Adaptive balancing method for workpieces in complex environment of transport vehicle

Abstract

The invention relates to a workpiece adaptive balancing method for a transport vehicle in a complex environment. The workpiece adaptive balancing method comprises the following steps that the current terrain and load conditions are determined through a multi-sensor fusion system; when the transport vehicle runs on a flat ground, left and right wheels of the transport vehicle are supported symmetrically, and the trunk of the transport vehicle is kept in a horizontal posture; when the transport vehicle is in contact with a slope terrain, the inclination direction and gradient of the terrain are sensed through the multi-sensor fusion system, the transport vehicle wheel feet close to the lower side are driven to extend, the transport vehicle wheel feet on the other side are driven to shorten, and the transport vehicle trunk is kept horizontal; when the foundation is subjected to differential settlement, a four-layer sensing framework is constructed, accurate sensing of the pose of the bearing table is achieved through multi-source information fusion, the pressure distribution of supporting legs of a transport vehicle is dynamically adjusted through an electro-hydraulic proportional control system, and the grounding pressure can be controlled within a foundation bearing threshold value. According to the control strategy of keeping dynamic balance of various road conditions and fusing multiple sensors, the adaptability of the transport vehicle to the environment during heavy-load transportation is improved.

Description

Technical Field

[0001] This application relates to the field of heavy-duty transport vehicle technology, and in particular to an adaptive balancing method for workpieces in complex environments of transport vehicles. Background Technology

[0002] With the continuous development of modern industry, heavy-duty transport vehicles are increasingly widely used in key fields such as infrastructure, energy, chemicals, and large-scale manufacturing. Tens of thousands of challenging transport tasks are required annually for transporting oversized and overweight materials such as wind power equipment, large modules, chemical containers, and heavy machinery. Traditional transportation methods often rely on manual direction, experience-based judgment, and conventional vehicles, facing problems such as insufficient turning radius, poor road adaptability, high safety risks, and low efficiency, making it difficult to meet the overall demands of current large-scale projects for efficient, safe, and intelligent transportation.

[0003] The patent, titled "A Self-Balancing Transport Vehicle, Control Method, and Shield Tunnel Construction Equipment," with publication number CN119705650A, describes a self-balancing transport vehicle comprising a chassis; a drive wheel assembly including a left drive wheel and a right drive wheel respectively disposed on the left and right sides of the chassis; a left drive component mounted on the chassis for driving the left drive wheel; a right drive component mounted on the chassis for driving the right drive wheel; a gyroscope sensor mounted on the chassis for detecting the chassis's lateral tilt angle ratio; and a control device configured to calculate adjustment parameters based on information from the gyroscope sensor, and adjust the output power of at least one of the left or right drive components according to these parameters. By utilizing the gyroscope sensor to detect the tilt angle information of the transport vehicle, the output power of at least one of the left or right drive components is controlled during turning and curved route transport, thereby ensuring that the transport vehicle can automatically correct its course to prevent overturning, thus securing the workpiece on the transport vehicle and significantly improving transport safety.

[0004] Existing heavy-duty transport vehicles have several shortcomings. First, they lack the ability to adapt to slopes of varying heights, making it highly susceptible to slippage of heavy loads. This can lead to unstable workpiece fixation or poor compatibility, affecting transport safety and accuracy. Second, they lack adaptability to different foundations. They lack specific adjustment mechanisms for complex foundations such as soft soil and frozen soil, making it difficult to control ground pressure, settlement rate, and vibration transmission. This can easily cause imbalance due to uneven foundation conditions. Third, they have weak sensing and fault tolerance capabilities. Sensor failures can easily lead to instability, making it difficult to cope with complex transport scenarios.

[0005] Therefore, we propose an adaptive balancing method for workpieces in complex environments of transport vehicles.

[0006] Application content To address the shortcomings of existing production technologies, the applicant provides an adaptive balancing method for workpieces in complex environments for transport vehicles, a control strategy that maintains dynamic balance under various road conditions and integrates multiple sensors, thereby increasing the adaptability of transport vehicles to the environment during heavy-load transport.

[0007] The technical solution adopted in this application is as follows: An adaptive balancing method for workpieces in complex environments of a transport vehicle includes: The ground slope and foundation type are perceived in real time through a multi-sensor fusion system to determine the current terrain and load conditions; When driving on flat ground, the left and right wheels of the transport vehicle are symmetrically supported, the body of the transport vehicle remains horizontal, the wheel-ground contact points are evenly distributed, and the center of gravity of the transport vehicle is located in the center of the support area, ensuring the balance on flat ground. When encountering sloping terrain, the multi-sensor fusion system senses the direction and slope of the terrain, drives the wheel of the transport vehicle on the side closer to the lower part to extend, and drives the wheel of the transport vehicle on the other side to shorten, so that the body of the transport vehicle remains horizontal. When uneven ground settlement occurs, a four-layer sensing architecture is constructed. Through multi-source information fusion, precise perception of the platform's position and orientation is achieved. The data layer uses a high-precision dual-axis tilt sensor to monitor the roll angle θ and pitch angle φ of the transport vehicle's body in real time. The feature layer integrates a fiber optic pressure array, mapping the hydraulic outrigger pressure distribution using the Bragg wavelength displacement principle, and eliminating local distortion errors through adjacent node data compensation. The decision layer uses a laser displacement sensor combined with a ground settlement prediction model to achieve millimeter-level deformation compensation. The fault-tolerant layer deploys three sets of redundant IMUs using a hardware watchdog + majority voting mechanism, ensuring short switching time and high system reliability in case of main sensor failure. The electro-hydraulic proportional control system dynamically adjusts the outrigger pressure distribution of the transport vehicle, keeping the ground pressure within the ground bearing capacity threshold.

[0008] Its further features are: The multi-sensor fusion system includes laser sensors, GNSS equipment, inertial navigation sensors, and magnetic navigation sensors, which together construct a comprehensive navigation perception network. The laser sensor provides high-precision ranging information, the GNSS equipment achieves global positioning, the inertial navigation sensor ensures stable tracking, and the magnetic navigation sensor assists in positioning.

[0009] The core of the four-layer sensing architecture lies in the development of the IGD-UKF dynamic perturbation suppression algorithm:

[0010] in, This represents the weighting coefficients dynamically assigned based on sensor confidence levels. For the nine-axis attitude transformation function, the regularization coefficient is... The value ranges from 0.05 to 0.2.

[0011] The center of gravity of the transport vehicle is projected to the center of the support area to ensure balance on the flat ground.

[0012] The data layer has a built-in temperature compensation algorithm that ensures minimal measurement error under certain temperature conditions.

[0013] The transport vehicle adopts a layered design, consisting of an upper layer and a lower layer. The lower layer uses a microcontroller to achieve motion control and includes four adjustable wheels and an electro-hydraulic proportional control system. The movement of the wheels is controlled by the electro-hydraulic proportional control system to ensure efficient and precise adaptive workpiece maintenance and transport vehicle movement control.

[0014] The upper layer is controlled by an industrial computer with powerful data processing capabilities. The upper layer includes the torso, on which the workpiece is placed. At the same time, the upper layer is equipped with a multi-sensor fusion system to provide the data support required for the positioning, navigation and path planning of the transport vehicle.

[0015] The laser rangefinder can be replaced by a millimeter-wave radar / binocular vision camera, both of which can work with the IMU to complete terrain perception; the fiber optic pressure array can be replaced by a strain gauge pressure sensor, which can meet the ground pressure monitoring requirements through multi-node data fusion.

[0016] The IGD-UKF algorithm is replaced by a fusion algorithm of extended Kalman filter (EKF) and adaptive fuzzy control. The majority voting mechanism can be replaced by a weighted voting mechanism, which judges the validity of data through weighted calculation. This is suitable for scenarios with large differences in sensor accuracy.

[0017] The four-layer perception architecture is simplified to a two-layer architecture of "data fusion layer - decision control layer". It integrates feature extraction and data fusion, and processes multi-sensor data in real time through the edge computing module. The fault tolerance is reduced, and the system latency is reduced by 10%, which is suitable for short-distance transportation scenarios with higher requirements for response speed.

[0018] The beneficial effects of this application are as follows: This application features a compact and rational structure, and is easy to operate. It uses an inertial navigation module and tilt sensors to collect slope information in real time, driving the wheels on both sides of the slope to achieve differentiated extension and retraction adjustments. Simultaneously, it adapts and adjusts the torso's posture, ensuring precise alignment of the torso's local coordinate system with the slope surface. This guarantees that the wheel-to-ground contact point is always within the effective working space, and the equipment's center of mass projection does not exceed the support boundary. The heavy-duty transport vehicle can stably adapt to multi-gradient slope terrain, significantly reducing the risk of workpiece slippage and fundamentally solving the core technical challenges of heavy-duty transport vehicles easily becoming unbalanced and workpieces easily slipping in slope environments.

[0019] In addition, this application also has the following advantages: (1) Multi-sensor fusion hierarchical perception and fault-tolerant architecture: A hierarchical perception architecture is constructed, consisting of a lower-level microcontroller responsible for motion control and an upper-level industrial control computer integrating lidar, GNSS, IMU, and magnetic navigation. Three redundant IMUs are deployed, and a hardware watchdog + majority voting mechanism is used to achieve sensor fault tolerance. When the GNSS signal is blocked, the IMU can work with the lidar to achieve continuous positioning; when the main sensor fails, it can quickly switch over. The system's mean time between failures (MTBF) is significantly improved compared to existing technologies, effectively ensuring the reliability of perception in complex scenarios such as tunnels and mountainous areas.

[0020] (2) Dynamic Disturbance Suppression and Pose Correction Algorithm Fusion of IGD-UKF: The fast convergence of improved gradient descent (IGD) and the anti-disturbance of unscented Kalman filtering (UKF) are combined to form the IGD-UKF algorithm, which is used for dynamic pose correction of heavy-duty transport vehicles. Compared with traditional Kalman filtering, this algorithm can significantly reduce tilt angle measurement error and greatly shorten the pose correction response time; when faced with sudden slope changes or instantaneous foundation settlement, it can effectively control the tilt angle fluctuation range of the bearing platform, providing accurate algorithmic support for the dynamic balance of the entire workpiece transportation process.

[0021] (3) Multi-dimensional sensing of foundation condition and coordinated control mechanism of hydraulic outriggers: The angle change of the bearing platform is captured by a dual-axis tilt sensor with temperature compensation, and the pressure distribution of the outriggers is mapped by a fiber optic pressure array. Combined with a laser displacement sensor and a settlement prediction model, the pressure and elongation of the outriggers are dynamically adjusted by an electro-hydraulic proportional control system. In this way, the ground pressure can be accurately controlled within the foundation bearing capacity threshold, significantly reducing the foundation settlement rate, effectively attenuating vibration transmission, and effectively solving the problem of heavy-duty transport vehicles easily tilting and becoming unbalanced in complex foundation environments such as soft soil and frozen soil.

[0022] (4) Wheel-to-ground contact point workspace constraint and optimization method: Through the adaptation calculation of the local coordinate system and the global coordinate system of the torso, the effective workspace boundary of the wheel-to-ground contact point is mathematically modeled and constrained to clarify the limit range of wheel extension and retraction. This method can avoid exceeding the mechanical workspace when the wheel is adjusted, and from the geometric constraint level, it helps to improve the stability and safety of adaptive ground height control, which is an important supplement to the core control method. Attached Figure Description

[0023] Figure 1 This is a terrain adaptation strategy diagram for the transport vehicle in this application.

[0024] Figure 2 This is a framework diagram for the adaptive adjustment of the ground height of the transport vehicle in this application.

[0025] Figure 3 This is a structural diagram of the transport vehicle sensor of this application.

[0026] Figure 4This is a schematic diagram of the multi-sensor fusion sensing system of this application. Detailed Implementation

[0027] The specific embodiments of this application are described below with reference to the accompanying drawings.

[0028] like Figures 1-4 As shown, an adaptive balancing method for workpieces in complex environments of a transport vehicle includes the following steps: The ground slope and foundation type are perceived in real time through a multi-sensor fusion system to determine the current terrain and load conditions; When driving on flat ground, the left and right wheels of the transport vehicle are symmetrically supported, the body of the transport vehicle remains horizontal, the wheel-ground contact points are evenly distributed, and the center of gravity of the transport vehicle is located in the center of the support area, ensuring the balance on flat ground. When encountering sloping terrain, the multi-sensor fusion system senses the direction and slope of the terrain, drives the wheel of the transport vehicle on the side closer to the lower part to extend, and drives the wheel of the transport vehicle on the other side to shorten, so that the body of the transport vehicle remains horizontal. like Figure 1 As shown, the terrain adaptation strategy of the transport vehicle is achieved through differential adjustment of wheel length and coordinated adaptation of torso posture. When driving on flat ground, the left and right wheels of the transport vehicle are supported by symmetrical lengths, and the local coordinate system Σ1 and the global coordinate system Σ0 of the torso are aligned. With the z-plane parallel, the torso maintained in a horizontal position, the wheel-to-ground contact points evenly distributed, and the center of mass projection located at the center of the support area, the initial balance of the flat terrain is ensured.

[0029] When uneven ground settlement occurs, a four-layer sensing architecture is constructed. Through multi-source information fusion, precise perception of the platform's position and orientation is achieved. The data layer uses a high-precision dual-axis tilt sensor to monitor the roll angle θ and pitch angle φ of the transport vehicle's body in real time. The feature layer integrates a fiber optic pressure array, mapping the hydraulic outrigger pressure distribution using the Bragg wavelength displacement principle, and eliminating local distortion errors through adjacent node data compensation. The decision layer uses a laser displacement sensor combined with a ground settlement prediction model to achieve millimeter-level deformation compensation. The fault-tolerant layer deploys three sets of redundant IMUs using a hardware watchdog + majority voting mechanism, ensuring short switching time and high system reliability in case of main sensor failure. The electro-hydraulic proportional control system dynamically adjusts the outrigger pressure distribution of the transport vehicle, keeping the ground pressure within the ground bearing capacity threshold.

[0030] Foundation settlement prediction models are commonly used models in engineering calculations and are existing models.

[0031] Specifically, the foundation settlement prediction model is a simple model of the relationship between load and settlement. It assumes that the soil is a linear elastic material and establishes a quantitative relationship between "load and settlement." In other words, it determines how much ground is subsided by a heavy vehicle.

[0032] The transport vehicle adopts a layered design, consisting of an upper and lower layer. The lower layer uses a microcontroller for motion control and includes four adjustable wheels and an electro-hydraulic proportional control system. This system controls the wheel movement, ensuring efficient and precise adaptive workpiece positioning and transport vehicle movement control. The upper layer is controlled by an industrial computer with powerful data processing capabilities. This upper layer includes a chassis where the workpiece is placed. It is also equipped with a multi-sensor fusion system, providing the necessary data support for positioning, navigation, and path planning. This enables precise positioning, navigation, and path planning in complex environments, as well as ensuring the stability and safety of the workpiece during transport, thus improving overall transport efficiency and providing strong technical support for heavy-duty transport in complex and changing environments.

[0033] The terrain adaptation strategy of the transport vehicle is achieved through differential adjustment of wheel length and coordinated adaptation of torso posture. When driving on flat ground, the left and right wheels of the transport vehicle are supported by symmetrical lengths, the torso maintains a horizontal posture, the wheel-ground contact points are evenly distributed, and the center of mass projection is located in the center of the support area, ensuring the initial balance of flat terrain.

[0034] For slope terrain perception and wheel-foot movement, when in contact with slope terrain, the slope is perceived through a multi-sensor fusion system.

[0035] During the adjustment of wheel length, it is always ensured that the wheel-ground contact point is always within the effective support range and that the overall center of mass projection does not exceed the support boundary, thereby adapting to the slope changes of the sloping terrain.

[0036] When a transport vehicle moves up a slope, its body may sway unsteadily due to gravity. To reduce the impact of these swaying movements on attitude adjustment, an inertial control mechanism is introduced into the attitude feedback. The parameter 'a' of this mechanism depends on the size and speed of the transport vehicle and can be adjusted according to the specific application scenario.

[0037] like Figure 2 As shown, the top part represents execution, and the bottom part represents feedback. The first input is the initial pose. Based on the rotation matrix of the ground slope and the pitch and roll angles of the ground, the input pose is adjusted for "slope matching" and an intermediate pose is output. Then, based on the rotation matrix of the vehicle's own attitude, the attitude is further adjusted and output to the vehicle. The vehicle transmits control flow to the attitude feedback through feedback, and further adjustments are made for slope adaptation.

[0038] The multi-sensor fusion system includes laser sensors, GNSS equipment, inertial navigation sensors, and magnetic navigation sensors. Each sensor leverages its strengths to jointly construct a comprehensive navigation perception network. The laser sensor provides high-precision ranging information, the GNSS equipment achieves global positioning, the inertial navigation sensor ensures stable tracking, and the magnetic navigation sensor assists in positioning. Through the fusion processing of multi-sensor data, the transport vehicle can achieve accurate and stable navigation in various environments. The built-in temperature compensation algorithm in the data layer ensures that the measurement error is extremely small under certain temperature conditions.

[0039] The core breakthrough of the four-layer sensing architecture lies in the development of the IGD-UKF dynamic perturbation suppression algorithm:

[0040] in, This represents the weighting coefficients dynamically assigned based on sensor confidence levels. For the nine-axis attitude transformation function, the regularization coefficient is... The value ranges from 0.05 to 0.2. This algorithm creatively combines the advantages of improved gradient descent (IGD) and unscented Kalman filtering (UKF). It successfully reduces the tilt angle measurement error to an extremely low level, showing a significant improvement over traditional Kalman filtering, and providing a high-precision pose base for hydraulic control.

[0041] Laser rangefinders can be replaced by millimeter-wave radar or binocular vision cameras, both of which can work with IMUs to complete terrain perception; fiber optic pressure arrays can be replaced by strain gauge pressure sensors, which, although slightly less accurate, can meet the needs of foundation pressure monitoring through multi-node data fusion.

[0042] The IGD-UKF algorithm can be replaced by a fusion algorithm of Extended Kalman Filter (EKF) and Adaptive Fuzzy Control. Although its accuracy is slightly lower than that of IGD-UKF under strong disturbances, it is easier to implement and can meet the pose correction requirements for terrains of medium to low complexity. The majority voting mechanism can be replaced by a weighted voting mechanism, which judges the validity of data through weighted calculation and is suitable for scenarios with large differences in sensor accuracy.

[0043] The four-layer perception architecture (data layer-feature layer-decision layer-fault tolerance layer) can be simplified to a two-layer architecture of "data fusion layer-decision control layer". It integrates feature extraction and data fusion, and processes multi-sensor data in real time through edge computing modules. Although the fault tolerance capability is slightly reduced, the system latency is reduced by 10%, which is suitable for short-distance transportation scenarios with higher requirements for response speed.

[0044] The adaptive ground-level height control method uses inertial navigation and tilt sensors to perceive the slope in real time, driving differentiated extension and retraction of the wheel feet on both sides of the slope, and simultaneously adjusting the body posture to adapt the local coordinate system to the slope surface. In practical applications, it can adapt to various slope terrains, reduce workpiece slippage, and significantly improve the stability and accuracy of slope transportation.

[0045] The constructed four-layer sensing architecture enables comprehensive monitoring of the foundation condition: the data layer uses a dual-axis tilt sensor to capture real-time changes in the roll and pitch angles of the support platform; the feature layer uses a fiber optic pressure array to map the pressure distribution of the outriggers, eliminating local distortion errors; and the decision layer uses a laser displacement sensor combined with a settlement prediction model to achieve millimeter-level deformation compensation. By dynamically adjusting the pressure distribution of the outriggers through an electro-hydraulic proportional control system, the ground pressure can be controlled within the foundation bearing capacity threshold, significantly reducing the settlement rate and effectively attenuating vibration transmission, thereby effectively avoiding the tilting problem of the transport vehicle caused by uneven foundation conditions.

[0046] A multi-sensor fusion system is adopted to achieve complementary advantages and mutual support for weaknesses, ensuring continuous and stable positioning. Redundant IMUs are deployed in the fault-tolerant layer, and a dual mechanism of "hardware watchdog + majority voting" is adopted to enable rapid switching when the main sensor fails. The mean time between failures of the system is significantly improved compared with existing technologies, and it can be well adapted to transportation scenarios with complex signals such as tunnels and mountainous areas.

[0047] To enhance pose correction accuracy and ensure stable dynamic balance of the workpiece, the core algorithm IGD-UKF integrates the fast convergence characteristics of Improved Gradient Descent (IGD) with the anti-disturbance advantages of Unscented Kalman Filtering (UKF), significantly reducing tilt angle measurement errors and greatly shortening pose correction response time. When encountering sudden slope changes or instantaneous foundation settlement, the system can effectively control the tilt angle fluctuation range of the support platform through multi-level damping mode switching and dynamic adjustment of the outriggers, ensuring the workpiece remains stable throughout the entire transportation process.

[0048] By collecting slope information in real time through an inertial navigation module and tilt sensors, the vehicle drives the wheels on both sides of the slope to achieve differentiated extension and retraction adjustments, while simultaneously adapting the posture of the vehicle body. This ensures that the local coordinate system of the vehicle body is precisely aligned with the slope surface, thereby guaranteeing that the wheel-to-ground contact point is always within the effective working space and that the projected center of gravity of the equipment does not exceed the support boundary. The heavy-duty transport vehicle can stably adapt to multi-gradient slope terrain, significantly reducing the risk of workpiece slippage and fundamentally solving the core technical challenges of heavy-duty transport vehicles being prone to imbalance and workpiece slippage in slope environments.

[0049] A multi-sensor fusion-based hierarchical perception and fault-tolerant architecture: A hierarchical perception architecture is constructed, consisting of a lower-level microcontroller responsible for motion control and an upper-level industrial control computer integrating LiDAR, GNSS, IMU, and magnetic navigation. Three redundant IMUs are deployed, and a hardware watchdog + majority voting mechanism is used to achieve sensor fault tolerance. When GNSS signals are blocked, the IMU can work with the LiDAR to achieve continuous positioning; when the main sensor fails, it can quickly switch over, significantly improving the system's mean time between failures (MTBF) compared to existing technologies, effectively ensuring perception reliability in complex scenarios such as tunnels and mountainous areas.

[0050] A dynamic perturbation suppression and pose correction algorithm fused with IGD-UKF: This algorithm combines the fast convergence of improved gradient descent (IGD) with the perturbation resistance of unscented Kalman filtering (UKF) to form the IGD-UKF algorithm, used for dynamic pose correction of heavy-duty transport vehicles. Compared to traditional Kalman filtering, this algorithm significantly reduces tilt angle measurement errors and greatly shortens pose correction response time. When faced with sudden slope changes or instantaneous foundation settlement, it effectively controls the tilt angle fluctuation range of the support platform, providing precise algorithmic support for the dynamic balance of the entire workpiece transportation process.

[0051] Multi-dimensional ground condition sensing and hydraulic outrigger coordinated control mechanism: A temperature-compensated dual-axis tilt sensor captures changes in the bearing platform angle, a fiber optic pressure array maps the outrigger pressure distribution, and a laser displacement sensor and settlement prediction model are combined to dynamically adjust the outrigger pressure and elongation via an electro-hydraulic proportional control system. This allows for precise control of the ground pressure within the foundation bearing capacity threshold, significantly reducing the foundation settlement rate, effectively attenuating vibration transmission, and effectively solving the problem of heavy-duty transport vehicles easily tilting and becoming unbalanced in complex ground environments such as soft soil and frozen soil.

[0052] Wheel-to-ground contact point workspace constraint and optimization method: Through adaptive calculations using the local and global coordinate systems of the torso, the effective workspace boundary of the wheel-to-ground contact point is mathematically modeled and constrained, clarifying the limit range of wheel extension and retraction. This method can prevent wheel adjustment from exceeding the mechanical workspace, and from a geometric constraint perspective, it helps to improve the stability and safety of adaptive ground clearance control, serving as an important supplement to the core control method.

[0053] The above description is an explanation of this application and not a limitation thereof. The scope of this application is defined by the claims. Within the scope of protection of this application, any form of modification may be made.

Claims

1. A method for adaptive balancing of workpieces in complex environments using a transport vehicle, characterized in that, include: The ground slope and foundation type are perceived in real time through a multi-sensor fusion system to determine the current terrain and load conditions; When driving on flat ground, the left and right wheels of the transport vehicle are symmetrically supported, the body of the transport vehicle remains horizontal, the wheel-ground contact points are evenly distributed, and the center of gravity of the transport vehicle is located in the center of the support area, ensuring the balance on flat ground. When encountering sloping terrain, the multi-sensor fusion system senses the direction and slope of the terrain, drives the wheel of the transport vehicle on the side closer to the lower part to extend, and drives the wheel of the transport vehicle on the other side to shorten, so that the body of the transport vehicle remains horizontal. When the foundation settles unevenly, a four-layer sensing architecture is constructed. Through multi-source information fusion, the position and posture of the support platform are accurately perceived. The data layer uses a high-precision dual-axis tilt sensor to monitor the roll angle θ and pitch angle φ of the transport vehicle body in real time. The feature layer integrates a fiber optic pressure array and maps the pressure distribution of the hydraulic outriggers through the Bragg wavelength displacement principle. Data compensation between adjacent nodes eliminates local distortion errors. The decision layer uses a laser displacement sensor combined with a foundation settlement prediction model to achieve millimeter-level deformation compensation. The fault-tolerant layer deploys three sets of redundant IMUs using a hardware watchdog + majority voting mechanism, resulting in short switching time and high system reliability in the event of a main sensor failure. By dynamically adjusting the pressure distribution of the outriggers of the transport vehicle through an electro-hydraulic proportional control system, the grounding pressure can be controlled within the foundation bearing capacity threshold.

2. The adaptive balancing method for workpieces in complex environments of a transport vehicle as described in claim 1, characterized in that: The multi-sensor fusion system includes laser sensors, GNSS equipment, inertial navigation sensors, and magnetic navigation sensors, which together construct a comprehensive navigation perception network. The laser sensor provides high-precision ranging information, the GNSS equipment achieves global positioning, the inertial navigation sensor ensures stable tracking, and the magnetic navigation sensor assists in positioning.

3. The adaptive balancing method for workpieces in complex environments of a transport vehicle as described in claim 1, characterized in that: The core of the four-layer sensing architecture lies in the development of the IGD-UKF dynamic perturbation suppression algorithm: in, This represents the weighting coefficients dynamically assigned based on sensor confidence levels. For the nine-axis attitude transformation function, the regularization coefficient is... The value ranges from 0.05 to 0.

2.

4. The adaptive balancing method for workpieces in complex environments of a transport vehicle as described in claim 1, characterized in that: The center of gravity of the transport vehicle is projected to the center of the support area to ensure balance on the flat ground.

5. The adaptive balancing method for workpieces in complex environments of a transport vehicle as described in claim 1, characterized in that: The data layer has a built-in temperature compensation algorithm that ensures minimal measurement error under certain temperature conditions.

6. The adaptive balancing method for workpieces in complex environments of a transport vehicle as described in claim 1, characterized in that: The transport vehicle adopts a layered design, consisting of an upper layer and a lower layer. The lower layer uses a microcontroller to achieve motion control and includes four adjustable wheels and an electro-hydraulic proportional control system. The movement of the wheels is controlled by the electro-hydraulic proportional control system to ensure efficient and precise adaptive workpiece maintenance and transport vehicle movement control.

7. The adaptive balancing method for workpieces in complex environments of a transport vehicle as described in claim 6, characterized in that: The upper layer is controlled by an industrial computer with powerful data processing capabilities. The upper layer includes the torso, on which the workpiece is placed. At the same time, the upper layer is equipped with a multi-sensor fusion system to provide the data support required for the positioning, navigation and path planning of the transport vehicle.

8. The adaptive balancing method for workpieces in complex environments of a transport vehicle as described in claim 1, characterized in that: The laser rangefinder can be replaced by a millimeter-wave radar / binocular vision camera, both of which can work with the IMU to complete terrain perception; the fiber optic pressure array can be replaced by a strain gauge pressure sensor, which can meet the ground pressure monitoring requirements through multi-node data fusion.

9. The adaptive balancing method for workpieces in complex environments of a transport vehicle as described in claim 1, characterized in that: The IGD-UKF algorithm is replaced by a fusion algorithm of extended Kalman filter (EKF) and adaptive fuzzy control. The majority voting mechanism can be replaced by a weighted voting mechanism, which judges the validity of data through weighted calculation. This is suitable for scenarios with large differences in sensor accuracy.

10. The adaptive balancing method for workpieces in complex environments of a transport vehicle as described in claim 1, characterized in that: The four-layer perception architecture is simplified to a two-layer architecture of "data fusion layer - decision control layer". It integrates feature extraction and data fusion, and processes multi-sensor data in real time through edge computing modules. The fault tolerance is reduced, and the system latency is reduced by 10%, which is suitable for short-distance transportation scenarios with higher requirements for response speed.

Images