A somatic robot for intelligent picking of warehouse materials and a control method thereof

Abstract

A kind of embodied robot for warehousing material intelligent picking and its control method, including embodied component and the mobile chassis for driving embodied component to move, lifting assembly for controlling embodied component is installed on the lifting of embodied component;The mobile chassis includes the movement wheel group for driving embodied component to move, the obstacle avoidance navigation for avoiding obstacles, the radar assembly of joint calibration of environmental coordinates is cooperated with the head camera in the obstacle avoidance navigation and the visual sensor for identifying surrounding environment;Relative to prior art, by the head camera and radar assembly, a fusion map with high-layered obstacle information is constructed, and in cooperation with the setting of visual sensor, when the embodied robot moves under the action of movement wheel group, obstacles can be identified by radar assembly and visual sensor, and the route of the embodied robot is selected according to the fusion map, and the embodied robot after bypassing obstacles is re-routed according to the fusion map.

Description

Technical Field

[0001] This invention relates to the field of sorting robot technology, specifically to an embodied robot for intelligent picking of warehouse materials and its control method. Background Technology

[0002] An embodied robot is an intelligent robot system that has a physical body, can interact with the environment in real time through perception, make autonomous decisions and execute actions, and can continuously learn in the process of interaction. Among them, a sorting embodied robot refers to a robot that is unmanned, has autonomous decision-making capabilities, and can automatically go to a designated location to perform logistics sorting operations.

[0003] In existing technologies, embodied robots used for material sorting typically employ wheeled chassis, such as Astribot S1 from Stardust Intelligent and X1 from Hangcha Group. However, due to the height limitations of the robot's structure, they cannot grasp materials on shelves higher than 2 meters in factory warehouses, hindering the large-scale application of embodied logistics robots in practical scenarios. Furthermore, there is currently a lack of stable and precise control methods for the coordinated control of the chassis and robotic arm during shelf material sorting. Simultaneously, the operating space of the robotic arm on the shelf is limited by the space constraints of shelf columns and beams, making it prone to encountering singularities during grasping path planning, leading to grasping failures.

[0004] Chinese patent CN116037471A discloses a sorting robot and its sorting method, including a moving mechanism and a sorting mechanism and a detection mechanism disposed on the moving mechanism. The moving mechanism includes a moving component and a detector, with the detector disposed on the moving component. The sorting mechanism includes a mounting frame, a robotic arm assembly, and a gripping assembly. The mounting frame is connected to the moving component, the robotic arm assembly is disposed on the mounting frame, and the gripping assembly is connected to the robotic arm assembly. The mounting frame is also provided with multiple goods storage boxes to improve the sorting efficiency of goods. The detection mechanism detects the labels and characteristic information of the goods, with dual detection to improve the situation of goods sorting errors.

[0005] The sorting robot disclosed above sorts goods by pre-planning its movement path. However, when there are obstacles in the movement path of the sorting robot, the sorting robot will directly collide with the obstacles, affecting the normal sorting operation of the sorting robot. At the same time, there is a risk of damage to the sorting robot or the detection device caused by the obstacles. Summary of the Invention

[0006] The present invention aims to overcome the deficiencies in the prior art and provide an embodied robot and its control method for intelligent picking of warehouse materials, which has a wide range of operating scenarios, accurate material picking, and automatic obstacle avoidance.

[0007] To achieve the above-mentioned objectives, the present invention adopts the following technical solution: a embodied robot for intelligent picking of warehouse materials, comprising an embodied component and a mobile chassis for moving the embodied component. The embodied component is equipped with a lifting component for controlling the lifting and lowering of the embodied component and a rotary support for controlling the rotation of the embodied component. The embodied component includes a torso, a head camera mounted on the torso, and a gripping robotic arm mounted on the torso. The mobile chassis includes a set of wheels for moving the embodied component and an obstacle avoidance navigation system for avoiding obstacles. The obstacle avoidance navigation system includes a radar component that works in conjunction with the head camera to jointly calibrate environmental coordinates and a visual sensor for identifying the surrounding environment. The embodied robot also includes a controller for constructing a navigation route in real time based on environmental coordinates, planning the driving mode of the mobile chassis, and planning the gripping path of the gripping robotic arm.

[0008] As a preferred embodiment of the present invention, the radar assembly includes a front-end lidar installed at the front end of the moving chassis in the direction of movement and a rear-end lidar installed at the rear end of the moving chassis in the direction of movement. The vision sensor includes a front-end vision sensor corresponding to the front-end lidar and a rear-end vision sensor corresponding to the rear-end lidar. The front-end lidar and the rear-end lidar are diagonally arranged on the moving chassis.

[0009] As a preferred embodiment of the present invention, the movable wheel set includes omnidirectional wheels for supporting the movement of the movable wheel set, a first steering wheel and a second steering wheel for adjusting the movement direction of the movable chassis.

[0010] As a preferred embodiment of the present invention, the head camera is mounted on the top of the torso body, and the torso body is provided with a pitch motor and a rotation motor for adjusting the recognition angle of the head camera.

[0011] As a preferred embodiment of the present invention, the gripping robotic arm includes a left robotic arm and a right robotic arm symmetrically arranged on both sides of the torso body. The torso body is provided with two robotic arm bases that control the left robotic arm and the right robotic arm respectively. The left robotic arm is provided with a left gripper, and the right robotic arm is provided with a right gripper.

[0012] As a preferred embodiment of the present invention, the ends of the left and right robotic arms are each provided with a vision camera for recognizing the gripping states of the left and right grippers, respectively.

[0013] As a preferred embodiment of the present invention, the lifting assembly includes a lifting body connected to a vertically mounted mobile chassis, a body assembly installed on the top of the lifting body, and a temporary storage box corresponding to the gripping robotic arm on the lifting body.

[0014] As a preferred embodiment of the present invention, the lifting body is provided with a temporary storage platform for supporting the temporary storage box, and the lifting body is provided with a lifting rod for controlling the lifting of the temporary storage platform.

[0015] A control method for an embodied robot used for intelligent picking of warehouse materials, based on the embodied robot for intelligent picking of warehouse materials, includes the following steps:

[0016] Step S1: Map building. The radar component on the mobile chassis scans the environment to generate a near-ground global map, and then the head camera supplements the height information in the map to build a fused map with height-layered obstacle information.

[0017] Step S2: Plan the travel path and move. Based on the fused map, plan the travel path of the embodied robot to the shelf, and control the embodied robot to move through the moving wheel set until it reaches the set position next to the shelf.

[0018] Step S3: The spatial positions of multiple points of the embodied robot are calibrated. Using the moving wheel set as the reference coordinate, the coordinate transformation relationship between the gripping robot base, the head camera and the moving wheel set is obtained. After obtaining the coordinate transformation relationship, the changes of the head camera, the left robot arm and the right robot arm relative to the moving wheel set can be obtained by the changes of the motor codes carried by each component during the robot's operation.

[0019] Step S4: Adjust the relative position of the shelf and the embodied robot, turn the embodied component to face the shelf using the slewing support, and detect the relative distance between the embodied robot and the shelf using the head camera;

[0020] Step S5: Construct an electronic fence, identify the location of obstacles in the shelf using a head camera and radar components, and mark the obstacles at the reference coordinates in step S3;

[0021] Step S6: Identify the material, adjust the height of the body component through the lifting component, identify the shape of the material and its relative coordinates under the reference coordinates in step S3 through the head camera, and plan the grasping path of the grasping robot arm to avoid the electronic fence based on the electronic fence obtained in step S5.

[0022] Step S7: By calculating the condition number κ(J) during the grasping process of the gripping robot, check whether there are singularities in the grasping path. If there are no singularities, the controller guides the gripping robot to grasp according to the planned path; if there are singularities, the controller controls the moving wheel group to slightly translate the robot body and adjust the relative position of the robot body and the material. Then repeat steps 4 to 6 until there are no singularities.

[0023] in: , Let denote the 2-norm of the Jacobian matrix. The 2-norm of a matrix is ​​equal to the largest singular value of the matrix.

[0024] Step S8: After the gripping is completed, the gripping robotic arm retracts and approaches the torso body, rotates back to the initial orientation through the rotary support, and controls the lifting component to return the body component to the initial position;

[0025] Step S9: The position of the temporary storage bin is obtained through the head camera, the controller generates the material feeding path of the gripping robot arm, and places the material into the temporary storage bin according to the generated path;

[0026] Step S10: If multiple materials are picked up from the same bin on the shelf, repeat steps S6-S9. After completion, the robot will first move laterally away from the shelf and then move forward along the aisle to reach the next picking position to repeat the operation.

[0027] As a preferred embodiment of the present invention, during the movement of the embodied robot in step S2, obstacles affecting the movement of the embodied robot are detected in real time by the head camera and radar components. When an obstacle is encountered, the movement direction of the moving wheel group is adjusted by the controller until the embodied robot bypasses the obstacle. After bypassing the obstacle, the movement path is replanned until the designated position next to the shelf is reached.

[0028] As a preferred embodiment of the present invention, before the spatial position is calibrated in step S3, the embodied component is facing the direction of travel of the embodied robot, and the grasping robotic arm is in a retracted and folded state on the torso body.

[0029] Compared to existing technologies, this method constructs a fusion map with highly layered obstacle information using a head camera and radar components. Combined with the setup of visual sensors, the robot can identify obstacles as it moves with the help of the moving wheel assembly. The robot can then select a detour route based on the fusion map and replan its route after detouring the obstacles.

[0030] Simultaneously, with the help of the constructed fusion map, the coordinates of the shelves and materials are marked on the fusion map, and electronic fences are set according to the shelves to achieve the goal of preventing the gripping robot arm from colliding with the shelves while gripping the materials, thus ensuring the stable gripping of the gripping robot arm. Attached Figure Description

[0031] Figure 1 This is the front view of the present invention;

[0032] Figure 2 This is a side view of the present invention;

[0033] Figure 3 This is a bottom view of the present invention;

[0034] Figure 4 This is a diagram showing the usage state of the present invention;

[0035] Reference numerals: Body assembly 1, head camera 11, pitch motor 12, rotation motor 13, torso body 14, lifting assembly 2, lifting body 21, lifting rod 22, temporary storage platform 23, temporary storage bin 24, mobile chassis 3, mobile wheel set 31, omnidirectional wheel 32, first steering wheel 33, second steering wheel 34, gripping robotic arm 4, left robotic arm 41, right robotic arm 42, left gripper 43, right gripper 44, vision camera 45, robotic arm base 46, obstacle avoidance and navigation 5, radar assembly 6, front-end lidar 61, rear-end lidar 62, vision sensor 7, front-end vision sensor 71, rear-end vision sensor 72, controller 8, slewing support 9. Detailed Implementation

[0036] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0037] like Figures 1-4 As shown, an embodied robot for intelligent picking of warehouse materials includes an embodied component 1 and a mobile chassis 3 for moving the embodied component 1. The embodied component 1 is equipped with a lifting component 2 for controlling the lifting and lowering of the embodied component 1 and a rotary support 9 for controlling the rotation of the embodied component 1. The embodied component 1 includes a torso body 14, a head camera 11 mounted on the torso body 14, and a gripping robotic arm 4 mounted on the torso body 14. The mobile chassis 3 includes a set of moving wheels 31 for moving the embodied component 1 and an obstacle avoidance navigation 5 for avoiding obstacles. The obstacle avoidance navigation 5 includes a radar component 6 that works with the head camera 11 to jointly calibrate the environmental coordinates and a visual sensor 7 for identifying the surrounding environment. The embodied robot also includes a controller 8 for constructing a navigation route in real time based on the environmental coordinates, planning the driving mode of the mobile chassis 3, and planning the gripping path of the gripping robotic arm 4.

[0038] The mobile chassis 3, lifting assembly 2 and body assembly 1 are connected sequentially from bottom to top to form a body robot. The torso body 14 is used to install the gripping robotic arm 4 and the head camera 11. The mobile wheel set 31 drives the lifting assembly 2 and body assembly 1 to move synchronously.

[0039] The radar assembly 6 includes a front-end lidar 61 installed at the front end of the moving chassis 3 in the direction of movement and a rear-end lidar 62 installed at the rear end of the moving chassis 3 in the direction of movement. The vision sensor 7 includes a front-end vision sensor 71 corresponding to the front-end lidar 61 and a rear-end vision sensor 72 corresponding to the rear-end lidar 62. The front-end lidar 61 and the rear-end lidar 62 are diagonally arranged on the moving chassis 3.

[0040] With the front-end LiDAR 61 and the rear-end LiDAR 62 set diagonally, 360° blind-spot-free coverage is achieved, improving the reliability of positioning and obstacle avoidance, and optimizing cost and structure. Similarly, the front-end vision sensor 71 and the rear-end vision sensor 72 are also set diagonally to achieve 360° blind-spot-free coverage.

[0041] The movable wheel assembly 31 includes omnidirectional wheels 32 for supporting the movement of the movable wheel assembly 31, and a first steering wheel 33 and a second steering wheel 34 for adjusting the movement direction of the movable chassis 3.

[0042] The first steering wheel 33 and the second steering wheel 34 can be, but are not limited to, steering wheels, differential wheels, Mecanum wheels, etc., to meet the requirements of an omnidirectional movement wheel system structure.

[0043] The number of casters 32 is set according to actual needs. The first steering wheel 33 and the second steering wheel 34 are located in the middle of the mobile chassis 3, and the first steering wheel 33 and the second steering wheel 34 are symmetrically arranged along the center of the mobile chassis 3. Under the rotation of the first steering wheel 33 and the second steering wheel 34, the mobile chassis 3 can achieve 360° self-rotation in place. It can also realize the horizontal movement, diagonal movement and movement in any direction of the mobile chassis 3, which meets the needs of the mobile chassis 3 in dense storage environment. The first steering wheel 33 and the second steering wheel 34 are independently controlled by the controller 8 to control the rotation direction and rotation speed.

[0044] With the omnidirectional wheels 32 providing support, multiple omnidirectional wheels 32 can be installed at various corners of the mobile chassis 3. After the first steering wheel 33 and the second steering wheel 34 adjust the steering and speed, the omnidirectional wheels 32 act as passive wheels, rubbing against the ground to support the mobile chassis 3 and ensure the stable movement of the mobile chassis 3.

[0045] The outer ring of the mobile chassis 3 is equipped with anti-collision beams and anti-collision strips for protecting the internal components of the mobile chassis 3.

[0046] The head camera 11 is a 3D depth camera. The head camera 11 is mounted on the top of the torso body 14, and the torso body 14 is equipped with a pitch motor 12 and a rotation motor 13 for adjusting the recognition angle of the head camera 11.

[0047] The pitch motor 12 controls the vertical head-mounted camera 11 to tilt up or down, adjusting the pitch angle of the head-mounted camera 11 to achieve vertical tracking of the target and adjustment of the viewing angle.

[0048] The rotary motor 13 controls the left and right rotation of the head camera 11 in the horizontal direction, adjusts the horizontal angle of the head camera 11, and realizes panoramic scanning and horizontal tracking.

[0049] A base is installed on the top of the torso body 14, and a rotating platform controlled by a rotary motor 13 is installed on the base. Then, a pitch bracket is installed on the rotating platform. The motor shaft of the pitch motor 12 on the pitch bracket is connected to the head camera 11. The pitch state and rotation angle of the head camera 11 are controlled by the pitch motor 12 and the rotary motor 13, respectively.

[0050] The gripping robotic arm 4 includes a left robotic arm 41 and a right robotic arm 42 symmetrically arranged on both sides of the torso body 14. The torso body 14 is provided with two robotic arm bases 46 that control the left robotic arm 41 and the right robotic arm 42 respectively. The left robotic arm 41 is provided with a left gripper 43, and the right robotic arm 42 is provided with a right gripper 44. The ends of the left robotic arm 41 and the right robotic arm 42 are provided with vision cameras 45 for recognizing the gripping state of the left gripper 43 and the right gripper 44 respectively.

[0051] The left robotic arm 41 and the right robotic arm 42 are independently controlled by the controller 8. The left robotic arm 41 and the right robotic arm 42 can grasp objects individually or simultaneously, depending on actual needs.

[0052] Depending on the environment, the left gripper 43 and the right gripper 44 can be selected with suction cups, grippers, dexterous hands or other structures suitable for the shape of the material, and the material can be grasped by the coordinated movement of the controller 8.

[0053] The vision cameras 45 at the ends of the left robotic arm 41 and the right robotic arm 42 work in conjunction with the head camera 11 to identify the position of the material, thereby improving the recognition accuracy.

[0054] The lifting assembly 2 includes a lifting body 21 connected to a vertically mounted mobile chassis 3, a body assembly 1 mounted on the top of the lifting body 21, and a temporary storage box 24 corresponding to the gripping robotic arm 4 on the lifting body 21.

[0055] The lifting body 21 is provided with a temporary storage platform 23 for supporting the temporary storage box 24, and the lifting body 21 is provided with a lifting rod 22 for controlling the lifting of the temporary storage platform 23.

[0056] The lifting body 21 can be driven by a servo motor or hydraulic system, and in conjunction with the extension and retraction of the multi-stage telescopic linkage, it can drive the body component 1 to move up and down, thereby realizing the adjustment and control of the height of the head camera 11, the left robotic arm 41 and the right robotic arm 42.

[0057] The temporary storage bin 24 is used for temporary storage of materials during sorting operations. When the temporary storage platform 23 is lowered to its lowest position, it is convenient for the temporary storage bin to connect with other conveying equipment or manual labor to realize the transfer of materials. The lifting rod 22 is driven by hydraulic or servo motor to drive the temporary storage platform 23 to adjust its position up and down.

[0058] The slewing support 9 is used to connect the lifting assembly 2 and the torso body 14, and controls the torso body 14 to rotate 360° via the controller 8.

[0059] A control method for an embodied robot used for intelligent picking of warehouse materials, based on the embodied robot for intelligent picking of warehouse materials, includes the following steps:

[0060] Step S1: Construct a map. The radar component 6 on the mobile chassis 3 scans the environment to generate a near-ground global map. The head camera 11 then supplements the map with height information to construct a fused map with height-layered obstacle information.

[0061] The radar component 6 and the head camera 11 are jointly calibrated, and parameters such as scanning frequency and exposure time are calibrated simultaneously to ensure that the data of the radar component 6 and the head camera 11 correspond in real time.

[0062] Step S2: Plan the travel path and move. Based on the fusion map, plan the travel path of the embodied robot to the shelf, and control the embodied robot to move through the moving wheel set 31 through the controller 8 until it reaches the set position next to the shelf.

[0063] During the movement of the embodied robot, obstacles affecting its movement are detected in real time by the head camera 11 and radar component 6. When an obstacle is encountered, the movement direction of the moving wheel set 31 is adjusted by the controller 8 until the embodied robot bypasses the obstacle. After bypassing the obstacle, the movement path is replanned. If an obstacle is detected again during the movement, the movement direction of the moving wheel set 31 is adjusted by the controller 8 until the embodied robot bypasses the obstacle and the movement path is replanned again until it reaches the designated position next to the shelf.

[0064] Step S3: The spatial positions of multiple points of the embodied robot are calibrated. Using the moving wheel set 31 as the reference coordinate, the coordinate transformation relationship between the gripping robot base 46, the head camera 11 and the moving wheel set 31 is obtained. After obtaining the coordinate transformation relationship, the changes of the head camera 11, the left robot arm 41 and the right robot arm 42 relative to the moving wheel set 31 can be obtained by the changes of the motor codes carried by each component during the robot's operation.

[0065] Specifically, let:

[0066] The moving chassis coordinate system (w) is a fixed reference coordinate system with the center of the moving chassis 3 as the origin;

[0067] The torso coordinate system (p) is as follows: The bottom of the body assembly 1 is fixed on the lifting assembly 2, and it rotates and rises and falls with the moving base 3 and the lifting assembly 2.

[0068] Robotic arm base coordinate system (a): Located on the torso coordinate system, moving with the torso;

[0069] Head camera coordinate system (c): Controlled by rotary motor 13 and pitch motor 12, it rotates and pitches relative to the torso coordinate system;

[0070] The end effector coordinate system (e) of the robotic arm is located on the left gripper 43 or the right gripper 44.

[0071] The coordinate transformation of the torso 14 relative to the movable chassis 3 is as follows:

[0072]

[0073] Where: θ1 is the horizontal rotation angle controlled by the moving wheel set 31;

[0074] z2 represents the vertical lifting height controlled by the lifting motor 13;

[0075] The robotic arm base 46 is at a fixed position (xa, ya, za) relative to the movable chassis 3, that is:

[0076]

[0077] The coordinates of the head camera 11 relative to the torso 14 at a fixed position (xc, yc, zc) are transformed as follows:

[0078]

[0079] Where: θ2 is the horizontal rotation angle controlled by the rotary motor 13.

[0080] The transformation of the robotic arm's end effector relative to the robotic arm base 46 is as follows:

[0081]

[0082] Where R is the rotation matrix and T is the translation vector, both of which are related to the joint angles φ1, φ2, …, φ6 of the robotic arm.

[0083] The transformation of the robotic arm's end effector relative to the mobile chassis's 3D coordinate system is as follows:

[0084]

[0085] The transformation of the head camera 11 relative to the coordinate system of the moving chassis 3 is as follows:

[0086]

[0087] During the picking operation, the coordinates of the materials on the shelf relative to the head camera 11 are obtained by taking pictures using the head camera 11:

[0088]

[0089] The coordinate transformation of the material relative to the moving chassis 3 is as follows:

[0090]

[0091] Before spatial positioning is determined, the embodied component 1 is facing the direction of travel of the embodied robot, and the gripping robotic arm 4 is in a retracted and folded state on the torso body 14, so as to minimize the collision range of the embodied component 1.

[0092] Step S4: Adjust the relative position of the shelf and the embodied robot. Rotate the embodied component 1 to face the shelf using the slewing support 9. Detect the relative distance between the embodied robot and the shelf using the head camera 11. When the head camera 11 detects that the embodied robot is too far from the shelf, control the moving chassis 3 to move closer to the shelf.

[0093] Step S5: Construct an electronic fence, identify the location of obstacles in the shelf using head camera 11 and radar component 6, and calibrate the obstacles at the reference coordinates in step S3.

[0094] To construct the electronic fence information composed of obstacles such as shelf uprights and beams, firstly, in the coordinate system of the moving chassis in step S3, the radar component 6 and the head camera 11 are jointly calibrated to obtain the coordinate information of obstacles such as uprights and beams. Then, based on the relative coordinate relationship between the material and the mobile chassis 3 of the robot, the range of the electronic fence is defined for the left robotic arm 4 and the right robotic arm 42.

[0095] Step S6: Identify the material, adjust the height of the body component 1 through the lifting component 2, identify the shape of the material and its relative coordinates under the reference coordinates in step S3 through the head camera 11, and plan the grasping path of the grasping robot arm 4 to avoid the electronic fence according to the electronic fence obtained in step S5.

[0096] Using the electronic fence range obtained in step S5, the positional relationship between obstacles and materials is calculated, and then the optimal grasping path that can avoid the electronic fence is generated through path planning.

[0097] Step S7: By calculating the condition number κ(J) during the grasping process of the gripping robot arm 4, check whether there are singularities in the grasping path. If there are no singularities, the controller 8 guides the gripping robot arm 4 to grasp according to the planned path; if there are singularities, the controller 8 controls the moving wheel group 31 to slightly translate the robot body and adjust the relative position of the robot body and the material. Then repeat steps 4 to 6 until there are no singularities.

[0098] in: , Let denote the 2-norm of the Jacobian matrix. The 2-norm of a matrix is ​​equal to its largest singular value, i.e.:

[0099]

[0100] Step S8: After the gripping is completed, the gripping robotic arm 4 retracts and approaches the torso body 14, rotates the body assembly 1 to return to the initial orientation through the rotary support 9, and controls the lifting assembly 2 to return the body assembly 1 to the initial position.

[0101] Step S9: The position of the temporary storage bin 24 is obtained through the head camera 11, the controller 8 generates the material release path of the gripping robotic arm 4, and places the material in the temporary storage bin 24 according to the generated path.

[0102] Step S10: If multiple materials are picked up from the same bin on the shelf, repeat steps S6-S9. After completion, the robot will first move laterally away from the shelf and then move forward along the aisle to reach the next picking position to repeat the operation.

[0103] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention; therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0104] Although this document uses numerous reference numerals from the figures, such as: body assembly 1, head camera 11, pitch motor 12, rotation motor 13, torso body 14, lifting assembly 2, lifting body 21, lifting rod 22, temporary storage platform 23, temporary storage bin 24, mobile chassis 3, mobile wheel set 31, omnidirectional wheel 32, first steering wheel 33, second steering wheel 34, gripping robotic arm 4, left robotic arm 41, right robotic arm 42, left gripper 43, right gripper 44, vision camera 45, robotic arm base 46, obstacle avoidance and navigation 5, radar assembly 6, front-end lidar 61, rear-end lidar 62, vision sensor 7, front-end vision sensor 71, rear-end vision sensor 72, controller 8, etc., the possibility of using other terms is not excluded. The use of these terms is merely for the convenience of describing and explaining the essence of the invention; interpreting them as any additional limitation would contradict the spirit of the invention.

Claims

1. A embodied robot for intelligent picking of warehouse materials, comprising an embodied component (1) and a mobile chassis (3) for moving the embodied component (1), wherein the embodied component (1) is equipped with a lifting component (2) for controlling the lifting of the embodied component (1) and a rotary support (9) for controlling the rotation of the embodied component (1); characterized in that, The embodied component (1) includes a torso (14), a head camera (11) mounted on the torso (14), and a gripping robotic arm (4) mounted on the torso (14); the mobile chassis (3) includes a set of moving wheels (31) for moving the embodied component (1) and an obstacle avoidance navigation (5) for avoiding obstacles. The obstacle avoidance navigation (5) includes a radar component (6) that works with the head camera (11) to jointly calibrate the environmental coordinates and a visual sensor (7) for identifying the surrounding environment; the embodied robot also includes a controller (8) for constructing a navigation route in real time based on the environmental coordinates, planning the driving mode of the mobile chassis (3), and planning the gripping path of the gripping robotic arm (4).

2. The embodied robot for intelligent picking of warehouse materials according to claim 1, characterized in that, The radar assembly (6) includes a front-end lidar (61) installed at the front end of the moving chassis (3) in the direction of movement and a rear-end lidar (62) installed at the rear end of the moving chassis (3) in the direction of movement. The vision sensor (7) includes a front-end vision sensor (71) corresponding to the front-end lidar (61) and a rear-end vision sensor (72) corresponding to the rear-end lidar (62). The front-end lidar (61) and the rear-end lidar (62) are diagonally arranged on the moving chassis (3).

3. The embodied robot for intelligent picking of warehouse materials according to claim 1, characterized in that, The moving wheel assembly (31) includes a universal wheel (32) for supporting the movement of the moving wheel assembly (31), a first steering wheel (33) and a second steering wheel (34) for adjusting the movement direction of the moving chassis (3).

4. The embodied robot for intelligent picking of warehouse materials according to claim 1, characterized in that, The head camera (11) is mounted on the top of the torso body (14), and the torso body (14) is provided with a pitch motor (12) and a rotation motor (13) for adjusting the recognition angle of the head camera (11).

5. The embodied robot for intelligent picking of warehouse materials according to claim 1, characterized in that, The gripping robotic arm (4) includes a left robotic arm (41) and a right robotic arm (42) symmetrically arranged on both sides of the torso body (14). The torso body (14) is provided with two robotic arm bases (46) that control the left robotic arm (41) and the right robotic arm (42) respectively. The left robotic arm (41) is provided with a left gripper (43), and the right robotic arm (42) is provided with a right gripper (44). The ends of the left robotic arm (41) and the right robotic arm (42) are provided with vision cameras (45) for recognizing the gripping state of the left gripper (43) and the right gripper (44) respectively.

6. The embodied robot for intelligent picking of warehouse materials according to claim 1, characterized in that, The lifting assembly (2) includes a lifting body (21) connected to a vertically mounted mobile chassis (3), a body assembly (1) installed on the top of the lifting body (21), and a temporary storage box (24) corresponding to the gripping robotic arm (4) on the lifting body (21).

7. A holographic robot for intelligent picking of warehouse materials according to claim 6, characterized in that, The lifting body (21) is provided with a temporary storage platform (23) for supporting the temporary storage box (24), and the lifting body (21) is provided with a lifting rod (22) for controlling the lifting of the temporary storage platform (23).

8. A control method for an embodied robot for intelligent picking of warehouse materials, based on the embodied robot for intelligent picking of warehouse materials as described in any one of claims 1-7, characterized in that, Includes the following steps: Step S1: Construct a map by scanning the environment with the radar component (6) on the mobile chassis (3) to generate a near-ground global map, and then supplementing the height information in the map with the head camera (11) to construct a fused map with height-layered obstacle information. Step S2: Plan the travel path and move. Based on the fusion map, plan the travel path of the embodied robot to the shelf, and control the embodied robot to move through the moving wheel set (31) through the controller (8) until it reaches the set position next to the shelf; Step S3: The spatial positions of multiple points of the embodied robot are calibrated. Using the moving wheel set (31) as the reference coordinate, the coordinate transformation relationship between the gripping robot base (46), the head camera (11) and the moving wheel set (31) is obtained. After obtaining the coordinate transformation relationship, the changes of the head camera (11), the left robot arm (41) and the right robot arm (42) relative to the moving wheel set (31) can be obtained by the changes of the motor codes carried by each component during the robot's operation. Step S4: Adjust the relative position of the shelf and the embodied robot, turn the embodied component (1) to face the shelf by using the slewing support (9), and detect the relative distance between the embodied robot and the shelf by the head camera (11); Step S5: Construct an electronic fence, identify the location of obstacles in the shelf using a head camera (11) and radar component (6), and mark the obstacles at the reference coordinates in step S3; Step S6: Identify the material, adjust the height of the body component (1) through the lifting component (2), identify the shape of the material and its relative coordinates under the reference coordinates in step S3 through the head camera (11), and plan the grasping path of the grasping robot arm (4) to avoid the electronic fence according to the electronic fence obtained in step S5. Step S7: By calculating the condition number κ(J) during the grasping process of the gripping robot (4), check whether there are singularities in the grasping path. If there are no singularities, the controller (8) guides the gripping robot (4) to grasp according to the planned path; if there are singularities, the controller (8) controls the moving wheel group (31) to move the robot body slightly to adjust the relative position of the robot body and the material. Then repeat steps 4 to 6 until there are no singularities. in: , Let denote the 2-norm of the Jacobian matrix. The 2-norm of a matrix is ​​equal to the largest singular value of the matrix. Step S8: After the grasping is completed, the grasping robotic arm (4) retracts and approaches the torso body (14), rotates back to the initial direction through the rotary support (9), and controls the lifting component (2) to return the body component (1) to the initial position; Step S9: The position of the temporary storage box (24) is obtained by the head camera (11), the controller (8) generates the material feeding path of the gripping robot arm (4), and places the material in the temporary storage box (24) according to the generated path; Step S10: If multiple materials are picked up from the same bin on the shelf, repeat steps S6-S9. After completion, the robot will first move laterally away from the shelf and then move forward along the aisle to reach the next picking position to repeat the operation.

9. A control method for an embodied robot for intelligent picking of warehouse materials according to claim 8, characterized in that, In step S2, during the movement of the embodied robot, obstacles affecting the movement of the embodied robot are detected in real time by the head camera (11) and radar component (6). When an obstacle is encountered, the movement direction of the moving wheel group (31) is adjusted by the controller (8) until the embodied robot bypasses the obstacle. After bypassing the obstacle, the movement path is replanned until the designated position next to the shelf is reached.

10. A control method for an embodied robot for intelligent picking of warehouse materials according to claim 8, characterized in that, Before step S3 calibrates the spatial position, the embodied component (1) is facing the direction of travel of the embodied robot, and the gripping mechanical arm (4) is in a retracted and folded state on the torso body (14).

Images