An omnidirectional intelligent mobile forklift control system

Abstract

The application relates to the technical field of warehouse logistics equipment, and discloses an omnidirectional intelligent mobile forklift control system. The system comprises a dynamic path planning module, an omnidirectional motion control module, a load stability monitoring module, an environment self-adaptive adjustment module and an intelligent obstacle avoidance decision module. The dynamic path planning module collects multi-dimensional space data of a warehouse environment, and constructs a dynamic environment topological map; the omnidirectional motion control module analyzes driving wheel parameters based on the map, and generates an omnidirectional mobile compound motion instruction set; the load stability monitoring module collects a fork pressure distribution matrix and a cargo gravity center offset amount according to the instruction set, and calculates a stability compensation parameter; the environment self-adaptive adjustment module fuses illumination and slope data, and generates motion control constraint conditions; and the intelligent obstacle avoidance decision module predicts a dynamic obstacle trajectory, outputs an obstacle avoidance path re-planning instruction and updates the map. The system can meet the demand of modern intelligent warehousing.

Description

Technical Field

[0001] This invention relates to the field of warehousing and logistics equipment technology, specifically to an omnidirectional intelligent mobile forklift control system. Background Technology

[0002] In modern warehousing and logistics systems, the efficiency and safety of cargo handling directly impact the overall supply chain's operational efficiency. Forklifts, as core handling equipment, play a crucial role in the quality of warehousing operations. Currently, most mainstream forklift control systems on the market employ fixed-path planning. When building environmental maps, these systems typically only collect two-dimensional coordinates of the racks and static obstacle locations, making it difficult to handle dynamic changes in the warehousing environment, such as the temporary appearance of personnel or handling equipment, or minor adjustments to rack positions. When these dynamic factors occur, forklifts with fixed-path planning often require manual intervention to adjust their routes, impacting operational efficiency and potentially leading to safety hazards due to human error.

[0003] Current forklift motion control systems mostly employ a single drive mode, making omnidirectional movement difficult. Turning and reversing are complex in narrow warehouse aisles, requiring significant operating space. This has become a major factor limiting warehouse operational efficiency, especially given the increasing demand for efficient warehouse space utilization. Furthermore, existing systems have simplistic monitoring of load stability during cargo handling, typically relying solely on fork lifting height, neglecting factors such as cargo center of gravity shifts and uneven pressure distribution on the load-bearing surface. When forklifts encounter uneven ground, changes in slope, or accelerate / decelerate, the cargo center of gravity can easily shift, leading to minor issues like cargo tilting and damage, or even serious problems like cargo falling and threatening personnel and equipment safety.

[0004] Current forklift control systems have weak environmental adaptability and cannot adjust motion control strategies based on environmental parameters such as light intensity, ground friction coefficient, and slope in different storage areas. For example, in dimly lit storage areas, the forklift's sensors may experience reduced data acquisition accuracy due to insufficient light; in areas with low ground friction coefficients, forklifts may slip if they continue to travel at normal speeds; and in sloping areas, failure to adjust torque distribution may result in insufficient power or excessive energy consumption. These problems make existing forklift control systems unable to meet the demands of modern intelligent warehousing in terms of operational efficiency, safety, and adaptability. There is an urgent need for an intelligent mobile forklift control system capable of handling dynamic environments, achieving omnidirectional movement, ensuring stable loads, and possessing environmental adaptability. Summary of the Invention

[0005] The purpose of this invention is to provide an omnidirectional intelligent mobile forklift control system to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides an omnidirectional intelligent mobile forklift control system, the system comprising:

[0007] The dynamic path planning module acquires multi-dimensional spatial data of the warehouse environment and constructs a dynamic environment topology map by collecting real-time three-dimensional coordinates of shelves, point clouds of obstacle outlines, and ground friction coefficients.

[0008] Based on the dynamic environment topology map, the omnidirectional motion control module analyzes the independent steering angle and torque distribution ratio of each drive wheel of the forklift chassis and generates a composite motion command set for omnidirectional movement.

[0009] The load stability monitoring module collects the pressure distribution matrix of the fork bearing surface and the offset of the center of gravity of the cargo in real time according to the composite motion instruction set, and calculates the stability compensation parameters during the dynamic handling process.

[0010] Based on the stability compensation parameters, the environmental adaptive adjustment module integrates the current ambient light intensity and ground slope data to generate motion control constraints for different storage areas.

[0011] The intelligent obstacle avoidance decision module performs probability prediction of the motion trajectory of dynamic obstacles based on the motion control constraints, outputs obstacle avoidance path replanning instructions, and updates the dynamic environment topology map.

[0012] Preferably, the dynamic path planning module includes:

[0013] The spatial data acquisition submodule acquires the geometric feature point cloud of the shelf uprights simultaneously through LiDAR and depth camera, and extracts the calibration parameters of shelf height and aisle width;

[0014] The topology modeling submodule converts the point cloud data into a rasterized node network based on the calibration parameters, and marks the topological connection relationship between passable areas and fixed obstacles.

[0015] The path optimization submodule uses a bidirectional search algorithm to calculate the shortest path node sequence from the starting point to the target point based on the gridded node network, generates an initial navigation path, and writes it into the dynamic environment topology map.

[0016] Preferably, the omnidirectional motion control module includes:

[0017] The kinematic analysis submodule decomposes the coupling relationship between the four-wheel independent steering angle and the drive wheel speed of the forklift chassis based on the curvature change points of the initial navigation path;

[0018] Based on the coupling relationship, the torque distribution submodule calculates the torque distribution weight of each drive wheel during the acceleration phase and generates the composite motion command set that includes the steering angle increment and torque adjustment threshold.

[0019] Preferably, the load stability monitoring module includes:

[0020] The pressure sensing submodule acquires the real-time pressure values ​​of each monitoring point on the fork bearing surface through an array of pressure sensors, and constructs a two-dimensional pressure distribution matrix.

[0021] The center of gravity calculation submodule calculates the lateral and longitudinal offsets of the cargo's center of gravity relative to the fork center axis based on the two-dimensional pressure distribution matrix and the three-dimensional scanning data of the cargo's outline.

[0022] The compensation generation submodule derives the overturning moment compensation coefficient of the forklift when turning based on the lateral and longitudinal offsets, and outputs the stability compensation parameters.

[0023] Preferably, the environment adaptive adjustment module includes:

[0024] The illumination analysis submodule collects intensity distribution data of the ceiling light source in the storage area and identifies the navigation and positioning error correction amount in low-illuminance areas;

[0025] The slope compensation submodule obtains the pitch and roll angles of the ground slope through the inertial measurement unit and calculates the influence factor of gravity on cargo slippage.

[0026] The constraint generation submodule integrates the navigation and positioning error correction amount and the influence factor to generate the motion control constraints that limit the maximum omnidirectional movement speed and steering angle rate.

[0027] Preferably, the intelligent obstacle avoidance decision-making module includes:

[0028] The trajectory prediction submodule identifies the motion direction and velocity vector of dynamic obstacles through a multi-target tracking algorithm, and predicts the collision probability distribution within a future time window.

[0029] The path update submodule marks high-risk areas in the dynamic environment topology map based on the collision probability distribution and replans the obstacle avoidance path node sequence, and outputs the obstacle avoidance path replanning instruction.

[0030] Preferably, the system further includes:

[0031] The shelf recognition calibration module scans the texture features of the shelf identification signs with the laser radar and matches them with the target shelf location data in the pre-stored shelf number database;

[0032] The dynamic path planning module corrects the endpoint coordinates of the initial navigation path based on the target shelf location data.

[0033] Preferably, the system further includes:

[0034] The battery management module monitors the remaining charge and output current fluctuations of the forklift's power battery in real time and predicts the sustainable operating time within the current task cycle.

[0035] The dynamic path planning module filters accessible shelf operation areas in the dynamic environment topology map based on the sustainable operating time.

[0036] Preferably, the system further includes:

[0037] The fault diagnosis module collects temperature and vibration spectrum data of each drive motor and identifies fault characteristic patterns of abnormally worn parts;

[0038] The omnidirectional motion control module dynamically disables abnormal drive wheels and adjusts the torque distribution strategy of the composite motion command set based on the fault characteristic mode.

[0039] Preferably, the system further includes:

[0040] The wireless communication module receives multi-forklift collaborative task instruction sets issued by the central dispatch system;

[0041] The intelligent obstacle avoidance decision-making module marks the real-time location and planned path of other forklifts in the dynamic environment topology map according to the multi-forklift collaborative task instruction set.

[0042] Compared with the prior art, the beneficial effects of the present invention are:

[0043] This omnidirectional intelligent mobile forklift control system acquires multi-dimensional spatial data of the warehouse environment through a dynamic path planning module. It collects the three-dimensional coordinates of the racks, the point cloud of obstacle outlines, and the ground friction coefficient in real time and constructs a dynamic environmental topology map. It can accurately capture dynamic changes in the warehouse environment. Whether it is a temporary obstacle or a slight adjustment of the rack position, it can be perceived in time and integrated into the map update. It can achieve dynamic path adaptation without manual intervention, allowing the forklift to always maintain a reasonable driving path in the complex and ever-changing warehouse environment, reducing operation interruptions caused by environmental changes and improving the overall warehouse handling efficiency.

[0044] The omnidirectional motion control module analyzes the independent steering angle and torque distribution ratio of each drive wheel of the forklift chassis based on a dynamic environment topology map, generating a composite motion command set for omnidirectional movement. This breaks through the limitations of the traditional single drive mode of forklifts, enabling the forklift to achieve flexible omnidirectional movement. Whether it is lateral movement, rotation in place, or diagonal travel in narrow warehouse aisles, it can be easily completed. This significantly reduces the forklift's operating space requirements and effectively improves the utilization rate of warehouse space. Especially in high-density storage scenarios, it can significantly reduce the time and space required for forklifts to turn and adjust, further optimizing the work process.

[0045] The load stability monitoring module, based on a composite motion command set, collects the pressure distribution matrix of the fork bearing surface and the offset of the cargo's center of gravity in real time. It calculates stability compensation parameters during dynamic handling, monitoring the load status from multiple dimensions, moving beyond traditional fork height detection to accurately capture potential risks such as cargo center of gravity offset and uneven pressure on the bearing surface. The calculated stability compensation parameters allow for timely adjustments to the forklift's motion. For example, when cargo center of gravity offset is detected, the travel speed, steering angle, or torque distribution can be adjusted appropriately to prevent cargo tilting or falling during handling, ensuring safety, reducing economic losses from cargo damage, and mitigating personnel and equipment safety risks caused by falling cargo.

[0046] The environmental adaptive adjustment module, based on stability compensation parameters, integrates current ambient light intensity and ground slope data to generate motion control constraints for different storage areas, giving the forklift strong environmental adaptability. In areas with insufficient light intensity, the accuracy of data acquisition can be ensured by adjusting sensor sensitivity or auxiliary lighting strategies. In areas with low ground friction coefficient, slippage can be avoided by constraining travel speed and optimizing torque distribution. In areas with slopes, the torque distribution of the drive wheels can be adjusted according to slope data to ensure that the forklift has sufficient power when climbing and maintains stable travel when descending, avoiding problems such as excessive energy consumption or insufficient power. This allows the forklift to maintain a stable and efficient operating state in storage areas with different environmental conditions.

[0047] The intelligent obstacle avoidance decision module predicts the trajectory of dynamic obstacles based on motion control constraints, outputs obstacle avoidance path replanning instructions, and updates the dynamic environment topology map, further enhancing the forklift's active obstacle avoidance capabilities. This module not only identifies dynamic obstacles but also predicts their trajectories, anticipates potential collision risks, and promptly outputs obstacle avoidance path replanning instructions. This ensures the forklift can quickly and safely adjust its path when encountering dynamic obstacles, avoiding collisions. Simultaneously, the obstacle avoidance path replanning instructions also update the dynamic environment topology map, providing accurate environmental data for subsequent path planning, creating a virtuous cycle and further improving the safety and efficiency of forklift operations. Attached Figure Description

[0048] Figure 1 This is a timing diagram of the omnidirectional intelligent mobile forklift control system described in this invention;

[0049] Figure 2 A schematic diagram illustrating the working principle of the dynamic path planning module;

[0050] Figure 3 This is a schematic diagram of the working principle of the omnidirectional motion control module. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] Please see Figure 1 This invention provides an omnidirectional intelligent mobile forklift control system. The system comprises multiple modules working collaboratively. A dynamic path planning module uses LiDAR, a depth camera, and a friction coefficient sensor to collect real-time multi-dimensional spatial data from the warehouse environment, including the three-dimensional coordinates of the shelves, obstacle outlines, point clouds, and ground friction coefficients. This data, after filtering and fusion, is used to construct a dynamic environmental topology map, which represents the connection between passable areas and fixed obstacles in a rasterized form. An omnidirectional motion control module receives the dynamic environmental topology map as input, analyzes the independent steering angles and torque distribution ratios of each drive wheel of the forklift chassis, and generates a composite motion command set containing steering sequences and speed curves through a kinematic model. A load stability monitoring module triggers an array of pressure sensors on the fork bearing surface based on the composite motion command set, collects the pressure distribution matrix, and calculates the cargo center of gravity offset using three-dimensional scanning data, thereby deriving stability compensation parameters during dynamic handling. An environmental adaptive adjustment module integrates a light sensor and an inertial measurement unit to acquire ambient light intensity and ground slope data, and generates motion control constraints for low-light or sloping areas after fusing the stability compensation parameters. The intelligent obstacle avoidance decision module is based on motion control constraints. It uses a multi-target tracking algorithm to predict the trajectory probability of dynamic obstacles, outputs obstacle avoidance path replanning instructions, and updates the dynamic environment topology map in real time to form a closed-loop control.

[0053] Example 1: See Figure 2The LiDAR and depth camera are integrated into the forklift's protective bracket, simultaneously initiating the data acquisition process in a hardware-synchronized manner. The LiDAR employs a rotating mechanical structure, emitting a laser beam that continuously scans 360 degrees horizontally and covers a certain elevation angle range vertically through a multi-beam layout, thereby capturing the 3D point cloud data of the rack uprights. The depth camera, in conjunction with LEDs, projects a speckle pattern and calculates the depth value of each pixel using the principle of binocular vision. The data from both sensors are aligned using timestamps and then sent to the preprocessing unit. The point cloud data preprocessing includes denoising, registration, and segmentation steps. The denoising algorithm uses a statistical outlier removal method to filter out interference points caused by dust or flying insects. The registration process aligns the LiDAR point cloud and the depth camera point cloud using an iterative nearest-point algorithm, forming a dense point cloud model in a unified coordinate system. The segmentation algorithm uses region growing technology to extract the continuous surface of the rack uprights, fitting geometric features such as edge lines and corner points from them. The calibration parameters for shelf height and aisle width are automatically calculated from the segmented point cloud. Shelf height is obtained by detecting the height difference of the point cloud clusters of adjacent shelf beams, while aisle width is determined by analyzing the minimum passable distance between the point clouds of the shelf uprights on both sides. These parameters are recorded as metadata and stored together with the point cloud. The topology modeling submodule receives the preprocessed point cloud data, first projects the 3D point cloud onto a 2D horizontal plane, and generates an occupied grid map. The size of each grid is set to 0.1 meters square based on the forklift body size and safety margin. The grid value is marked as idle, occupied, or unknown based on the point cloud density.

[0054] The rasterized map is converted into a node network graph, with the center point of each free grid cell serving as a node. Connections between nodes are established based on eight-neighborhood or four-neighborhood rules, forming the vertices and edges of the graph structure. Grids corresponding to fixed obstacles such as shelf bases or walls are marked as impassable nodes, and their connecting edges are removed. The calculation of topological connections uses graph traversal algorithms, such as breadth-first search, to verify reachability between nodes and to specially mark dead ends or narrow passages, generating a map representation in the form of an adjacency matrix or adjacency list. This network graph is dynamically updated to reflect environmental changes. The path optimization submodule, based on the constructed rasterized node network, initializes the node numbers corresponding to the start and target points. A bidirectional search algorithm simultaneously searches from both the start and target nodes, maintaining open and closed lists for each search direction. The open list stores nodes to be expanded, while the closed list records visited nodes. During the search, each node evaluates its heuristic cost to the start and target points, with the cost function considering Euclidean distance and a turning penalty factor. When the open lists of the two search directions intersect, the algorithm terminates and backtracks the path. The generated node sequence is converted into a continuous coordinate point sequence to form an initial navigation path. This path includes location points and recommended orientation, and is written into the path layer of the dynamic environment topology map for use by other modules.

[0055] The LiDAR scanning frequency of the spatial data acquisition submodule is set to 10 Hz, the point cloud density can be configured to tens of thousands of points per frame, the depth camera resolution is maintained at 1920x1080 pixels, the exposure time adapts to changes in ambient light intensity, and the point cloud registration uses the quaternion method to calculate the rotation and translation matrix. During calibration parameter extraction, the shelf height is calculated by determining the peak range through histogram analysis of the vertical direction of the point cloud; the passage width uses the convex hull algorithm to detect the boundary point clouds of obstacles on both sides, and then calculates the nearest pair distance; all parameters are smoothed using Kalman filtering to reduce jitter. The topology modeling submodule uses a sliding window method to update the local map in real time during rasterization. Each node is assigned a unique identifier during node network construction, and the weight of connecting edges is adjusted according to the actual distance between grids and the difficulty of passage, such as increasing the weight in slope areas. Topology connectivity verification uses the Floyd-Warshall algorithm to calculate the shortest path for all node pairs, ensuring global reachability analysis; fixed obstacle labeling combined with historical data avoids misjudgment. When implementing the bidirectional search algorithm in the path optimization submodule, the heuristic function uses Manhattan distance to accelerate convergence, the steering penalty factor is set according to the minimum turning radius of the forklift, and B-spline interpolation is used to smooth the node sequence during path backtracking to eliminate jagged paths.

[0056] The initial navigation path's coordinate point sequence is appended with timestamps and confidence information. The dynamic environment topology map update mechanism employs a differential update strategy, modifying only changed areas to reduce computational load. The spatial data acquisition submodule achieves microsecond-level precision through GPS clock signals for hardware synchronization. Point cloud denoising uses radius filtering to remove isolated points, and the registration algorithm employs feature point matching to improve efficiency. Geometric feature extraction utilizes principal component analysis to calculate the point cloud's normal vector and curvature. The calibration parameter calculation results are stored in a circular buffer, supporting multi-frame data fusion to improve robustness. The topology modeling submodule dynamically adjusts the grid size based on environmental complexity. Virtual nodes are added to the node network graph to handle dynamic obstacles. Topological connectivity uses a disjoint-set data structure to maintain connected components, ensuring the integrity of the path search. The path optimization submodule's search algorithm automatically inserts intermediate nodes in sparse path point regions to avoid the difficulty of controlling long straight paths. After path generation, collision detection is performed for iterative verification to ensure safety and feasibility.

[0057] Example 2: See Figure 3The kinematics analysis submodule receives the initial navigation path from the path planning module. This path consists of a series of continuous path points with curvature information. The analysis process first discretizes the path, extracting key path points at forklift control cycles. Each key point contains global coordinates, orientation angle, and path curvature data. The forklift chassis uses a Mecanum wheel structure with four-wheel independent steering and drive. The kinematic parameters of each wheel are calculated by establishing a transformation relationship between the vehicle coordinate system and the global coordinate system. The analysis algorithm is based on the principle of rigid body kinematics, decomposing the overall motion velocity vector of the vehicle body into the linear velocity and steering angle requirements of each wheel. The handling of curvature change points is particularly critical; in high-curvature sections, the differential speed and steering angle change sequence of each wheel needs to be calculated in advance. The coupling relationship analysis considers the motion constraints between wheel sets, describing the mapping relationship between vehicle body speed and wheel speed through a Jacobian matrix. The matrix elements contain the geometric parameters of the wheel mounting position and the instantaneous value of the steering angle. The analysis results generate the theoretical steering angle and speed reference values ​​of each drive wheel at each point on the path. The torque distribution submodule performs dynamic weight calculation based on the coupling relationship output by kinematic analysis. The weight distribution is based on the estimated wheel-ground adhesion coefficient and the actual load on each wheel. The load information is determined by combining the vehicle tilt angle sensor and air spring pressure data.

[0058] The torque distribution during acceleration employs a feedforward and feedback control structure. The feedforward part calculates the basic torque requirement based on acceleration commands and wheel inertia, while the feedback part corrects the torque output through a closed-loop motor speed adjustment. The distribution algorithm ensures that the slip ratio of each wheel remains within the optimal range. The generation process of the composite motion command set integrates steering angle increments and torque adjustment thresholds. The steering angle increment is represented as a differential expression of the angle change between adjacent control cycles. The torque adjustment threshold is set as a dynamic calculation of the upper limit of torque based on the motor thermal model and battery discharge capacity. The command set is sent to each wheel hub motor controller via a fixed period through the CAN bus. The pressure sensing submodule of the load stability monitoring module manages 32 pressure sensing units on the fork bearing surface through an embedded system. The sensing units are arranged in a matrix covering the main bearing area of ​​the fork. Each unit includes a temperature-compensated strain gauge and signal conditioning circuitry. The data acquisition cycle is synchronized with the motion control cycle. The construction of the two-dimensional pressure distribution matrix uses an interpolation algorithm to enhance spatial resolution. The matrix element values ​​are normalized to represent the relative pressure intensity of each region. The matrix data is updated in real time and includes a timestamp.

[0059] The center of gravity calculation submodule synchronously acquires 3D scan data of the cargo's outline, collected by a 3D TOF camera mounted on the mast. The point cloud data is processed using plane fitting and contour extraction algorithms to obtain the cargo's circumscribed cube model. The center of gravity offset calculation combines the pressure distribution matrix with the cargo's geometric model, employing a multi-sensor fusion algorithm to weight the pressure distribution center of gravity and the geometric center, ultimately outputting the lateral and longitudinal offsets of the cargo's center of gravity relative to the fork centerline. These offsets are represented as vectors in terms of magnitude and direction. After the center of gravity offset is input into the compensation generation submodule, a dynamic model of the forklift-cargo system is first established, including vehicle mass parameters, cargo mass parameters, and suspension system characteristics. The overturning moment calculation considers the coupling effect of centrifugal force generated by steering acceleration and cargo gravity. The compensation coefficient is derived based on the moment balance equation, comparing the theoretical stability boundary with actual measured values ​​to generate compensation parameters. These parameters include hydraulic system pressure adjustment values, speed limit suggestions, and steering angular rate correction values. The stability compensation parameters are output in a structured data package format, containing the compensation action time window, parameter priority flags, and failure protection conditions. The kinematic analysis submodule uses differential geometry to calculate path curvature, with the curvature radius threshold set in relation to the forklift's minimum turning radius. The analytical algorithm automatically inserts transition curves for sharp bends to ensure smooth motion. The torque distribution submodule incorporates an adaptive algorithm for weight calculation, adjusting the distribution ratio in real-time based on motor current feedback. Dual threshold values ​​are set for torque adjustment to prevent overload. The pressure sensing submodule's signal conditioning circuit includes an anti-aliasing filter and a programmable gain amplifier; sampled values ​​are digitally filtered to eliminate mechanical vibration interference. The center of gravity calculation submodule's fusion algorithm uses a Kalman filter to reduce measurement noise, and spatiotemporal registration of the geometric model and pressure data is achieved through hardware synchronization signals.

[0060] The dynamic model parameters of the compensation generation submodule can be calibrated on-site via a calibration program. Before outputting the compensation parameters, a rate-of-change limitation process is applied to prevent abrupt changes. Data exchange between the omnidirectional motion control module and the load stability monitoring module is achieved through shared memory, with a unified control cycle of 10 milliseconds. A heartbeat detection mechanism is established between modules to ensure communication reliability. Verification of the composite motion command set is performed using hardware-in-the-loop testing in a simulation environment, with an instruction caching mechanism ensuring control continuity even with network latency. The pressure sensing matrix is ​​calibrated using standard weight blocks for linearity calibration, and temperature drift compensation coefficients are stored in non-volatile memory. The calculation of the center of gravity offset incorporates confidence level assessment; at low confidence levels, a conservative control strategy is automatically switched. Stability compensation parameters are integrated with the vehicle stability control system to form a multi-layered safety protection architecture. The kinematic analysis submodule supports smooth transition processing during path replanning, automatically calculating interpolated trajectories after path point updates to avoid abrupt motion changes. The slip ratio control of the torque distribution submodule uses a model predictive control algorithm to predict changes in wheel-ground adhesion in advance. The fault detection function of the pressure sensing matrix can identify individual sensor failures and reconstruct the pressure distribution using data from adjacent sensors. The center of gravity calculation submodule optimizes the point cloud density in its 3D point cloud processing, reducing computational load while maintaining accuracy. The overturning moment calculation in the compensation generation submodule incorporates a road slope compensation factor to improve stability control accuracy under slope conditions.

[0061] Example 3: The illumination analysis submodule of the environmental adaptive adjustment module continuously collects ambient light data through a light intensity sensor array installed on the top of the forklift. This array consists of eight photosensitive elements arranged in a ring, with each element having a sampling frequency of 20 Hz. The data is transmitted to the signal processor after analog-to-digital conversion. The processor uses a Gaussian filtering algorithm to smooth the original illuminance value to eliminate instantaneous fluctuations. The identification of low-illuminance areas is based on a preset threshold comparison. When the average illuminance is below 50 lux, it is marked as a low-illuminance area. The calculation of the navigation and positioning error correction amount fuses the measurement deviations of the visual sensor and the lidar. The positioning drift caused by the illuminance change is estimated through a Kalman filter. The correction amount is adjusted in real time in differential form to adjust the reference input of the path tracking controller. The slope compensation submodule relies on a high-precision inertial measurement unit (IMU) to monitor changes in ground tilt. The IMU includes a three-axis accelerometer and gyroscope, outputting pitch and roll angle data at a 200 Hz frequency. This data is fused with wheel speed sensor information using a complementary filtering algorithm to calculate the vehicle's tilt angle relative to the horizontal plane. The influence factor of gravity on cargo slippage is derived using a physical model, considering the static friction coefficient between the cargo and fork surfaces and the load weight. The influence factor calculation is combined with real-time dynamic parameters to dynamically assess the stability risk of the cargo on the slope. The constraint generation submodule integrates the navigation and positioning error correction output from the illumination analysis submodule with the influence factors provided by the slope compensation submodule. A fuzzy logic control system is used for multi-source information fusion. The input variables of the fuzzy system include illumination error, slope angle, and load weight, while the output variables are the limits for omnidirectional maximum speed and steering angular rate. These constraints are adaptively adjusted according to environmental changes, forming a set of motion control constraints.

[0062] The trajectory prediction submodule of the intelligent obstacle avoidance decision-making module processes dynamic obstacle data from sensors using a multi-object tracking algorithm. The tracking algorithm is based on a detection-association framework. First, it uses a YOLOv4 model to identify moving objects in the image, such as people or other forklifts. Then, it uses Kalman filtering to predict the motion direction and velocity vector of each obstacle. The prediction time window is set to 5 seconds, and a sequence of future trajectory points is generated at 0.1-second intervals. The collision probability distribution is calculated using a logistic regression model, where the collision probability... Defined by the following formula:

[0063]

[0064] in: This represents the collision probability, with a value ranging from 0 to 1. It is the sensitivity coefficient, which controls the steepness of the probability curve. It is the real-time distance between the forklift and the obstacle. This is a distance threshold that is dynamically adjusted based on forklift speed and obstacle type. The path update submodule monitors the collision probability distribution in real time. When the probability exceeds 0.7, it is marked as a high-risk area and marked with red grids in the dynamic environment topology map. Path replanning uses the RRT* algorithm, which randomly expands the tree structure from the current path point until a new path node sequence that bypasses the high-risk area is found. The new path is smoothed using B-spline curves to ensure executability, and finally, the obstacle avoidance path replanning command is output to the control system.

[0065] The illumination analysis submodule calibrates the light intensity sensor array periodically using a standard light source. The response curve of each element is stored in non-volatile memory. The illuminance threshold is dynamically adjusted according to warehouse operation time; for example, the threshold is reduced to 30 lux during nighttime operations to adapt to low ambient light conditions. The application of navigation and positioning error correction adopts a gradual adjustment strategy to avoid abrupt changes that could lead to control instability. The correction data is strictly synchronized with the timestamp of the positioning system. The slope compensation submodule's inertial measurement unit data fusion algorithm incorporates wheel speedometer mileage information to correct integral drift. The pitch and roll angle calculation results are low-pass filtered to remove high-frequency noise. The gravity influence factor is calculated in real time with reference to fork tilt sensor data to ensure that the model matches actual operating conditions. The constraint generation submodule's fuzzy logic control system rule base contains dozens of empirical rules, such as "significantly reduce speed if illuminance is low and slope is steep." The defuzzification method uses the center of gravity method to output accurate constraint values. The trajectory prediction submodule's multi-target tracking algorithm supports obstacle ID maintenance, handles occlusion through appearance feature matching, and uses a Kalman filter to generate a state vector including position, velocity, and acceleration. The confidence level of the predicted trajectory is evaluated based on historical tracking accuracy. The path update submodule's RRT* algorithm adapts the number of sampling points to environmental complexity, increasing sampling density in narrow passages. Path smoothing considers the minimum turning radius constraint of forklifts to ensure physical feasibility. Data exchange between modules is achieved via a high-speed bus with a 10-millisecond cycle to ensure real-time performance.

[0066] In the in-depth implementation of the illumination analysis submodule, the layout optimization of the light intensity sensor array covers a 180-degree field of view in the forklift's forward direction. The spectral response of each sensor element matches the characteristics of artificial light sources. The data acquisition circuit includes a programmable gain amplifier to adapt to different illumination ranges. The Gaussian filter window size is dynamically adjusted according to the forklift speed, decreasing during high-speed movement to improve response speed. The navigation and positioning error correction model introduces an ambient light color temperature compensation factor because different light source color temperatures affect the exposure parameters of the visual sensor. The correction calculation is combined with inertial navigation system data to improve robustness. The inertial measurement unit installation position of the slope compensation submodule is dynamically balanced to reduce the impact of vehicle vibration on angle measurement. The pitch and roll angles are calculated using quaternions to avoid universal joint lock issues. The calculation of the gravity influence factor integrates fork hydraulic pressure data to estimate the friction state between the cargo and the forks in real time. The influence factor output is a dimensionless value ranging from 0 to 1, representing the slip risk level.

[0067] The fuzzy logic system input variables of the constraint generation submodule are normalized, illumination error is mapped to the [0,1] interval, slope angle is converted to angle value, and defuzzified output constraint values ​​are amplitude-limited to prevent extreme cases. The motion control constraint set is sent to the chassis controller via the CAN bus, with the instruction format including timestamps and priority flags. The YOLOv4 model training of the trajectory prediction submodule uses a warehouse scenario dataset to optimize the detection accuracy of forklifts and personnel, and the process noise and observation noise parameters of the Kalman filter are adaptively adjusted online. Collision probability distribution calculation is accelerated using parallel computing to meet real-time requirements. The RRT* algorithm implementation of the path update submodule incorporates heuristic guidance to improve search efficiency, and the path verification stage performs collision detection loops until safety is achieved. The obstacle avoidance path replanning instruction includes path point sequences and control parameters to ensure a smooth transition. The collaboration between the environmental adaptive adjustment module and the intelligent obstacle avoidance decision module is coordinated through a central scheduler, with a module status monitoring heartbeat mechanism that switches to a safe mode in case of anomalies. The entire system implementation emphasizes real-time performance and reliability to meet the complex operational needs of intelligent forklifts. The lighting analysis submodule incorporates an outlier detection algorithm in its data preprocessing to eliminate interference from instantaneous light sources. The slope compensation submodule considers dynamic load changes in its influence factor calculations, such as weight distribution adjustments during cargo swaying. The constraint generation submodule's fuzzy rules support online learning, optimizing rule weights based on historical operation data. The collision probability formula parameters in the trajectory prediction submodule are calibrated using measured data, with the sensitivity coefficient λ set to different values ​​based on obstacle type: λ=0.5 for personnel and λ=0.3 for forklifts. The path update submodule's replanning results undergo stability verification to ensure the new path conforms to dynamic constraints. The inter-module communication protocol uses a custom binary format to reduce transmission latency, and data integrity is guaranteed through CRC checksum verification.

[0068] Example 4: The collaborative workflow of the shelf recognition calibration module and the battery management module is demonstrated through a specific task scenario: Assume a forklift receives an instruction from the central dispatch system to retrieve goods from shelf A-05-12. The shelf recognition calibration module first activates the LiDAR to scan the surrounding environment. The laser beam is emitted at a frequency of 10 Hz and the return signal is received. After voxel filtering and noise reduction, the point cloud data is segmented into possible shelf column structures using a region growing algorithm. When a metal structure with a sign is detected, the scanning focus is automatically adjusted to the sign area, and the point cloud density is increased to twice the standard value to capture richer surface details. The texture feature extraction of the sign uses a local binary mode algorithm, which converts the point cloud intensity value into rotation-invariant texture feature codes. These feature codes are matched against a pre-stored shelf number database. The database is stored in the vehicle's solid-state drive and records the global coordinates, height specifications, and special operational requirements corresponding to each shelf number. The matching process uses an approximate nearest neighbor search algorithm. When the similarity between the feature code and the database record A-05-12 exceeds a threshold, the system determines that the target shelf has been successfully identified. At this point, the module reads the precise coordinates of the shelf stored in the database (X=125.63m, Y=88.47m, Z=5.21m) and uses these coordinates as the new endpoint for path planning. After receiving the target shelf location data, the dynamic path planning module immediately initiates the path correction procedure.

[0069] The correction algorithm first calculates the deviation vector between the current navigation path endpoint and the original target point, and then uses gradient descent to iteratively adjust the path node sequence. During the iteration process, the adjustment amount of each path point is inversely proportional to its distance to the new endpoint; the closer the point, the larger the adjustment, ensuring a smooth path transition. The accuracy of the corrected path endpoint coordinates is controlled within ±2 cm, and the path curvature is recalculated to ensure compliance with forklift kinematic constraints. The entire correction process is completed within 200 milliseconds, achieving seamless path switching. The battery management module's monitoring function is implemented through a high-precision coulomb counter, which samples the total battery current and individual cell voltage in real time at a sampling frequency of 1 kHz. The current data is averaged and filtered through a sliding window, and the remaining power is estimated using a battery temperature compensation model. Output current fluctuation analysis uses Fast Fourier Transform to identify the main harmonic components and their energy distribution. Sustainable operating time prediction is based on a dynamic power consumption model, which considers the forklift's current load weight, changes in the planned path slope, and historical energy consumption data. The prediction algorithm divides the remaining battery power by the average power consumption over the past five minutes to obtain the base running time, and then makes adjustments based on path characteristics: for example, increasing the time consumption coefficient for uphill sections and decreasing the coefficient for downhill sections. See Table 1 for the battery key parameter monitoring records during task execution.

[0070] Table 1: Battery Parameter Monitoring and Running Time Prediction

[0071] Timestamp Remaining battery power (%) Average current (A) Temperature (°C) Load weight (kg) Predict the remaining time (min) 14:30:15 78.3 25.6 28.4 0 126 14:32:40 76.8 43.2 29.1 350 118 14:35:20 75.1 38.9 29.8 350 115

[0072] The dynamic path planning module performs reachability area filtering in a dynamic environmental topology map based on the sustainable operating time data provided by the battery management module. The filtering algorithm first calculates the minimum energy consumption for a round trip from the current location to each racking area. The energy consumption model includes basic movement energy consumption, lifting energy consumption, and additional turning energy consumption. Then, the energy consumption value is converted into time requirements and compared with the predicted remaining time. Only racking areas with time requirements less than 70% of the predicted remaining time are marked as reachable areas. This conservative threshold ensures the forklift has sufficient power to return to the charging station. Reachable area information is overlaid on the environmental map with different color layers for the path planning algorithm to prioritize. The racking recognition calibration module's database update mechanism is designed in dual modes: in online mode, it receives incremental updates from the central database via a wireless network; in offline mode, it relies on locally cached data to maintain operation. Backup solutions for sign recognition failures include rack outline matching and relative position estimation algorithms to improve system robustness. LiDAR scanning parameters are automatically adjusted according to ambient light, increasing transmission power under strong light conditions to ensure point cloud quality. The power consumption model parameters of the battery management module are adaptively optimized through a learning algorithm, recording the deviation between actual energy consumption and predicted values ​​under different operating conditions, and gradually correcting the model coefficients. The temperature compensation model considers battery aging and adjusts the compensation curve after more than 500 charge-discharge cycles. The predicted sustainable operating time is updated every 30 seconds, triggering a system alarm upon significant changes. The reachable area filtering results from the dynamic path planning module influence path decisions in real time; when the predicted remaining time falls below a threshold, the system automatically plans the optimal path to the charging station. The gradient descent method used in the path correction process has a maximum number of iterations to prevent infinite loops, and the convergence condition is based on the quadratic norm of the coordinate error.

[0073] The signage texture feature extraction also incorporates optical character recognition technology as an auxiliary verification method. After spatiotemporal alignment of point cloud data and visual images, a convolutional neural network is used to recognize the numbered characters on the signs. Database queries employ a multi-level caching architecture, storing recently accessed records in memory to accelerate response speed. Coordinate correction values ​​are smoothed using Kalman filtering to avoid frequent path adjustments due to recognition jitter. The current sampling circuit of the battery management module uses a 24-bit high-precision ADC, and the current sensor utilizes the closed-loop Hall effect principle to improve measurement linearity. The slope influence coefficient in the power consumption model is dynamically calculated using data from the inertial measurement unit, reflecting actual road conditions in real time. The prediction algorithm automatically switches to a conservative mode when abnormally high current consumption is detected, extending the remaining prediction time by 20% as a safety margin. The reachability filtering algorithm of the dynamic path planning module incorporates a dynamic obstacle influence factor, appropriately increasing the time budget when moving obstacles exist in the path. The filtering results are processed in conjunction with task priorities, allowing for more relaxed reachability conditions for high-priority tasks. The entire system implementation emphasizes data consistency between modules, ensuring that decisions by each module are based on data from the same time segment through a unified time scale.

[0074] Example 5: Assume that in aisle B2 of the warehouse area, forklift AGV-03 is performing a transport task from the picking station to rack A-05. The temperature sensor of the fault diagnosis module is embedded inside the windings of the forklift's four drive motors, using a PT100 platinum resistance element to collect temperature data at a frequency of 10 times per second. Vibration monitoring is achieved through a triaxial accelerometer mounted on the motor housing, with a sampling frequency set to 1kHz. When AGV-03 passes through a bend in the aisle, the temperature of its right front wheel drive motor rises continuously from the normal operating temperature of 65 degrees Celsius to 82 degrees Celsius within 30 seconds. Simultaneously, the vibration accelerometer detects an abnormal vibration component at a frequency of 87Hz, which matches the bearing fault characteristic frequency. The wavelet transform analysis unit of the fault diagnosis module performs a 5-level decomposition of the vibration signal, extracting a significant impact component in the third level of detail coefficients. The matching degree with the bearing inner ring wear pattern in the pre-stored fault feature library reaches 0.76, indicating that the system determines that the motor has an early fault risk. Upon receiving the fault characteristic pattern, the omnidirectional motion control module immediately initiates the fault-tolerant control program. The control algorithm first marks the abnormal drive wheel as disabled and fixes its steering angle through a mechanical locking device to prevent free rotation from interfering with the vehicle's attitude.

[0075] The torque distribution strategy was adjusted to a three-wheel drive mode, recalculating the torque distribution weights of the remaining three drive wheels. The new distribution scheme considered mass redistribution and steering center offset, employing the principle of torque balance to maintain the driving direction. Restrictions were added to the generation of the composite motion command set, reducing the maximum speed from 1.5 meters per second to 1.0 meter per second, and limiting the steering angle rate to 60% of the normal value. These adjustments were transmitted in real-time to each motor controller via the control area network. During this period, the wireless communication module maintained a continuous 5G connection with the central dispatch system, receiving collaborative task command sets containing multi-forklift location information. The command set was encapsulated in JSON format, containing information such as task priority, path reservation time window, and emergency braking protocol. When AGV-03 entered the intersection of the shelving area, the communication module received a warning that AGV-07 was approaching from the side passage; this vehicle had higher priority and carried fragile items. The intelligent obstacle avoidance decision module immediately activated the collaborative obstacle avoidance mode, marking the real-time position of AGV-07 (coordinates X=135.72, Y=92.45) and its predetermined path point sequence on the dynamic environmental topology map.

[0076] The system employs a time-window conflict detection algorithm to calculate the spatiotemporal overlap probability of the two vehicles' paths in the intersection area, identifying a collision risk within the next 8 seconds. The path update submodule initiates dynamic path replanning, generating an avoidance path for AGV-03 that includes temporary waiting points: first, it pauses at its current position for 3 seconds to allow AGV-07 to pass, then completes the remaining tasks along the corrected path. The node sequence of the replanned path is shared with AGV-07 via V2X communication, forming a collaborative obstacle avoidance scheme. The fault diagnosis module's deep monitoring function continuously enhances data acquisition after detecting anomalies, increasing the temperature sampling frequency to 20Hz and adding envelope spectrum detection to vibration analysis. The severity assessment of bearing faults uses trend analysis, calculating the rate of temperature rise and the rate of vibration energy increase; a level-two alarm is triggered when the indicators exceed the threshold. The three-wheel drive model of the fault-tolerant control system is pre-verified through offline simulation, storing multiple torque distribution schemes for different faulty wheel positions to ensure the smoothness of the switching process. The wireless communication module's data link establishes a dual verification mechanism; each data packet includes a sequence number and a timestamp, and automatically switches to a backup frequency band when the packet loss rate exceeds 5%. The parsing of collaborative task instruction sets adopts a state machine model to verify instruction integrity and execution conditions. Invalid instructions are stored in the log for subsequent analysis.

[0077] The collision detection algorithm in the intelligent obstacle avoidance decision-making module incorporates predictive confidence assessment. When the positioning signal of AGV-07 fluctuates significantly, the detection time window is automatically expanded to 12 seconds to improve safety. In practical implementation, the bearing fault feature library of the fault diagnosis module is constructed through accelerated life testing in the laboratory, containing vibration spectrum templates for different wear stages. The basis function for wavelet transform is the Db4 wavelet, and the number of decomposition levels is adaptively adjusted according to the motor speed. The trend analysis of temperature data uses an exponentially weighted moving average algorithm to eliminate random fluctuation interference. The fault-tolerant control switching process of the omnidirectional motion control module includes a 200-millisecond transition period, during which fuzzy control is used to progressively adjust torque distribution. The path tracking algorithm in three-wheel drive mode incorporates lateral error compensation, offsetting the yaw torque caused by asymmetric driving force by adjusting the rear wheel steering angle. The speed limit parameter is dynamically adjusted according to the severity of the fault, further reducing the speed to 0.8 m / s when the temperature exceeds 85 degrees Celsius. The 5G connection of the wireless communication module uses a dedicated network frequency band, with transmission latency controlled within 50 milliseconds. The message queue of the collaborative task instruction set is configured with priority management, and emergency avoidance instructions can interrupt regular status reporting. The path replanning results of the intelligent obstacle avoidance decision module have been stability verified to ensure that the forklift can still meet dynamic constraints even in a fault state.

[0078] The fault diagnosis module interfaces with the vehicle management system using a publish-subscribe model, with fault events broadcast to relevant subsystems via a message middleware. The omnidirectional motion control module's torque distribution strategy pre-stores solutions for various fault scenarios, including extreme cases such as single-wheel failure and dual-wheel failure. The wireless communication module's signal strength monitoring and path planning are linked, pre-planning conservative paths in signal blind spots. The intelligent obstacle avoidance decision-making module's collaborative algorithm supports a multi-vehicle negotiation mechanism, automatically employing a pre-set obstacle avoidance rule base when simultaneous avoidance by other vehicles is detected. The entire implementation process balances real-time performance, reliability, and safety. Early warnings from fault diagnosis provide the system with fault-tolerant processing time, wireless communication ensures multi-agent collaborative efficiency, and intelligent obstacle avoidance decision-making achieves a balance between safety and efficiency. Data flows between modules are synchronized using a unified timescale, and the execution results of control commands are verified through sensor feedback, forming a closed-loop control system.

[0079] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0080] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An omnidirectional intelligent mobile forklift control system, characterized in that, include: The dynamic path planning module acquires multi-dimensional spatial data of the warehouse environment and constructs a dynamic environment topology map by collecting real-time three-dimensional coordinates of shelves, point clouds of obstacle outlines, and ground friction coefficients. Based on the dynamic environment topology map, the omnidirectional motion control module analyzes the independent steering angle and torque distribution ratio of each drive wheel of the forklift chassis and generates a composite motion command set for omnidirectional movement. The load stability monitoring module collects the pressure distribution matrix of the fork bearing surface and the offset of the center of gravity of the cargo in real time according to the composite motion instruction set, and calculates the stability compensation parameters during the dynamic handling process. Based on the stability compensation parameters, the environmental adaptive adjustment module integrates the current ambient light intensity and ground slope data to generate motion control constraints for different storage areas. The intelligent obstacle avoidance decision module performs probability prediction of the motion trajectory of dynamic obstacles based on the motion control constraints, outputs obstacle avoidance path replanning instructions, and updates the dynamic environment topology map.

2. The omnidirectional intelligent mobile forklift control system according to claim 1, characterized in that, The dynamic path planning module includes: The spatial data acquisition submodule acquires the geometric feature point cloud of the shelf uprights simultaneously through LiDAR and depth camera, and extracts the calibration parameters of shelf height and aisle width; The topology modeling submodule converts the point cloud data into a rasterized node network based on the calibration parameters, and marks the topological connection relationship between passable areas and fixed obstacles. The path optimization submodule uses a bidirectional search algorithm to calculate the shortest path node sequence from the starting point to the target point based on the gridded node network, generates an initial navigation path, and writes it into the dynamic environment topology map.

3. The omnidirectional intelligent mobile forklift control system according to claim 2, characterized in that, The omnidirectional motion control module includes: The kinematic analysis submodule decomposes the coupling relationship between the four-wheel independent steering angle and the drive wheel speed of the forklift chassis based on the curvature change points of the initial navigation path; Based on the coupling relationship, the torque distribution submodule calculates the torque distribution weight of each drive wheel during the acceleration phase and generates the composite motion command set that includes the steering angle increment and torque adjustment threshold.

4. The omnidirectional intelligent mobile forklift control system according to claim 3, characterized in that, The load stability monitoring module includes: The pressure sensing submodule acquires the real-time pressure values ​​of each monitoring point on the fork bearing surface through an array of pressure sensors, and constructs a two-dimensional pressure distribution matrix. The center of gravity calculation submodule calculates the lateral and longitudinal offsets of the cargo's center of gravity relative to the fork center axis based on the two-dimensional pressure distribution matrix and the three-dimensional scanning data of the cargo's outline. The compensation generation submodule derives the overturning moment compensation coefficient of the forklift when turning based on the lateral and longitudinal offsets, and outputs the stability compensation parameters.

5. The omnidirectional intelligent mobile forklift control system according to claim 4, characterized in that, The environmental adaptive adjustment module includes: The illumination analysis submodule collects intensity distribution data of the ceiling light source in the storage area and identifies the navigation and positioning error correction amount in low-illuminance areas; The slope compensation submodule obtains the pitch and roll angles of the ground slope through the inertial measurement unit and calculates the influence factor of gravity on cargo slippage. The constraint generation submodule integrates the navigation and positioning error correction amount and the influence factor to generate the motion control constraints that limit the maximum omnidirectional movement speed and steering angle rate.

6. The omnidirectional intelligent mobile forklift control system according to claim 5, characterized in that, The intelligent obstacle avoidance decision-making module includes: The trajectory prediction submodule identifies the motion direction and velocity vector of dynamic obstacles through a multi-target tracking algorithm, and predicts the collision probability distribution within a future time window. The path update submodule marks high-risk areas in the dynamic environment topology map based on the collision probability distribution and replans the obstacle avoidance path node sequence, and outputs the obstacle avoidance path replanning instruction.

7. The omnidirectional intelligent mobile forklift control system according to claim 6, characterized in that, Also includes: The shelf recognition calibration module scans the texture features of the shelf identification signs with the laser radar and matches them with the target shelf location data in the pre-stored shelf number database; The dynamic path planning module corrects the endpoint coordinates of the initial navigation path based on the target shelf location data.

8. The omnidirectional intelligent mobile forklift control system according to claim 7, characterized in that, Also includes: The battery management module monitors the remaining charge and output current fluctuations of the forklift's power battery in real time and predicts the sustainable operating time within the current task cycle. The dynamic path planning module filters accessible shelf operation areas in the dynamic environment topology map based on the sustainable operating time.

9. The omnidirectional intelligent mobile forklift control system according to claim 8, characterized in that, Also includes: The fault diagnosis module collects temperature and vibration spectrum data of each drive motor and identifies fault characteristic patterns of abnormally worn parts; The omnidirectional motion control module dynamically disables abnormal drive wheels and adjusts the torque distribution strategy of the composite motion command set based on the fault characteristic mode.

10. The omnidirectional intelligent mobile forklift control system according to claim 9, characterized in that, Also includes: The wireless communication module receives multi-forklift collaborative task instruction sets issued by the central dispatch system; The intelligent obstacle avoidance decision-making module marks the real-time location and planned path of other forklifts in the dynamic environment topology map according to the multi-forklift collaborative task instruction set.

Images