AGV (Automatic Guided Vehicle) intelligent transfer control scheduling method and system for personnel body transfer

Abstract

The invention relates to the technical field of artificial intelligence and AGV control, and discloses an AGV intelligent transfer control scheduling method and system for personnel body transfer, and the method comprises the steps: obtaining a space pose relation between an AGV and a target bed; the Mecanum wheel chassis is driven to move in all directions, and the posture is finely adjusted to achieve millimeter-level butt joint; the gravity center and the supporting state of the body are recognized in real time through a multi-section flexible transfer mechanism in combination with a pressure distribution sensing array; and the speed of each conveying unit is dynamically distributed to realize zero relative sliding, and a local lifting device driven by a voice coil motor is used for performing millisecond-level vertical compensation to ensure zero clearance in the whole process. The system comprises a space pose sensing module, a chassis motion control module, a flexible transfer execution module, a pressure-speed coordinated regulation module, a vertical clearance compensation module and a task management module. The transfer safety, comfort and automation level are remarkably improved, and the system is suitable for high-sensitivity scenes such as severe illness, postoperative nursing and old-age nursing.

Description

Technical Field

[0001] This invention belongs to the field of artificial intelligence and automated guided vehicle (AGV) control technology, specifically relating to an AGV intelligent transfer control and scheduling method and system for the transfer of human bodies. Background Technology

[0002] Assistive care technologies for people with disabilities or limited mobility have become an important branch of modern medical devices and automated logistics. Within various medical institutions and rehabilitation centers, the safe and efficient transport of personnel is not only a fundamental aspect of ensuring a smooth patient treatment process, but also a key indicator for assessing nursing quality, reducing the workload of medical staff, and improving the operational efficiency of the institution. Automated Guided Vehicles (AGVs), as the core carrier of intelligent mobile platforms, are expanding their application in medical scenarios from material delivery to high-precision human body transport. This necessitates the integration of multi-dimensional motion control, human-machine interaction safety, and dynamic environmental adaptability to meet the stringent clinical requirements for non-invasive, stable, and precise transport.

[0003] Among them, the AGV system for transporting human bodies focuses on achieving seamless docking and non-destructive transfer between hospital beds, examination tables, and transport platforms. This technological approach relies on the deep synergy of an omnidirectional mobile chassis, a retractable load-bearing mechanism, and a synchronous transmission mechanism. It aims to achieve complex posture adjustments and stable heavy-load transport within a limited space through mechatronic design. Its core objective is to ensure that patients are not affected by shear forces, vibrations, or sudden displacements during transport, while avoiding the risks of manual handling. This is particularly suitable for special populations such as those with fragile skin, those recovering post-surgery, or those who are bedridden for extended periods.

[0004] Existing AGVs for personnel transfer still face multiple technical bottlenecks in practical applications: First, the transfer process generally uses unidirectional conveyor belts or rigid pusher structures. When the sliding plate retracts or extends, significant relative sliding occurs between the conveying medium and the human body due to speed mismatch. The resulting shear force can easily cause secondary injuries such as pressure sores and incision tears. Second, under heavy load conditions, the sliding friction resistance between the conveyor belt and the support surface increases sharply, leading to violent load fluctuations in the drive system and causing jerking and shaking during operation, seriously affecting riding comfort and structural stability. Third, traditional AGVs are limited by two-wheel differential speed or steering wheel mechanisms, making it difficult to achieve multi-degree-of-freedom flexible movement in narrow wards or environments with dense equipment. They cannot accurately align with beds at different heights, angles, or positions, often requiring multiple adjustments or even manual intervention. These problems collectively restrict the large-scale deployment of medical AGVs in high-safety-standard scenarios, urgently requiring an intelligent transfer control and scheduling method and system that integrates zero-shear force control, rolling drag reduction support, and omnidirectional high-precision docking. Summary of the Invention

[0005] This invention provides an intelligent transfer control and scheduling method and system for AGV (Automated Guided Vehicle) transport of human bodies, aiming to solve the problems of patient discomfort, secondary injury risk, and low operational efficiency caused by excessive shear force, insufficient operational stability, and low docking accuracy during the transfer process in existing technologies. This invention constructs a multi-degree-of-freedom collaborative motion platform, combining high-precision pose perception, dynamic friction compensation mechanism, and adaptive speed synchronization strategy to achieve smooth transfer of the human body without relative sliding or shear stress transmission during the transfer process, and to achieve millimeter-level precise docking with beds at different heights and angles in complex medical environments.

[0006] According to one aspect of the present invention, an intelligent transfer control and scheduling method for AGVs used for the transfer of human body parts is provided, comprising: Acquire spatial pose relationship data between the target bed and the current docking position of the AGV. The spatial pose relationship data includes three-dimensional coordinate offset, pitch angle deviation, roll angle deviation and heading angle deviation. Based on the spatial pose relationship data, the AGV chassis is driven to perform omnidirectional movement and attitude adjustment, so that the AGV's transfer platform and the target bed are aligned in the horizontal plane and matched in the vertical direction. After the transfer platform completes the initial docking with the target bed, the multi-segment flexible transfer mechanism is activated. The multi-segment flexible transfer mechanism consists of several parallel flexible transmission units. Each flexible transmission unit includes an independently servo-driven conveyor belt module, a pressure distribution sensor array, and a local fine-tuning lifting device. The pressure distribution sensor array collects the contact pressure distribution map of the human body on the transfer platform in real time, and identifies the center of gravity of the body, the boundary of the support area, and the local pressure peak points based on the map. Based on the contact pressure distribution map, the running speed and direction of each flexible conveyor unit are dynamically allocated so that all conveyor belt modules maintain zero relative velocity with the human body surface during the transfer process, thereby eliminating shear force. During the transfer process, the local fine-tuning lifting device is activated simultaneously to perform millisecond-level dynamic compensation on the vertical height of each flexible conveyor unit, so as to offset the gap changes between the conveyor belt and the support surface caused by local deformation of the human body or unevenness of the bed surface, and ensure a close fit throughout the process. Once the transfer is complete and the human body has been fully transferred to the target bed, the AGV automatically executes the evacuation path planning and returns to the standby area.

[0007] Preferably, the acquisition of spatial pose relationship data between the target bed and the current docking position of the AGV specifically includes: By deploying a binocular vision sensor array on top of the AGV, the structural feature points of the target bed are reconstructed in three dimensions, and the spatial coordinates of the bed edge, bed leg connection points and bed surface marking patterns are extracted. By installing laser ranging sensor groups at the four corners of the AGV chassis, the distance between the AGV and the surrounding walls, equipment and ground markings is measured. Combined with the pre-stored indoor digital twin map, the six-degree-of-freedom pose of the AGV itself in the global coordinate system is calculated. The coordinates of the feature points of the target bed are transformed with the global pose of the AGV, and the relative pose deviation between the two is calculated.

[0008] Preferably, the omnidirectional movement and attitude adjustment of the AGV chassis specifically includes: The AGV chassis adopts a Mecanum wheel omnidirectional drive configuration, with each of the four Mecanum wheels driven by an independent servo motor. Each servo motor is equipped with a high-resolution encoder and a closed-loop current feedback controller. Based on the calculated three-dimensional coordinate offset, translation speed commands are generated in the X and Y axis directions; Based on the pitch angle deviation and roll angle deviation, the speed difference between the front and rear wheel sets and the left and right wheel sets is controlled to achieve micro-angle tilt adjustment of the chassis around the X and Y axes; Based on the heading angle deviation, the rotation direction and speed of the four Mecanum wheels are controlled by differential steering to achieve precise homing around the Z-axis; Throughout the adjustment process, the actual attitude of the chassis is monitored in real time by the inertial measurement unit and closed-loop feedback correction is performed with the target attitude until the pose error is less than a preset threshold. The preset threshold is that the horizontal position error is no more than three millimeters, the height error is no more than two millimeters, and the angle error is no more than 0.5 degrees.

[0009] Preferably, in the multi-segment flexible transfer mechanism, the number of flexible transfer units is seven, arranged sequentially from head to foot along the longitudinal direction of the human body. Each segment is 250 mm long and 600 mm wide. Adjacent segments are connected by elastic hinges, allowing a maximum relative deflection of 15 degrees in the vertical direction. Each flexible conveyor unit's conveyor belt module is made of polyurethane composite material with a micron-level anti-slip texture on the surface. Its static friction coefficient is not less than 0.8, and the dynamic friction coefficient fluctuation range is controlled within ±5%. Each conveyor module is driven by a pair of synchronous pulleys, which are connected to a brushless DC servo motor via a planetary reducer. The servo motor has a control cycle of one millisecond and a speed adjustment resolution of 0.1 revolutions per minute.

[0010] Preferably, the pressure distribution sensing array consists of 320 piezoresistive sensing units arranged in a 16-row, 20-column matrix. Each sensing unit has a sensing area of ​​25 square millimeters, a range of 0 to 500 Newtons, and a sampling frequency of 200 Hz. The contact pressure distribution map is noise-suppressed using a sliding window filtering algorithm and local voids are filled using morphological closing operations to form a continuous pressure field image. Based on the pressure field image, the two-dimensional projected coordinates of the body's center of gravity are calculated using a centroid-weighted algorithm, as shown in the formula: , ; in, The pressure value of the sensing unit in the i-th row and j-th column is... and These are the horizontal and vertical coordinates of the corresponding sensing unit in the coordinate system of the transfer platform.

[0011] Preferably, the dynamic allocation of the operating speed and direction of each flexible transmission unit specifically includes: The transfer platform is divided into several speed control zones, and each control zone corresponds to a flexible transmission unit. Using the projection point of the body's center of gravity as a reference, calculate the longitudinal offset distance of each control zone relative to the center of gravity; Set the main teleport speed The reference speed for the transfer mission ranges from 10 mm to 50 mm per second. For the k-th flexible transmission unit, its operating speed Calculate using the following formula: ; in, The distance of the centroid offset of the k-th segment. The maximum effective offset length is set to 800 millimeters. This is the speed compensation coefficient, with a value of 0.2. when When positive, it indicates that the segment is located behind the center of gravity and the speed is slightly higher than the reference speed; when... When the value is negative, it indicates that the object is in front and its speed is slightly lower than the reference speed, thereby achieving synchronous movement between the overall transmission surface and the human body.

[0012] Preferably, the local fine-tuning lifting device uses a linear actuator driven by a voice coil motor, with a stroke range of ±20 mm, a response time of five milliseconds, and a positioning accuracy of 0.05 mm. Each flexible conveyor unit is equipped with three linear actuators arranged in a triangle at its bottom, which enable fine-tuning of the attitude of the local plane through three-point support; The fine-tuning command is determined by the local pressure gradient in the pressure distribution map. When the pressure gradient in a certain area exceeds the threshold of 100 Newtons per square meter, it is determined that there is local suspension or compression, and the corresponding area's rise and fall compensation is immediately activated to restore the pressure gradient to a safe range.

[0013] Preferably, the evacuation path planning is based on the improved A* algorithm, combined with real-time obstacle detection data and bed occupancy status, to generate a collision-free, shortest-time reversal trajectory; During the evacuation process, the AGV keeps the transfer platform at its lowest storage height and disables all conveying functions, only enabling the chassis motion control and environmental perception modules.

[0014] According to another aspect of the present invention, an AGV intelligent transfer control system for the transfer of human body parts is provided, comprising: The spatial pose perception module is used to acquire spatial pose relationship data between the target bed and the current docking position of the AGV; The chassis omnidirectional motion control module is used to drive the AGV chassis to perform omnidirectional movement and attitude adjustment to achieve initial docking with the target bed. The multi-segment flexible transfer execution module includes several flexible transmission units, a pressure distribution sensor array, and a local fine-tuning lifting device, used to perform human body transfer without shearing force. The pressure-speed coordinated control module is used to dynamically allocate the operating speed and direction of each flexible transmission unit based on the pressure distribution map; The vertical gap compensation module is used to activate the local fine-tuning lifting device according to the local pressure gradient, so as to achieve full contact between the conveying surface and the support surface. The task status management module is used to trigger the evacuation path planning and control the AGV to return to the standby area after the transfer is completed.

[0015] The spatial pose perception module further includes a binocular vision 3D reconstruction unit, a laser ranging fusion positioning unit, and a relative pose calculation unit. The chassis omnidirectional motion control module further includes a Mecanum wheel servo drive unit, an inertial attitude feedback unit, and a posture closed-loop correction unit. In the multi-segment flexible transfer execution module, the flexible transmission units are connected in series by elastic hinges, and each segment is independently configured with a servo motor, a conveyor belt, a pressure sensor array and a linear actuator driven by three voice coil motors. The pressure-speed coordinated control module has a built-in center of gravity calculation engine and speed allocation algorithm engine, and its real-time processing cycle is no more than ten milliseconds. The vertical clearance compensation module is equipped with a pressure gradient monitoring threshold judge and an actuator drive controller, and its response delay does not exceed eight milliseconds.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: This invention, through the construction of a multi-segment flexible transfer mechanism and a high-precision pressure sensing system, achieves for the first time zero-relative-velocity synchronous movement between the transfer medium and the human body surface during personnel transport. This fundamentally eliminates shear stress caused by speed mismatch and effectively avoids secondary damage to fragile skin areas or postoperative wounds. Millisecond-level dynamic compensation of the transfer surface via a local fine-tuning lifting device ensures seamless contact throughout the transfer process, significantly improving operational stability and eliminating vibrations and jerks caused by sudden changes in internal frictional resistance in traditional equipment. The fusion of a Mecanum wheel omnidirectional chassis and six-degree-of-freedom pose sensing enables the AGV to achieve millimeter-level precise docking with beds of any height and angle in confined medical environments, increasing the docking success rate to over 99.5%. The overall system, through pressure-speed-height triple closed-loop coordinated control, elevates the comfort, safety, and automation of the transfer process to a new level, significantly reducing the physical burden on medical staff while meeting the stringent requirements of highly sensitive scenarios such as intensive care, postoperative rehabilitation, and elderly care. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall technical architecture of the AGV intelligent transfer control and scheduling method and system for the transfer of human bodies proposed in this invention. Figure 2 This is a schematic diagram of the core principle framework of the multi-segment flexible transfer mechanism and the pressure-speed coordinated control mechanism in this invention; Figure 3 This is a logical flowchart of the collaborative docking between spatial pose perception and chassis omnidirectional motion control in this invention. Figure 4 This is a schematic diagram of the multi-level interaction relationship and data flow of dynamic velocity distribution and vertical clearance compensation based on pressure distribution map in this invention. Figure 5 This is a logical flowchart of the task status management and evacuation path planning in this invention. Detailed Implementation

[0018] Example 1. This invention provides an AGV intelligent transfer control and scheduling method and system for human body transport. Its core lies in achieving smooth transfer of the human body without relative sliding or shear stress transmission during the transfer process through a multi-degree-of-freedom cooperative motion platform, high-precision pose perception, dynamic friction compensation mechanism, and adaptive speed synchronization strategy. It also enables millimeter-level precise docking with beds of different heights and angles in complex medical environments. The specific implementation steps of this method will be described in detail below.

[0019] See Figure 1-5 The method includes the following steps: S1, obtain the spatial pose relationship data between the target bed and the current parking position of the AGV; S2, based on the spatial pose relationship data, drive the AGV chassis to perform omnidirectional movement and attitude adjustment; S3, after the transfer platform completes the initial docking with the target bed, the multi-segment flexible transfer mechanism is activated; S4, which collects the contact pressure distribution map of the human body on the transfer platform in real time through the pressure distribution sensor array; S5, based on the contact pressure distribution map, dynamically allocate the operating speed and direction of each flexible transmission unit; S6, during the transfer process, the local fine-tuning lifting device is activated synchronously to perform millisecond-level dynamic compensation for the vertical height of each flexible conveying unit. S7. Once the transfer is complete and the human body has been fully transferred to the target bed, the AGV automatically executes the evacuation path planning and returns to the standby area.

[0020] In step S1, spatial pose relationship data between the target bed and the current docking position of the AGV is acquired. This spatial pose relationship data includes three-dimensional coordinate offset, pitch angle deviation, roll angle deviation, and heading angle deviation. Specifically, this step includes: using a binocular vision sensor array deployed on top of the AGV to perform three-dimensional reconstruction of the structural feature points of the target bed, extracting the spatial coordinates of the bed edge, bed leg connection points, and bed surface markings; using laser ranging sensor groups installed at the four corners of the AGV chassis to measure the distance between the AGV and surrounding walls, equipment, and ground markings; combining this with a pre-stored indoor digital twin map to calculate the AGV's six-degree-of-freedom pose in the global coordinate system; and performing coordinate transformation between the feature point coordinates of the target bed and the AGV's global pose to calculate the relative pose deviation between the two. The binocular vision sensor array employs a stereo imaging configuration with a baseline length of 300 mm, an image resolution of 1280 x 960 pixels, and a frame rate of 30 Hz. A depth map is generated using a semi-global matching algorithm, and the bed surface normal vector and boundary contour are extracted via RANSAC plane fitting. The laser ranging sensor group consists of four single-point laser rangefinders, with a ranging range of 0.2 m to 5 m, an accuracy of ±1 mm, and a sampling frequency of 100 Hz. Its data is fused with the attitude angles output from the inertial measurement unit, and pose estimation is performed using an extended Kalman filter. The final output is the AGV's position coordinates (X, Y, Z) and Euler angles (pitch, roll, and yaw) in the global coordinate system. Relative pose calculation uses a homogeneous transformation matrix for coordinate system transformation to ensure the integrity and consistency of the pose relationship data.

[0021] In step S2, based on the spatial pose relationship data, the AGV chassis is driven to perform omnidirectional movement and attitude adjustment, so that the AGV's transfer platform and the target bed are aligned in the horizontal plane and matched in the vertical direction. The AGV chassis adopts a Mecanum wheel omnidirectional drive configuration, with each of the four Mecanum wheels driven by an independent servo motor. Each servo motor is equipped with a high-resolution encoder and a closed-loop current feedback controller. Translation speed commands in the X and Y axes are generated based on the calculated three-dimensional coordinate offset. The speed difference between the front and rear wheel sets and the left and right wheel sets is controlled based on the pitch and roll angle deviations to achieve micro-angle tilt adjustment of the chassis around the X and Y axes. Based on the heading angle deviation, the rotation direction and speed of the four Mecanum wheels are controlled by differential steering to achieve precise return to center around the Z axis. Throughout the adjustment process, the actual attitude of the chassis is monitored in real time by an inertial measurement unit, and closed-loop feedback correction is performed with the target attitude until the pose error is less than a preset threshold. The preset thresholds are: horizontal position error no greater than 3 mm, height error no greater than 2 mm, and angle error no greater than 0.5 degrees. The rolling direction of the Mecanum wheel forms a 45-degree angle with the vehicle's coordinate system, and its kinematic model is as follows:

[0022] in, , Let be the linear velocity component of the vehicle body in the global coordinate system. Let be the angular velocity about the Z-axis. Let the radius of the Mecanum wheel be 1. It is half the track width. to The angular velocities of the four Mecanum wheels are given. The control system updates motion commands with a period of ten milliseconds to ensure the continuity and stability of the attitude adjustment process.

[0023] In step S3, after the transfer platform completes the initial docking with the target bed, the multi-segment flexible transfer mechanism is activated. The multi-segment flexible transfer mechanism consists of seven parallel flexible transfer units arranged longitudinally from head to toe along the body's length. Each segment is 250 mm long and 600 mm wide, with adjacent segments connected by elastic hinges, allowing for a maximum relative deflection of 15 degrees in the vertical direction. Each flexible transfer unit includes an independently servo-driven conveyor belt module, a pressure distribution sensor array, and a local fine-tuning lifting device. The conveyor belt module is made of polyurethane composite material with a micron-level anti-slip texture on its surface. Its static friction coefficient is not less than 0.8, and the dynamic friction coefficient fluctuation range is controlled within ±5%. Each conveyor belt module is driven by a pair of synchronous pulleys, which are connected to a brushless DC servo motor via a planetary reducer. The servo motor's control cycle is one millisecond, and its speed adjustment resolution is 0.1 revolutions per minute. The flexible hinge is made of high elastic modulus silicone material, with an embedded metal spring sheet to provide restoring torque, ensuring that each segment can automatically return to its original position after being compressed, while allowing local deformation to adapt to the curvature of the human body.

[0024] In step S4, the contact pressure distribution map of the human body on the transfer platform is acquired in real time using the pressure distribution sensing array. The pressure distribution sensing array consists of 320 piezoresistive sensing units arranged in a 16x20 matrix. Each sensing unit has a sensing area of ​​25 square millimeters, a range of 0 to 500 Newtons, and a sampling frequency of 200 Hz. The contact pressure distribution map is noise-suppressed using a sliding window filtering algorithm, and morphological closure operations are used to fill local voids, forming a continuous pressure field image. Based on the pressure field image, the two-dimensional projected coordinates of the body's center of gravity are calculated using a centroid-weighted algorithm, as shown in the formula: , ; in, The pressure value of the sensing unit in the i-th row and j-th column is... and These are the horizontal and vertical coordinates of the corresponding sensing unit in the coordinate system of the transfer platform. and These are the x-coordinate and y-coordinate (two-dimensional projection) of the human body's center of gravity in the coordinate system of the transfer platform.

[0025] The sliding window filtering uses a five-point mean filter with a window size of three times three, which effectively suppresses high-frequency noise; the morphological closing operation uses a five-times-five square structuring element to eliminate data loss caused by local occlusion or sensor failure.

[0026] In step S5, based on the contact pressure distribution map, the operating speed and direction of each flexible conveyor unit are dynamically allocated, ensuring that all conveyor belt modules maintain zero relative velocity with the human body surface during the transfer process, thereby eliminating shear force. This step specifically includes: dividing the transfer platform into seven speed control zones, each corresponding to a flexible conveyor unit; calculating the longitudinal offset distance of each control zone relative to the center of gravity, using the body's center of gravity projection point as a reference; and setting the main conveying speed. The reference speed for the transfer task ranges from 10 mm to 50 mm per second; for the k-th flexible transfer unit, its operating speed... Calculate using the following formula: ; in, The distance of the centroid offset of the k-th segment. The maximum effective offset length is set to 800 millimeters. The speed compensation coefficient is 0.2; when When positive, it indicates that the segment is located behind the center of gravity and the speed is slightly higher than the reference speed; when... A negative value indicates that the object is positioned in front, with a speed slightly lower than the baseline speed, thus achieving synchronized movement between the overall conveyor surface and the human body. The speed allocation command is generated by the pressure-speed coordinated control module, with a processing cycle of ten milliseconds, ensuring real-time speed adjustment.

[0027] In step S6, during the transfer process, a local fine-tuning lifting device is simultaneously activated to dynamically compensate the vertical height of each flexible conveyor unit at the millisecond level. This compensates for changes in the gap between the conveyor belt and the support surface caused by local deformation of the human body or unevenness of the bed surface, ensuring a close fit throughout the process. The local fine-tuning lifting device uses a linear actuator driven by a voice coil motor, with a stroke range of ±20 mm, a response time of five milliseconds, and a positioning accuracy of 0.05 mm. Three linear actuators arranged in a triangle are installed at the bottom of each flexible conveyor unit, achieving local plane attitude fine-tuning through three-point support. The fine-tuning command is determined by the local pressure gradient in the pressure distribution map. When the pressure gradient in a certain area exceeds the threshold of 100 Newtons per square meter, it is determined that there is local suspension or compression, and the corresponding area's lifting compensation is immediately activated to restore the pressure gradient to a safe range. The pressure gradient is calculated using the central difference method, calculating the gradient amplitude of the pressure field image. If the gradient amplitude is greater than the threshold, a compensation action is triggered, and the compensation amount is obtained by linearly mapping the difference between the gradient amplitude and the preset safety gradient.

[0028] In step S7, once the transfer is complete and the human body is fully transferred to the target bed, the AGV automatically executes the evacuation path planning and returns to the standby area. The evacuation path planning is based on an improved A* algorithm, combining real-time obstacle detection data and bed occupancy status to generate a collision-free, shortest-time reversal trajectory. During the evacuation process, the AGV maintains the transfer platform at its lowest possible storage height and disables all conveying functions, activating only the chassis motion control and environmental perception modules. The improved A* algorithm introduces a dynamic weighting factor, adjusting the heuristic function based on obstacle density and path curvature to ensure a smooth path and reliable obstacle avoidance. The environmental perception module continuously receives data from lidar and ultrasonic sensors to construct a local obstacle map for real-time path correction.

[0029] See Figure 1 The AGV intelligent transfer control system for personnel body transport includes a spatial pose perception module, a chassis omnidirectional motion control module, a multi-segment flexible transfer execution module, a pressure-speed coordinated control module, a vertical clearance compensation module, and a task status management module. The spatial pose perception module further includes a binocular vision 3D reconstruction unit, a laser ranging fusion positioning unit, and a relative pose calculation unit; the chassis omnidirectional motion control module further includes a Mecanum wheel servo drive unit, an inertial attitude feedback unit, and a pose closed-loop correction unit; in the multi-segment flexible transfer execution module, the flexible transfer units are connected in series via elastic hinges, and each segment is independently configured with a servo motor, conveyor belt, pressure sensor array, and three voice coil motor-driven linear actuators; the pressure-speed coordinated control module has a built-in center of gravity calculation engine and speed allocation algorithm engine, with a real-time processing cycle of no more than ten milliseconds; the vertical clearance compensation module has a pressure gradient monitoring threshold judge and an actuator drive controller, with a response delay of no more than eight milliseconds; the task status management module is responsible for coordinating the working status of each module, triggering the evacuation process based on the transfer completion signal, and monitoring the system health status.

[0030] The various modules of the system interact via high-speed industrial Ethernet, with a communication cycle of one millisecond, ensuring the synchronization and real-time performance of control commands. All sensor data is timestamped and synchronized using a hardware-triggered mechanism to eliminate timing deviations caused by asynchronous multi-source acquisition. The system software architecture adopts a layered design: the bottom layer is a real-time operating system, the middle layer is a control algorithm library, and the top layer is a task scheduling engine. Each layer communicates with the others through standard interfaces, ensuring the system's maintainability and scalability.

[0031] This embodiment achieves shear-free transfer, highly stable operation, and millimeter-level precise docking during the transfer of human bodies using the aforementioned method and system, effectively solving problems such as patient discomfort, risk of secondary injury, and low operational efficiency in existing technologies.

[0032] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. An intelligent transfer control and scheduling method for AGVs used for the transfer of human bodies, characterized in that, include: Acquire spatial pose relationship data between the target bed and the current docking position of the AGV. The spatial pose relationship data includes three-dimensional coordinate offset, pitch angle deviation, roll angle deviation and heading angle deviation. Based on the spatial pose relationship data, the AGV chassis is driven to perform omnidirectional movement and attitude adjustment, so that the AGV's transfer platform and the target bed are aligned in the horizontal plane and matched in the vertical direction. After the transfer platform completes the initial docking with the target bed, the multi-segment flexible transfer mechanism is activated. The multi-segment flexible transfer mechanism consists of several parallel flexible transmission units. Each flexible transmission unit includes an independently servo-driven conveyor belt module, a pressure distribution sensor array, and a local fine-tuning lifting device. The pressure distribution sensor array collects the contact pressure distribution map of the human body on the transfer platform in real time, and identifies the center of gravity of the body, the boundary of the support area, and the local pressure peak points based on the map. Based on the contact pressure distribution map, the running speed and direction of each flexible conveyor unit are dynamically allocated so that all conveyor belt modules maintain zero relative velocity with the human body surface during the transfer process, thereby eliminating shear force. During the transfer process, the local fine-tuning lifting device is activated simultaneously to perform millisecond-level dynamic compensation on the vertical height of each flexible conveyor unit, so as to offset the gap changes between the conveyor belt and the support surface caused by local deformation of the human body or unevenness of the bed surface, and ensure a close fit throughout the process. Once the transfer is complete and the human body has been fully transferred to the target bed, the AGV automatically executes the evacuation path planning and returns to the standby area.

2. The AGV intelligent transfer control and scheduling method for personnel body transport according to claim 1, characterized in that, Obtain spatial pose relationship data between the target bed and the current docking position of the AGV, including: By deploying a binocular vision sensor array on top of the AGV, the structural feature points of the target bed are reconstructed in three dimensions, and the spatial coordinates of the bed edge, bed leg connection points and bed surface marking patterns are extracted. By installing laser ranging sensor groups at the four corners of the AGV chassis, the distance between the AGV and the surrounding walls, equipment and ground markings is measured. Combined with the pre-stored indoor digital twin map, the six-degree-of-freedom pose of the AGV itself in the global coordinate system is calculated. The coordinates of the feature points of the target bed are transformed with the global pose of the AGV, and the relative pose deviation between the two is calculated.

3. The AGV intelligent transfer control and scheduling method for personnel body transport according to claim 2, characterized in that, Drive the AGV chassis to perform omnidirectional movement and attitude adjustment, including: The AGV chassis adopts a Mecanum wheel omnidirectional drive configuration, with each of the four Mecanum wheels driven by an independent servo motor. Each servo motor is equipped with a high-resolution encoder and a closed-loop current feedback controller. Based on the calculated three-dimensional coordinate offset, translation speed commands are generated in the X and Y axis directions; Based on the pitch angle deviation and roll angle deviation, the speed difference between the front and rear wheel sets and the left and right wheel sets is controlled to achieve micro-angle tilt adjustment of the chassis around the X and Y axes; Based on the heading angle deviation, the rotation direction and speed of the four Mecanum wheels are controlled by differential steering to achieve precise homing around the Z-axis; Throughout the adjustment process, the actual attitude of the chassis is monitored in real time by the inertial measurement unit and closed-loop feedback correction is performed with the target attitude until the pose error is less than a preset threshold. The preset threshold is that the horizontal position error is no more than three millimeters, the height error is no more than two millimeters, and the angle error is no more than 0.5 degrees.

4. The AGV intelligent transfer control and scheduling method for personnel body transport according to claim 3, characterized in that, In the multi-segment flexible transfer mechanism, there are seven flexible transfer units arranged sequentially from head to feet along the longitudinal direction of the human body. Each segment is 250 mm long and 600 mm wide. Adjacent segments are connected by elastic hinges, allowing a maximum relative deflection of 15 degrees in the vertical direction. Each flexible conveyor unit's conveyor belt module is made of polyurethane composite material with a micron-level anti-slip texture on the surface. Its static friction coefficient is not less than 0.8, and the dynamic friction coefficient fluctuation range is controlled within ±5%. Each conveyor module is driven by a pair of synchronous pulleys, which are connected to a brushless DC servo motor via a planetary reducer. The servo motor has a control cycle of one millisecond and a speed adjustment resolution of 0.1 revolutions per minute.

5. The AGV intelligent transfer control and scheduling method for personnel body transport according to claim 4, characterized in that, The pressure distribution sensing array consists of 320 piezoresistive sensing units arranged in a 16-row, 20-column matrix. Each sensing unit has a sensing area of ​​25 square millimeters, a range of 0 to 500 Newtons, and a sampling frequency of 200 Hz. The contact pressure distribution map is noise-suppressed using a sliding window filtering algorithm and local voids are filled using morphological closing operations to form a continuous pressure field image. Based on the pressure field image, the two-dimensional projected coordinates of the body's center of gravity are calculated using a centroid-weighted algorithm, as shown in the formula: ， ； in, The pressure value of the sensing unit in the i-th row and j-th column is... and These are the horizontal and vertical coordinates of the corresponding sensing unit in the coordinate system of the transfer platform.

6. The AGV intelligent transfer control and scheduling method for transporting human bodies according to claim 5, characterized in that, Dynamically allocate the operating speed and direction of each flexible transmission unit, including: The transfer platform is divided into several speed control zones, and each control zone corresponds to a flexible transmission unit. Using the projection point of the body's center of gravity as a reference, calculate the longitudinal offset distance of each control zone relative to the center of gravity; Set the main teleport speed The reference speed for the transfer mission is 10 to 50 millimeters per second. For the k-th flexible transmission unit, its operating speed Calculate using the following formula: ; in, The distance of the centroid offset of the k-th segment. The maximum effective offset length is set to 800 millimeters. This is the speed compensation coefficient, with a value of 0.

2. when When positive, it indicates that the segment is located behind the center of gravity and the speed is slightly higher than the reference speed; when... When the value is negative, it indicates that the object is in front and its speed is slightly lower than the reference speed, thereby achieving synchronous movement between the overall transmission surface and the human body.

7. The AGV intelligent transfer control and scheduling method for transporting human bodies according to claim 6, characterized in that, The local fine-tuning lifting device uses a linear actuator driven by a voice coil motor, with a stroke range of ±20 mm, a response time of five milliseconds, and a positioning accuracy of 0.05 mm. Each flexible conveyor unit is equipped with three linear actuators arranged in a triangle at its bottom, which enable fine-tuning of the attitude of the local plane through three-point support; The fine-tuning command is determined by the local pressure gradient in the pressure distribution map. When the pressure gradient in a certain area exceeds the threshold of 100 Newtons per square meter, it is determined that there is local suspension or compression, and the corresponding area's rise and fall compensation is immediately activated to restore the pressure gradient to a safe range.

8. The AGV intelligent transfer control and scheduling method for transporting human bodies according to claim 7, characterized in that, The evacuation path planning is based on the improved A* algorithm, which combines real-time obstacle detection data and bed occupancy status to generate a collision-free, shortest-time reversal trajectory. During the evacuation process, the AGV keeps the transfer platform at its lowest storage height and disables all conveying functions, only enabling the chassis motion control and environmental perception modules.

9. An AGV intelligent transfer control system for transporting human bodies, characterized in that, include: The spatial pose perception module is used to acquire spatial pose relationship data between the target bed and the current docking position of the AGV; The chassis omnidirectional motion control module is used to drive the AGV chassis to perform omnidirectional movement and attitude adjustment to achieve initial docking with the target bed. The multi-segment flexible transfer execution module includes several flexible transmission units, a pressure distribution sensor array, and a local fine-tuning lifting device, used to perform human body transfer without shearing force. The pressure-speed coordinated control module is used to dynamically allocate the operating speed and direction of each flexible transmission unit based on the pressure distribution map; The vertical gap compensation module is used to activate the local fine-tuning lifting device according to the local pressure gradient, so as to achieve full contact between the conveying surface and the support surface. The task status management module is used to trigger the evacuation path planning and control the AGV to return to the standby area after the transfer is completed.

10. The AGV intelligent transfer control system for transporting human bodies according to claim 9, characterized in that, The spatial pose perception module further includes a binocular vision 3D reconstruction unit, a laser ranging fusion positioning unit, and a relative pose calculation unit. The chassis omnidirectional motion control module further includes a Mecanum wheel servo drive unit, an inertial attitude feedback unit, and a posture closed-loop correction unit. In the multi-segment flexible transfer execution module, the flexible transfer units are connected in series by elastic hinges, and each segment is independently configured with a servo motor, a conveyor belt, a pressure sensor array and a linear actuator driven by three voice coil motors. The pressure-speed coordinated control module has a built-in center of gravity calculation engine and speed allocation algorithm engine, and its real-time processing cycle is no more than ten milliseconds. The vertical clearance compensation module is equipped with a pressure gradient monitoring threshold judge and an actuator drive controller, and its response delay does not exceed eight milliseconds.

Images