Wheel-foot combined type picking robot and control method thereof

Abstract

The invention is suitable for the technical field of agricultural automation, and provides a wheel-foot combined type picking robot and a control method thereof.The robot comprises a wheel-foot dual-mode moving chassis, a sensing system and a working module; the wheel-foot dual-mode mobile chassis adopts a four-foot bionic layout, each foot end is wrapped by a foot end shell, wheels are arranged at the bottoms of the foot end shells, and hub driving motors are arranged in the foot end shells; the sensing system comprises a laser radar, a depth camera and a control unit, the laser radar is installed at the front end of the wheel-foot dual-mode mobile chassis, and the depth camera is installed at the execution end of the working module; and the working module is mounted on the wheel-foot dual-mode mobile chassis. Through an innovative wheel-foot composite structure and a layered cooperative intelligent control strategy, the picking robot achieves full-process automation from efficient movement to stable operation in an unstructured agricultural environment, and the picking robot has extremely high application value and popularization prospects.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of agricultural automation, and particularly relates to a wheel-foot combined picking robot and a control method thereof. BACKGROUND

[0002] With the increasing demand for automation and intelligence in modern agriculture, agricultural robots, especially picking robots, have become a research hotspot. However, the agricultural working environment such as orchard and mountainous area usually has unstructured characteristics, such as rugged ground, variable slope, existence of gully or obstacle, etc.

[0003] The existing agricultural robots are mainly divided into three types of wheel type, track type and foot type. The wheel type robot has high moving speed and energy efficiency, but has poor adaptability in rugged terrain and is easy to slip and overturn. The track type robot has improved obstacle crossing ability, but is inconvenient to turn and causes great damage to the ground. The foot type robot (such as quadruped robot) has excellent terrain adaptability and stability, and can cross obstacles, but has slow moving speed and high energy consumption, and is not suitable for large-scale and high-efficiency field operation.

[0004] Therefore, how to design a picking platform that can move quickly on flat road like a wheel type robot and walk stably on complex terrain like a foot type robot, and realize efficient and stable picking in all terrains, is a technical problem to be solved in the current field. SUMMARY

[0005] The purpose of the embodiment of the present application is to provide a wheel-foot combined picking robot and a control method thereof, which aims to solve the problems proposed in the background.

[0006] The embodiment of the present application is realized as follows: a wheel-foot combined picking robot, comprising a wheel-foot dual-mode moving chassis, a perception system and a working module; The wheel-foot dual-mode moving chassis adopts a quadruped bionic layout, each foot end is wrapped by a foot end shell, the bottom of the foot end shell is provided with a wheel, the inside of the foot end shell is provided with a hub drive motor for driving the wheel to rotate, and the hub drive motor has a servo locking function, when it is needed to switch to the foot type mode, the controller can make the hub drive motor enter a large damping or locking state, so as to fix the wheel and make the foot end become a stable support point; The perception system comprises a laser radar, a depth camera and a control unit, the laser radar is installed at the front end of the wheel-foot dual-mode mobile chassis for long-distance environment scanning and SLAM mapping, the depth camera is installed at the execution end of the working module for close-range fruit identification, positioning and fine navigation, the control unit is arranged inside the wheel-foot dual-mode mobile chassis for processing sensor data and executing decision and control algorithms, and the control unit is internally provided with an inertial measurement unit (IMU) for acquiring real-time posture information of the robot such as pitch and roll; The working module is installed on the wheel-foot dual-mode mobile chassis for picking work.

[0007] In a further technical solution, the working module comprises a six-axis mechanical arm, the six-axis mechanical arm is installed on the wheel-foot dual-mode mobile chassis, a gripper end effector for picking fruits is installed at the end of the six-axis mechanical arm, and the depth camera is also installed at the end of the six-axis mechanical arm.

[0008] In a further technical solution, a scissor lifting mechanism is arranged on the wheel-foot dual-mode mobile chassis, and the six-axis mechanical arm is installed on the wheel-foot dual-mode mobile chassis through the scissor lifting mechanism.

[0009] In a further technical solution, the prediction balance control algorithm of the wheel-foot composite picking robot comprises the following steps: First, the data of the laser radar and the IMU are fused by a terrain data processing module in the control unit; The processed data is divided into two paths: one path is used for ground concave-convex analysis to evaluate the flatness of the area to be stepped on in front; the other path is used for balance state evaluation to judge the inclination degree and stability of the current vehicle body; The two paths of information are jointly input into a four-leg adjustment algorithm, which is based on a prediction control model to calculate the joint adjustment amount of each of the four legs required to maintain or restore balance; Finally, the joint control calculation module in the control unit converts these adjustment amounts into specific control instructions for each leg servo motor to drive the robot to complete posture adjustment; The whole process forms a closed loop, and the actual posture of the wheel-foot dual-mode mobile chassis is fed back to the control unit through the IMU for continuous and stable balance control.

[0010] In a further technical solution, the 2.5D terrain mapping of the wheel-foot composite picking robot comprises the following steps: First, after the original point cloud data collected by the laser radar is preprocessed and the coordinates are unified, the ground is segmented by an algorithm to extract pure terrain surface point cloud; Subsequently, the control unit performs grid processing and elevation interpolation in parallel, and fuses the two to generate a DEM (digital elevation model); Finally, by analyzing the DEM, the key terrain features including slope and ruggedness are extracted to form a 2.5D terrain map; wherein the ruggedness can be quantified by using terrain ruggedness index (TRI) and other indicators.

[0011] Further technical solutions, the multi-mode motion control steps of the wheel-foot combined picking robot are as follows: The motion mode decision module in the control unit receives the calculated TRI value, slope information and real-time obstacle detection results from the 2.5D terrain map; When the robot perceives that the terrain in front is flat, i.e. TRI < 0.01, and there is no obstacle, it switches to the wheeled high-speed mode; in this mode, the hub drive motor drives the wheels to move the robot, and the leg joints remain in a fixed posture; When it is perceived that the terrain is slightly uneven, i.e. 0.01≤TRI<0.05, it switches to the wheeled + active suspension mode; in this mode, the robot is still driven by the wheels, but the leg joints will be stretched and contracted in a small range and at a high frequency according to the predicted balance control algorithm to absorb the impact of the ground, playing the role of active suspension; When it is perceived that the terrain is rugged or needs to cross obstacles, i.e. TRI≥0.05, it switches to the foot-type slow-speed mode; in this mode, the hub drive motor enters the locking state to fix the wheels, and the robot moves through the complex area by alternately lifting and moving the four legs in a gait; When navigating to the predetermined picking point, the robot enters the leveling mode, the four legs are stretched to support and accurately adjust the leg length, and the upper platform is adjusted to be completely horizontal to provide a zero-inclination reference for the mechanical arm operation.

[0012] In any mode, if the sensor detects an emergency obstacle, the emergency obstacle avoidance mode is triggered, and the robot will immediately stop or perform a round action.

[0013] Further technical solutions, the cooperative precise positioning steps of the wheel-foot combined picking robot are as follows: The SLAM module of the laser radar is responsible for constructing a global map in a large range and providing pose estimation; at the same time, the visual odometry (V-O) module of the depth camera provides relative motion estimation at close range; The pose data from the IMU, the laser radar pose and the visual odometry pose are jointly input into the extended Kalman filter (EKF) fusion module in the control unit; The EKF fusion module combines the data of various sensors to suppress noise and drift, and finally outputs a 6-degree-of-freedom (6DoF) pose for the six-axis mechanical arm to perform coordinate transformation and precise positioning of the picking target.

[0014] Beneficial effects: The present invention provides a wheel-legged composite harvesting robot and its control method, the beneficial effects of which are as follows: (1) High terrain adaptability and high efficiency: By combining the wheel-foot composite structure with multi-mode motion control strategy, the robot can move at high speed on flat ground and switch to stable walking on foot on rugged ground, thus taking into account both movement efficiency and terrain adaptability.

[0015] (2) Strong active balancing and operational stability: Through predictive balancing control algorithms, the robot can perceive terrain undulations in real time and adjust its posture to ensure the platform remains stable during dynamic movement. At the picking point, the platform leveling function provides a stable operating benchmark for the robotic arm, significantly improving the picking success rate.

[0016] (3) Precise positioning and reliable navigation: By integrating the global mapping capability of lidar and the local fine perception capability of depth camera, high-precision and robust positioning is achieved in orchard environments with weak or missing GPS signals, which provides a guarantee for the precise operation of robotic arm.

[0017] (4) The algorithm is complete and implementable: It provides a detailed calculation process and parameter calibration method to ensure that those skilled in the art can implement the present invention based on the description, thus solving the problem of insufficient algorithm disclosure. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the structure of a wheel-legged composite harvesting robot provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the foot end of a wheel-foot hybrid harvesting robot provided in an embodiment of the present invention; Figure 3 Block diagram of the predictive balance control algorithm; Figure 4 Flowchart for 2.5D terrain mapping; Figure 5 This is a state transition diagram for multi-mode motion control; Figure 6 This is a flowchart of the collaborative localization process using lidar and depth cameras.

[0019] In the attached diagram: 1-Dual-mode mobile chassis with wheels and feet; 2-Control unit; 3-Scissor lifting mechanism; 4-Six-axis robotic arm; 5-Depth camera; 6-Grip end effector; 7-LiDAR; 8-Foot end housing; 9-Wheel hub drive motor; 10-Wheel. Detailed Implementation

[0020] In order to make the objects, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application.

[0021] The specific implementation of the present application is described in detail below in combination with specific embodiments.

[0022] As shown in Figure 1 and Figure 2 , a wheel-foot combined picking robot provided by an embodiment of the present application comprises a wheel-foot dual-mode mobile chassis 1, a perception system and a working module; The wheel-foot dual-mode mobile chassis 1 adopts a four-foot bionic layout, each foot end is wrapped by a foot end shell 8, the bottom of the foot end shell 8 is provided with a wheel 10, and the inside of the foot end shell 8 is provided with a hub drive motor 9 for driving the wheel 10 to rotate, and the hub drive motor 9 has a servo locking function, when it is needed to switch to the foot mode, the controller can make the hub drive motor 9 enter a large damping or locking state, so as to fix the wheel 10 and make the foot end become a stable support point; The perception system comprises a laser radar 7, a depth camera 5 and a control unit 2, the laser radar 7 is installed at the front end of the wheel-foot dual-mode mobile chassis 1 and is used for long-distance environment scanning and SLAM mapping; the depth camera 5 is installed at the execution end of the working module and is used for close-range fruit identification, positioning and fine navigation; the control unit 2 is arranged inside the wheel-foot dual-mode mobile chassis 1 and is used for processing data of various sensors (i.e. the laser radar 7, the depth camera 5 and the IMU) and executing decision and control algorithms; the control unit 2 is built-in with an IMU and is used for acquiring real-time attitude information of the robot such as pitch and roll; The working module is installed on the wheel-foot dual-mode mobile chassis 1 and is used for picking work.

[0023] As shown in Figure 1 , as a preferred embodiment of the present application, the working module comprises a six-axis mechanical arm 4, the six-axis mechanical arm 4 is installed on the wheel-foot dual-mode mobile chassis 1, a gripper end effector 6 for grabbing fruits is installed at the end of the six-axis mechanical arm 4, and the depth camera 5 is also installed at the end of the six-axis mechanical arm 4.

[0024] As shown in Figure 1 , as a preferred embodiment of the present application, a scissor lifting mechanism 3 is arranged on the wheel-foot dual-mode mobile chassis 1, and the six-axis mechanical arm 4 is installed on the wheel-foot dual-mode mobile chassis 1 through the scissor lifting mechanism 3, and the vertical working range of the mechanical arm can be expanded through the scissor lifting mechanism 3.

[0025] As shown in Figure 3As a preferred embodiment of the present application, the predictive balance control algorithm steps of the wheel-foot combined picking robot are as follows: First, the data of the laser radar 7 and the IMU are fused by the terrain data processing module in the control unit 2. The processed data is divided into two paths: one path performs ground concave-convex analysis to evaluate the flatness of the area to be stepped on in front; the other path performs balance state evaluation to judge the inclination degree and stability of the current vehicle body.

[0026] The two paths of information are jointly input into the four-leg adjustment algorithm, which is based on a predictive control model to calculate the joint adjustment amount of each of the four legs required to maintain or restore balance.

[0027] Finally, the joint control calculation module in the control unit 2 converts these adjustment amounts into specific control instructions for each leg servo motor to drive the robot to complete posture adjustment. The whole process forms a closed loop, and the actual posture of the wheel-foot dual-mode mobile chassis 1 is fed back to the control unit 2 through the IMU to realize continuous and stable balance control.

[0028] As shown in Figure 4 As a preferred embodiment of the present application, the 2.5D terrain mapping steps of the wheel-foot combined picking robot are as follows: In order to realize intelligent path planning and mode switching, the robot needs to have an accurate understanding of the surrounding terrain.

[0029] First, after pre-processing (filtering and noise reduction) and coordinate unification (sensor to world coordinate system) of the original point cloud data collected by the laser radar 7, ground segmentation is performed through an algorithm to extract pure terrain surface point cloud.

[0030] Subsequently, the control unit 2 performs grid processing (divides the terrain into grids and calculates the average elevation of each grid) and elevation interpolation (fills in data in sparse areas of the point cloud) in parallel, and fuses the two to generate a DEM.

[0031] Finally, by analyzing the DEM, key terrain features such as slope and ruggedness are extracted to form a 2.5D terrain map for subsequent motion planning and mode decision modules. The ruggedness can be quantified using TRI, etc.

[0032] As shown in Figure 5 As a preferred embodiment of the present application, the multi-mode motion control steps of the wheel-foot combined picking robot are as follows: The motion mode decision module in the control unit receives the calculated TRI value, slope information from the 2.5D terrain map, and real-time obstacle detection results.

[0033] When the robot perceives that the terrain ahead is flat (e.g., TRI < 0.01) and free of obstacles, it switches to the wheeled high-speed mode. In this mode, the hub drive motor 9 drives the wheels 10 to move the robot quickly, and the leg joints remain in a fixed posture, resulting in low energy consumption and high efficiency.

[0034] When the robot perceives a slightly uneven terrain (e.g., 0.01 ≤ TRI < 0.05), it switches to the wheeled + active suspension mode. In this mode, the robot is still driven by the wheels 10, but the leg joints will stretch and contract in a small range and at a high frequency according to the predictive balance control algorithm, absorbing ground impact and acting as an active suspension to ensure the platform's stability.

[0035] When the robot perceives a rugged terrain or needs to cross obstacles (e.g., TRI ≥ 0.05), it switches to the legged slow-speed mode. In this mode, the hub drive motor 9 enters a locked state to fix the wheels 10, and the robot moves through complex areas slowly by alternately lifting and moving the four legs in a stable gait.

[0036] When navigating to the predetermined picking point, the robot enters the leveling mode, with the four legs stretched to support and accurately adjust the leg length, adjusting the upper platform to be completely horizontal and providing a zero-inclination reference for the robotic arm operation.

[0037] In any mode, if the sensor detects an emergency obstacle, it will trigger the emergency obstacle avoidance mode, and the robot will immediately stop or perform a detour action.

[0038] As shown in Figure 6 As a preferred embodiment of the present application, the cooperative precise positioning steps of the wheel-foot combined picking robot are as follows: To achieve precise picking, the robot must know its precise position and posture in the orchard. Therefore, a multi-sensor fusion positioning scheme is adopted. The SLAM module of the laser radar 7 is responsible for constructing a global map in a large range and providing a rough, low-drift pose estimation. At the same time, the visual odometry (V-O) module of the depth camera 5 provides high-frequency, high-precision relative motion estimation at close range. The attitude data from the IMU, the laser radar pose, and the visual odometry pose are jointly input into the extended Kalman filter (EKF) fusion module in the control unit 2. The EKF fusion module combines the advantages of each sensor, suppresses noise and drift, and finally outputs a high-precision, high-frequency (> 100 Hz) 6-DOF pose for the six-axis robotic arm 4 to perform coordinate transformation and precise positioning of the picking target.

[0039] As a preferred embodiment of the present application, the predictive balance control algorithm has the following specific calculation process: 1) Ground concave-convex analysis calculation: For the front terrain point cloud data, least squares method is used for surface fitting; The surface equation is: ; By solving the least squares problem ; Finally, the surface coefficient matrix is obtained: ; Where: is the three-dimensional coordinate of the point in the world coordinate system in the terrain point cloud, and are horizontal coordinates, is the elevation coordinate. is the coordinate of the th terrain point in the world coordinate system. is the surface fitting coefficient, obtained by least squares method. is the surface coefficient matrix, containing all the fitting coefficients.

[0040] 2) Curvature calculation: Based on the fitted surface, the average curvature is calculated: ; The first and second order partial derivatives involved in the formula are calculated from the terrain fitting surface equation , and the specific definitions are as follows: ; ; ; ; ; Where, is the average curvature, used to describe the macroscopic bending degree of the terrain surface. and are the first order partial derivatives of the terrain surface in and directions, which physically represent the slope of the point on the terrain surface. 、 and are the second order partial derivatives of the terrain surface, which describe the rate of change of the slope and are the core components for calculating the curvature.

[0041] 3) Zero Moment Point (ZMP) stability calculation: ; ; Where: is the ground reaction force vector of the th leg. 、 and is the coordinate of the i-th leg foot in the robot coordinate system. is the coordinate of the i-th leg foot in the robot coordinate system. , and are the components of the ground reaction force in the x, y and z directions. is the coordinate of the i-th leg foot in the robot coordinate system. is the coordinate of the i-th leg foot in the robot coordinate system. is the coordinate of the i-th leg foot in the robot coordinate system. is the coordinate of the i-th leg foot in the robot coordinate system.

[0042] 4) Four-legged adjustment amount optimization calculation: using the quadratic programming method: ; where: is the joint angle adjustment amount vector; is the ground reaction force adjustment amount vector; is the joint adjustment weight matrix, is the ground reaction force adjustment weight matrix, calibrated by experiment; is the optimization objective function.

[0043] The constraint conditions are as follows: Dynamics constraint: , which is an approximation of the robot dynamics equation based on the current state linearization.

[0044] Stability constraint: ; where the support polygon refers to the smallest convex hull area formed by the position points of all feet in the landing state, ensuring that is located in this area is a necessary condition for dynamic stability of the robot; is the mass matrix; is the damping matrix; is the stiffness matrix; Joint limit: ; where, is the current joint angle vector; and are the minimum and maximum limit values of the joint angle.

[0045] As a preferred embodiment of the present application, the specific calculation process of 2.5D terrain mapping is as follows: 1) Ground segmentation algorithm: RANSAC plane fitting is used: For point cloud P, randomly select 3 points to calculate the plane equation: ; Calculate the distance of all points to the plane : ; Inlier set: ; where: is the plane equation coefficient. is the distance from the th point to the plane. is the distance threshold (m) for RANSAC to determine inliers, usually determined by experiment or experience.

[0046] Repeat iterations, select the plane with the most inliers as the ground.

[0047] 2) DEM generation: Grid division: divide the terrain into 0.1m x 0.1m grids; Elevation interpolation: adopt inverse distance weighting method (IDW) ; where: is the weight coefficient, inversely proportional to the square of the distance. is the distance from the th data point to the center of the grid. is the elevation interpolation result at the center of the grid, calculated based on the point cloud data in the world coordinate system.

[0048] 3) TRI calculation: for each grid cell, calculate the elevation change with its 8 neighborhood grids to evaluate the terrain ruggedness: ; The specific calculation process is as follows: ; ; Then:

[0049] ; where: is the index coordinate of the current grid cell; is the elevation value of the current grid cell ; is the th to th, the elevation value of the 8 neighborhood grid cells of the current grid cell; to is the absolute value of the elevation difference between the current grid cell and each neighborhood grid cell; is the terrain ruggedness index of the current grid cell . ​

[0050] 4) Slope calculation: Slope is calculated based on DEM data: ; Where: Rate of elevation change in direction ; Rate of elevation change in direction ; and Size of the grid in and directions, here ; is the slope value.

[0051] As a preferred embodiment of the invention, the multi-mode motion control threshold is calculated as follows: 1) TRI threshold determination: Based on the physical parameters of the robot: Wheel mode can pass over height: ; Active suspension compensation range: ; Foot mode crossing height: ; Where: is the wheel mode passable height, determined by the wheel radius ; is the maximum compensation range of the active suspension system; is the maximum crossing height of the foot mode; is the thigh length, ; is the calf length, ; and are the hip and knee joint angles, here .

[0052] 2) Mode switching logic: if TRI < 0.01 and no obstacles: Wheel high-speed mode: = 1.0 m / s elif 0.01 ≤ TRI < 0.05: Wheel + active suspension: v = 0.5 m / s, active suspension frequency f = 2-5 Hz elif TRI ≥ 0.05: Foot slow-speed mode: v = 0.2 m / s, gait cycle T = 2 s in: This is the maximum speed in wheel mode; This represents the movement speed in the current mode. For active suspension frequency; The gait cycle is in the foot-based mode.

[0053] The TRI threshold mentioned above is a typical value calculated based on the specific parameters of the robot in this embodiment. In practical applications, it can be precisely calibrated through experiments according to the physical size, weight distribution, and specific working environment of different robots.

[0054] In a preferred embodiment of the present invention, the calculation process for cooperative precise positioning is as follows: 1) EKF state model: State vector: ; Process model: ; Observation model: ; in: The coordinates of the robot's position in the world coordinate system are in meters (m). This represents the robot's velocity components in the world coordinate system, expressed in meters per second (m / s). Let be the robot's pose quaternion; To control the input vector; The noise is a process noise that follows a Gaussian distribution. ; To observe the noise, it follows a Gaussian distribution. ; The process noise covariance matrix is... To observe the noise covariance matrix.

[0055] This is the system state transition function (which may be nonlinear); The observation function maps the state to the sensor measurement space.

[0056] 2) EKF prediction steps: ; ; in: For a moment Prior state estimation, To estimate the covariance matrix a priori, Let be the Jacobian matrix of the state transition function.

[0057] 3) EKF update steps: ; ; ; wherein: is the Kalman gain matrix; is the Jacobian matrix of the observation function; is the identity matrix.

[0058] 4) Sensor data time synchronization: Interpolation method is adopted in combination with hardware time stamp to solve the time stamp difference of different sensor data collection and transmission, and ensure the time consistency of data fusion: For the time , the data required, if : ; wherein: is the interpolation data at time . and are the sensor data of adjacent time stamps and .

[0059] As a preferred embodiment of the present application, the kinematics calculation is as follows: 1) Inverse kinematics calculation: Given the foot end position , the joint angle is calculated:

[0060] ; ; ; ; ; ; ; wherein: is the coordinate of the foot end in the hip joint coordinate system; is the yaw angle of the hip joint; is the thigh pitch angle; is the knee pitch angle; is the projection distance of the foot end on the horizontal plane; is the straight line distance from the hip joint to the foot end; is the included angle between the foot end vector and the horizontal plane; is the included angle between the thigh and the hip joint-foot end connecting line.

[0061] 2) Jacobian matrix calculation: For speed mapping and force control: ; Joint speed to foot end speed: ; Foot end force to joint torque: ; Where: J is the Jacobian matrix of the mechanical leg, which is a matrix; v is the foot end linear velocity vector; is the joint angular velocity vector; is the joint torque vector; is the force vector acting on the foot end.

[0062] As a preferred embodiment of the present application, the algorithm parameter calibration step is as follows: Based on the actual mechanical parameters of the robot: thigh length ; calf length ; wheel radius ; total mass ; single leg mass ; motor maximum torque ; Control parameters calibrated through experiments: Joint adjustment weight matrix: ; Ground reaction force adjustment weight matrix: ; EKF process noise covariance : ; EKF observation noise covariance : ; Verify algorithm performance through simulation and experiment: Balance control accuracy: on a 10° slope, platform inclination control error < 0.5°; Positioning accuracy: absolute positioning error < 0.05m, relative positioning error < 0.01m; Mode switching response: from detecting terrain changes to completing mode switching < 0.5s; Picking success rate: in the leveling mode, the success rate of the robotic arm picking > 95%.

[0063] In summary, through the innovative wheel-foot composite structure and hierarchical collaborative intelligent control strategy, the picking robot successfully realizes full-process automation from efficient movement to stable operation in the unstructured agricultural environment, effectively improving the operation efficiency and stability in the unstructured environment.

[0064] The above description is only the preferred embodiment of the present application, and is not intended to limit the present application. Any modification, equivalent replacement and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A wheel-foot hybrid picking robot, characterized in that, The wheel-foot dual-mode mobile chassis, a perception system, and a working module are included. The wheel-foot dual-mode mobile chassis adopts a quadruped bionic layout, each foot end is wrapped by a foot end shell, the bottom of the foot end shell is provided with a wheel, and the inside of the foot end shell is provided with a hub drive motor for driving the wheel to rotate, the hub drive motor has a servo locking function, when it is needed to switch to the foot mode, the controller of the hub drive motor will make the hub drive motor enter a large damping or locking state, so as to fix the wheel and make the foot end become a stable support point. The perception system includes a laser radar, a depth camera, and a control unit, the laser radar is installed at the front end of the wheel-foot dual-mode mobile chassis, is used for long-distance environment scanning and SLAM mapping, the depth camera is installed at the execution end of the working module, is used for close-range fruit identification, positioning, and fine navigation, the control unit is arranged inside the wheel-foot dual-mode mobile chassis, is used for processing sensor data and executing decision and control algorithms, and the control unit is internally provided with an IMU, which is used for acquiring real-time posture information of the robot. The working module is installed on the wheel-foot dual-mode mobile chassis and is used for picking work.

2. The wheel-foot hybrid picking robot according to claim 1, characterized in that, The working module includes a six-axis mechanical arm, the six-axis mechanical arm is installed on the wheel-foot dual-mode mobile chassis, a gripper end effector for picking fruits is installed at the end of the six-axis mechanical arm, and the depth camera is also installed at the end of the six-axis mechanical arm.

3. The wheel-foot hybrid picking robot according to claim 2, characterized in that, A scissor lifting mechanism is arranged on the wheel-foot dual-mode mobile chassis, and the six-axis mechanical arm is installed on the wheel-foot dual-mode mobile chassis through the scissor lifting mechanism.

4. The wheel-foot hybrid picking robot according to any one of claims 1-3, characterized in that, The prediction balance control algorithm of the wheel-foot composite picking robot is as follows: Firstly, the data of the laser radar and the IMU are fused by a terrain data processing module in the control unit. The processed data is divided into two paths: one path is used for ground concave-convex analysis to evaluate the flatness of the area to be stepped on in front; the other path is used for balance state evaluation to judge the inclination degree and stability of the current vehicle body. The two paths of information are jointly input into a four-leg adjustment algorithm, which is based on a prediction control model to calculate the joint adjustment amount of each of the four legs required to maintain or restore balance. Finally, the joint control calculation module in the control unit converts these adjustment amounts into specific control instructions for each leg servo motor to drive the robot to complete the posture adjustment. The whole process forms a closed loop, the actual posture of the wheel-foot dual-mode mobile chassis is fed back to the control unit through the IMU for continuous balance control.

5. The wheel-foot hybrid picking robot according to claim 4, characterized in that, The 2.5D terrain mapping steps of the wheel-foot composite picking robot are as follows: Firstly, after the original point cloud data collected by the laser radar is preprocessed and the coordinates are unified, the ground surface point cloud is extracted through an algorithm; Subsequently, the control unit performs grid processing and elevation interpolation in parallel, and fuses the two to generate a DEM; Finally, by analyzing the DEM, the key terrain features including the slope and ruggedness are extracted to form a 2.5D terrain map; the ruggedness is quantified by the TRI index, which can be adjusted according to the specific application scenario and the physical parameters of the robot.

6. The wheel-foot hybrid picking robot according to claim 5, wherein, The multi-mode motion control steps of the wheel-foot hybrid picking robot are as follows: The motion mode decision module in the control unit receives the calculated TRI value, slope information and real-time obstacle detection results from the 2.5D terrain map; When the robot perceives that the terrain ahead is flat, i.e. TRI < 0.01, and there is no obstacle, it switches to the wheeled high-speed mode; in this mode, the wheel hub drive motor drives the wheels to move the robot, and the leg joints remain in a fixed posture; When it is perceived that the terrain is slightly uneven, i.e. 0.01 ≤ TRI < 0.05, it switches to the wheeled + active suspension mode; in this mode, the robot is still driven by the wheels, but the leg joints will stretch and contract in a small range and at a high frequency according to the predicted balance control algorithm to absorb the impact of the ground, playing the role of active suspension; When it is perceived that the terrain is rugged or needs to cross obstacles, i.e. TRI ≥ 0.05, it switches to the foot-type slow-speed mode; in this mode, the wheel hub drive motor enters the locking state to fix the wheels, and the robot moves through the complex area by alternately lifting and moving the four legs in a gait; When navigating to the predetermined picking point, the robot enters the leveling mode to adjust the upper platform to be completely horizontal, providing a zero-inclination reference for the mechanical arm operation; In any mode, if the sensor detects an emergency obstacle, the emergency obstacle avoidance mode will be triggered, and the robot will immediately stop or perform a roundabout action.

7. The wheel-foot hybrid picking robot according to claim 6, characterized in that, The cooperative fine positioning steps of the wheel-foot hybrid picking robot are as follows: The SLAM module of the laser radar is responsible for constructing a global map in a wide range and providing pose estimation; at the same time, the visual odometry (V-O) module of the depth camera provides relative motion estimation at close range; The pose data from the IMU, the laser radar pose and the visual odometry pose are jointly input into the EKF fusion module in the control unit; The EKF fusion module combines the data of various sensors to suppress noise and drift, and finally outputs a 6-degree-of-freedom pose for the six-axis mechanical arm to perform coordinate transformation and fine positioning of the picking target.

Images