordinate system as state variables, and using the additional yaw torque as the control input; predicting the state variables in the future time domain based on the discrete state-space model of vehicle dynamics, and constructing a cost function to comprehensively evaluate trajectory tracking error, yaw stability and control smoothness; adaptively adjusting the lateral error weight and yaw error weight in the cost function, increasing the lateral error weight and yaw error weight when the track curvature increases to improve curve tracking accuracy and stability, and decreasing the lateral error weight when the track curvature decreases to optimize control smoothness; in each control cycle, performing secondary programming based on the adaptively adjusted cost function and introducing control input constraints and control increment constraints, solving online to obtain the optimal additional yaw torque, and calculating the torque of each wheel through the underlying torque distribution controller.

[0066] Specifically, since this type of vehicle lacks a steering mechanism and traditional wheel slip angle control, this invention selects the vehicle's longitudinal velocity, lateral velocity, yaw rate, yaw angle, and longitudinal and lateral positions in the global coordinate system as state variables, and selects an additional yaw torque as the control input. The vehicle state is adjusted by the additional yaw torque. This method predicts the trajectory deviation for several future steps based on discrete state-space equations, and simultaneously considers trajectory tracking error, yaw stability, and control smoothness in the cost function. By dynamically adjusting the optimization objective with adaptive weight factors according to the track curvature, high-precision trajectory tracking and stable control of the track-mounted unmanned vehicle are achieved.

[0067] The state-space equations of a model predictive control system can be simplified as follows: ; in, for The system state variables at any given time; For system outputs; select vehicle yaw angle, yaw rate, and lateral displacement; .

[0068] State vector: ; in, The longitudinal speed of the vehicle; This refers to the lateral speed. This refers to the yaw rate; It is the horizontal swing angle; This refers to the vertical position in the global coordinate system. This represents the horizontal position in the global coordinate system.

[0069] Control input: ; Linearization yields the linear time-varying state equation: ; Discretization using the forward Euler method yields: ; Change to incremental: ; in, ; Represents the identity matrix; Represents the system sampling interval; To calculate the step size for discretization; .

[0070] make ; The new system space state expression is then obtained as follows: ; in, ; ; This is the output matrix; This is the system output.

[0071] The matrix expression is: ; Among them, system output System input ; ; ; Construct a cost function with the objectives of trajectory tracking, suppressing yaw instability, and ensuring control smoothness: ; in, This represents the desired output vector; To constrain slack variables; Indicates the weights of slack variables; For prediction in the time domain; To control the time domain. This is the weight matrix for the state tracking error; The matrix used to control the incremental smoothing weights.

[0072] To enhance lateral accuracy and stability at curves, a curvature-based approach is proposed. The adaptive weight adjustment strategy is as follows: when the curvature increases, the weights of lateral error, yaw rate error and yaw angle error are increased; when the curvature decreases, the lateral weights are reduced to prioritize smoothness.

[0073] This application employs adaptive predictive control based on a dynamic model in the steering condition. By predicting the future trajectory and optimizing torque distribution, it reduces lateral deviation and heading error during cornering, ensuring smooth steering of the unmanned vehicle on the circular track and avoiding the risk of sideslip or derailment.

[0074] In some embodiments, the adaptive adjustment function is: ; ; in, For curvature; As the reference curvature; This is the saturation value; Baseline horizontal weights; Weighting of baseline yaw rate error; As the baseline yaw angle error weight; This is the horizontal weighting amplification factor; For horizontal error weights; As the weight for the yaw rate error; This is the weight for the yaw angle error; This represents the weighting of the lateral accuracy error.

[0075] Set the control increment constraint and control quantity constraint as follows: ; The aforementioned constrained optimization problem can be solved using quadratic programming. The general form of quadratic programming is as follows: ; in, Represents the coefficient matrix of the quadratic form. Represents the vector of coefficients of linear terms. , All are zero-matrix dimension fits and are inequality constraint matrices. , These represent the upper and lower limits of the constraint conditions, respectively.

[0076] The optimal control increment sequence in the control time domain can be obtained through quadratic programming. The first element of the control sequence is extracted to obtain the optimal control quantity at the current sampling time. The optimal output is calculated, and the torque of each wheel is calculated through the underlying torque distribution controller.

[0077] The torque of each wheel is shown below: ; in, This represents the total required torque. This refers to the torque of the left front wheel; This refers to the torque of the left rear wheel; This refers to the torque of the right front wheel; This refers to the torque of the right rear wheel; The radius of the wheel's rolling radius; This refers to the wheel track.

[0078] This application optimizes the performance trade-off in cornering control by adaptively adjusting the weights of lateral and yaw errors based on curvature. When curvature increases, it prioritizes tracking accuracy and stability, while focusing on smoothness when curvature is low. This improves the overall performance of the autonomous vehicle on different track sections and achieves a dynamic balance between safety and smoothness.

[0079] In some embodiments, the step of independently controlling the torque of each wheel using an improved sliding mode variable structure control algorithm with a preset slip ratio as the target specifically includes: calculating the slip ratio based on the wheel speed and vehicle speed, constructing a sliding mode surface based on the slip ratio and the target slip ratio, and designing a sliding mode control law to generate a control signal. The calculation formula for the sliding mode control law is as follows: ;in, The rate of change of the sliding surface over time; For sliding surface functions; This is the gain of the isotropic approaching term; The gain is the exponentially approaching term; This represents the actual slip ratio of the wheel; The target slip ratio, i.e., the preset slip ratio, is used. Based on the control signal, vehicle motion state parameters, and wheel dynamic parameters, the target control torque applied to the hub motor is calculated. The formula for calculating the target control torque is as follows: ;in, To control the torque to the target; The moment of inertia of the wheel; The longitudinal velocity at the center of the wheel; For sliding surface functions; This is the gain of the isotropic approaching term; The gain is the exponentially approaching term; This represents the actual slip ratio of the wheel; The radius of the wheel; It is a saturation function; The longitudinal reaction force of the contact surface to the wheel; This is the longitudinal acceleration at the center of the wheel.

[0080] The target control torque An application is made to the corresponding hub motor so that the actual slip ratio of the drive wheel converges to the preset slip ratio.

[0081] Specifically, with a preset slip ratio as the target and the slip ratio of the four wheels as the input, a slip mode controller is designed to control the torque of the four wheel hub motors, so that the actual slip ratio of the wheels is close to the preset slip ratio.

[0082] The simplified kinematic model of a single wheel of a vehicle is as follows: ; in, The moment of inertia of the wheel; This refers to the wheel's angular acceleration; The reaction force provided to the contact surface; Wheel torque; The radius is the wheel radius.

[0083] With the objective of keeping the wheel slip ratio near a preset slip ratio, the switching function for sliding mode control is defined as follows: ; in, This is the switching function for the sliding surface; This represents the actual slip ratio of the wheel; The target slip ratio is the preset slip ratio.

[0084] Taking the first derivative of the switching function, we get: ; Further, we can conclude that: ; The comprehensive sliding mode control system is as follows: ; in, The moment of inertia of the wheel; This is the exponential convergence law; This refers to the longitudinal acceleration at the center of the wheel. This refers to the wheel's angular acceleration; The reaction force provided to the contact surface; Wheel torque; The radius is the wheel radius.

[0085] Choose the exponential approach law: ; in, These are terms that approach at a constant velocity. This is the exponential term.

[0086] To effectively reduce chattering in sliding mode controllers, a continuous saturation function is used. Replacement symbol function The saturation function is defined as: ; in, This represents the boundary layer width.

[0087] Therefore, the target control torque for sliding mode control is: ; in, To control the torque to the target; The moment of inertia of the wheel; The longitudinal velocity at the center of the wheel; For sliding surface functions; This is the gain of the isotropic approaching term; The gain is the exponentially approaching term; This represents the actual slip ratio of the wheel; The radius of the wheel; It is a saturation function; The longitudinal reaction force of the contact surface to the wheel; This is the longitudinal acceleration at the center of the wheel.

[0088] This application employs an improved sliding mode variable structure control to suppress wheel slippage, enabling rapid restoration of wheel traction and preventing further slippage. By adjusting the braking torque in real time through a sliding mode control law, it ensures that the wheels always operate within the optimal slip ratio range, enhancing the traction of the unmanned vehicle under wet, slippery, or steep slope conditions, thereby improving safety performance in emergency situations.

[0089] Please refer to Figure 4 On the other hand, the present invention also provides a control system for an unmanned concrete transport vehicle facing the wall of a cooling tower, comprising: an acquisition module for acquiring vehicle operating parameters and track characteristic parameters in real time; a working condition determination module for calculating the total required torque based on the vehicle operating parameters and the track characteristic parameters, and determining whether the vehicle is currently in a driving condition, braking condition, steering condition, or slip condition; a driving condition module for, if the vehicle is determined to be in a driving condition, distributing the total required torque to each wheel using an adaptive particle swarm optimization algorithm based on simulated annealing, with the goal of reducing motor energy consumption; and a braking condition module. The system includes three modules: a braking module and a steering module. The steering module is used to dynamically balance maximizing energy recovery with braking stability using a multi-objective genetic algorithm based on variable weight coefficients, and distribute the total required torque to each wheel. The steering module is used to solve for the optimal additional yaw torque and calculate the torque of each wheel through an underlying torque distribution controller if the vehicle is determined to be in a steering condition. The slippage module is used to independently control the torque of each wheel using an improved sliding mode variable structure control algorithm with a preset slip ratio as the target if the vehicle is determined to be in a slippage condition.

[0090] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part 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 of the various embodiments of the present invention. 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.

[0091] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, database, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0092] The above are merely embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent modifications made based on the content of the present invention's specification and drawings, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A control method for an unmanned concrete transport vehicle facing the wall of a cooling tower, characterized in that the steps include... include: Real-time acquisition of vehicle operating parameters and track characteristic parameters; The total required torque is calculated based on the vehicle operating parameters and the track characteristic parameters, and it is determined whether the vehicle is currently in driving, braking, steering, or slipping condition. If the vehicle is determined to be in driving mode, an adaptive particle swarm optimization algorithm based on simulated annealing is used to distribute the total required torque to each wheel with the goal of reducing motor energy consumption. If the vehicle is determined to be in braking condition, a multi-objective genetic algorithm based on variable weight coefficients is used to dynamically balance maximizing energy recovery and braking stability, and to distribute the total required torque to each wheel. If the vehicle is determined to be in a steering condition, an adaptive model predictive control based on a dynamic model is used to solve for the optimal additional yaw torque and calculate the torque of each wheel through the underlying torque distribution controller. If the vehicle is determined to be in a slip condition, an improved sliding mode variable structure control algorithm is adopted to independently control the torque of each wheel with a preset slip ratio as the target.

2. The unmanned concrete transport vehicle control method for the cooling tower wall as described in claim 1, characterized in that, The step of calculating the total required torque based on the vehicle operating parameters and the track characteristic parameters specifically includes: The vehicle operating parameters include vehicle speed, wheel speed, and wheel load; The track characteristic parameters include track curvature and track slope; Based on the track slope, a longitudinal dynamic model is constructed to calculate the feedforward torque; Based on the error between the vehicle speed and the expected vehicle speed, adaptive PID control is performed according to the wheel load and the track slope to obtain the feedback torque. The total required torque is calculated by combining the feedforward torque and the feedback torque.

3. The unmanned concrete transport vehicle control method for the cooling tower wall as described in claim 2, characterized in that, The steps for determining whether the vehicle is currently in driving, braking, steering, or slipping mode specifically include: When the track curvature is lower than a preset curvature threshold and the total required torque is positive, it is determined to be a driving condition; When the track curvature is lower than a preset curvature threshold and the total required torque is negative, it is determined to be a braking condition. When the track curvature is higher than a preset curvature threshold, it is determined to be a turning condition; The wheel slip ratio is calculated based on the vehicle speed and the wheel rotation speed. When the wheel slip ratio exceeds a preset slip ratio threshold and the duration exceeds a preset duration, it is determined to be a slip condition.

4. The unmanned concrete transport vehicle control method for the cooling tower wall as described in claim 1, characterized in that, The adaptive particle swarm optimization algorithm based on simulated annealing includes: The torque distribution coefficient is initialized as a particle, and the particle swarm size, particle velocity boundary, upper and lower limits of distribution coefficient, inertia weight, learning factor, and simulated annealing initial temperature and decay coefficient are set, and a search space of [0,1] is constructed. A fitness function is constructed with energy consumption minimization as the optimization objective; the torque distribution between the front and rear axles is calculated by inputting allocation coefficients, and the total power loss is obtained. The simplified form of the fitness function is: ； in, The fitness function value; This represents the total drive torque of the front axle; This represents the total drive torque of the rear axle; The front axle single-wheel drive torque is distributed evenly. The rear axle single-wheel drive torque is distributed evenly. For the efficiency of the front axle motor; For the efficiency of the rear axle motor; This refers to the front axle motor speed; This refers to the speed of the rear axle motor. The torque distribution coefficient is updated based on the individual historical best and the group historical best solutions. An annealing mechanism is introduced to linearly reduce the inertia weight based on the number of iterations in the early stage of the iteration to guide global exploration. In the later stage of the iteration, the learning factor is adaptively adjusted using the current temperature value and the probability of poor solution is accepted for local optimization. When the maximum number of iterations or the convergence condition is met, the iteration is terminated, the optimal torque distribution coefficient is output, and the total required torque is distributed to each wheel according to the optimal torque distribution coefficient.

5. The unmanned concrete transport vehicle control method for the cooling tower wall as described in claim 1, characterized in that, The multi-objective genetic algorithm based on variable weight coefficients includes: A multi-objective optimization function is constructed based on energy recovery rate and stability index. The energy recovery rate is calculated by the ratio of the actual recovered energy to the ideal recovered energy during the braking process, and the stability index is calculated by the matching degree between the front and rear axle normal load distribution and the braking torque distribution. A multi-objective genetic algorithm model is constructed by setting the population size, crossover probability, mutation probability, and number of iterations. A variable weight mechanism of fuzzy control is introduced to adaptively adjust the weight coefficients of the energy recovery rate and the braking stability according to the vehicle's operating status. Based on preset physical constraints and preset safety conditions, and by introducing an elite retention strategy, the individual with the highest fitness in the current population is regarded as an elite individual and retained in the next generation population, while the individual with the lowest fitness in the next generation population is eliminated. The front and rear axle torque distribution coefficients are iteratively optimized to generate a target solution set that satisfies the preset constraints. The target solution set is processed offline to generate a query mapping table indexed by vehicle speed and braking intensity, which stores the optimal front and rear axle torque distribution coefficients. When the vehicle brakes, the current vehicle speed and braking intensity are obtained in real time, and the corresponding optimal distribution coefficient is called by querying the mapping table to achieve comprehensive optimization control of braking energy recovery and vehicle stability.

6. The unmanned concrete transport vehicle control method for the cooling tower wall as described in claim 5, characterized in that, The variable weighting mechanism of the fuzzy control specifically includes: Vehicle speed V and braking intensity Z are used as input parameters of the fuzzy controller, and the output is the weight level K1 of energy recovery and the weight level K2 of braking stability. The universe of discourse of vehicle speed V is set to [0, 30], and the fuzzy subset is {VL, VM, VH}. The universe of discourse of braking intensity Z is set to [0, 1], and the fuzzy subset is {ZL, ZM, ZH}. The universe of discourse of weight coefficients is set to [0, 1], and the fuzzy subset is {LL, L, M, H, HH}. The weight levels of LL, L, M, H, and HH increase progressively. When the vehicle is at a low speed VL, the weighting level K1 of braking stability gradually increases from LL to M as the braking intensity Z increases; When the vehicle is at medium speed VM, energy recovery level is prioritized under low braking intensity ZL, energy recovery is maximized under medium braking intensity ZM, and braking stability level is prioritized under high braking intensity ZH. When the vehicle is traveling at high speed (VH), the weighting level of braking stability under low braking intensity (ZL) is increased to M, the weighting level of braking stability under medium braking intensity (ZM) is increased to H, and the weighting level of braking stability under high braking intensity is increased to HH, while the weighting level of energy recovery is reduced to LL.

7. The unmanned concrete transport vehicle control method for the cooling tower wall as described in claim 1, characterized in that, The steps of the adaptive model predictive control based on the dynamic model specifically include: A discrete state-space model of vehicle dynamics is constructed based on vehicle operating parameters and track characteristic parameters. The longitudinal speed, lateral speed, yaw rate, yaw angle, longitudinal position and lateral position in the global coordinate system are used as state variables, and the additional yaw torque is used as the control input. Based on the vehicle dynamics discrete state-space model, predict the state variables in the future time domain, and construct a cost function to comprehensively evaluate trajectory tracking error, yaw stability and control smoothness; Adaptive weight adjustment is performed on the lateral error weight and yaw error weight in the cost function. When the track curvature increases, the lateral error weight and the yaw error weight are increased to improve the curve tracking accuracy and stability. When the track curvature decreases, the lateral error weight is decreased to optimize control smoothness. In each control cycle, a secondary programming is performed based on the adaptively adjusted cost function and by introducing control input constraints and control increment constraints. The optimal additional yaw torque is obtained online, and the torque of each wheel is calculated through the underlying torque distribution controller.

8. The unmanned concrete transport vehicle control method for the cooling tower wall according to claim 7, characterized in that, The adaptive adjustment function is: ； ； in, For curvature; As the reference curvature; This is the saturation value; Baseline horizontal weights; Weighting of baseline yaw rate error; As the baseline yaw angle error weight; This is the horizontal weighting amplification factor; For horizontal error weights; As the weight for the yaw rate error; This is the weight for the yaw angle error; This represents the weighting of the lateral accuracy error.

9. The unmanned concrete transport vehicle control method for the cooling tower wall as described in claim 1, characterized in that, The steps of independently controlling the torque of each wheel using the improved sliding mode variable structure control algorithm with a preset slip ratio as the target specifically include: The slip ratio is calculated based on the wheel speed and vehicle speed, and a sliding surface is constructed based on the slip ratio and the target slip ratio. A sliding control law is then designed to generate a control signal. The calculation formula for the sliding control law is as follows: ； in, The rate of change of the sliding surface over time; For sliding surface functions; This is the gain of the isotropic approaching term; The gain is the exponentially approaching term; This represents the actual slip ratio of the wheel; The target slip ratio, i.e., the preset slip ratio; Based on the control signal, vehicle motion state parameters, and wheel dynamic parameters, the target control torque applied to the hub motor is calculated. The formula for calculating the target control torque is as follows: ； in, To control the torque to the target; The moment of inertia of the wheel; The longitudinal velocity at the center of the wheel; For sliding surface functions; This is the gain of the isotropic approaching term; The gain is the exponentially approaching term; This represents the actual slip ratio of the wheel; The radius of the wheel; It is a saturation function; The longitudinal reaction force of the contact surface to the wheel; The longitudinal acceleration at the center of the wheel; The target control torque An application is made to the corresponding hub motor so that the actual slip ratio of the drive wheel converges to the preset slip ratio.

10. A control system for an unmanned concrete transport vehicle facing the wall of a cooling tower, characterized in that, include: The acquisition module is used to acquire vehicle operating parameters and track characteristic parameters in real time; The working condition determination module is used to calculate the total required torque based on the vehicle operating parameters and the track characteristic parameters, and to determine whether the vehicle is currently in a driving condition, braking condition, steering condition or slip condition. The driving condition module is used to distribute the total required torque to each wheel by adopting an adaptive particle swarm algorithm based on simulated annealing if the vehicle is determined to be in driving condition, with the goal of reducing motor energy consumption. The braking condition module is used to dynamically balance maximizing energy recovery and braking stability by employing a multi-objective genetic algorithm based on variable weight coefficients if the vehicle is determined to be in a braking condition, and to distribute the total required torque to each wheel. The steering condition module is used to determine that the vehicle is in a steering condition, and then adopts adaptive model predictive control based on the dynamic model to solve for the optimal additional yaw torque and calculate the torque of each wheel through the underlying torque distribution controller. The slip condition module is used to independently control the torque of each wheel if the vehicle is determined to be in a slip condition by adopting an improved sliding mode variable structure control algorithm with a preset slip ratio as the target.

Images