Composite mobile assembly robot and workpiece grabbing method

Abstract

The application discloses a composite mobile assembly robot and a workpiece grabbing method, and belongs to the technical field of robots. The chassis adopts a configuration that four groups of single-degree-of-freedom wheel legs are symmetrically arranged on the two sides of the chassis frame, each group of wheel legs is driven by a leg joint motor to swing a leg main body around a horizontal axis, and a rudder system structure and a wheel group arranged at the tail end of the leg main body are cooperatively used for steering and moving. Meanwhile, a six-degree-of-freedom mechanical arm adopts a series layout that a yaw gimbal, a first pitching joint, a first mechanical arm, a second pitching joint, a second mechanical arm, an end effector assembly and a gripper structure are sequentially hinged, so that the six-degree-of-freedom mechanical arm has sufficient movement flexibility and structural rigidity in the pitching and yaw directions. Meanwhile, the movement of each wheel leg and the mechanical arm is uniformly coordinated through a control module, so that the terrain passing capacity and operation stability are ensured.

Description

Technical Field

[0001] This invention belongs to the field of robotics technology, and particularly relates to a composite mobile assembly robot and a workpiece gripping method. Background Technology

[0002] In applications such as warehousing and logistics, power facility inspection and maintenance, construction, and emergency rescue, robots are often required to transport modular workpieces in unstructured terrain environments including steps, undulating surfaces, and ramps, and assemble these workpieces into high-level assembly stations with plug-in interfaces. Examples of tasks include replacing battery modules, installing testing equipment, and docking plug-in functional components. Such operations place high demands on the mobile platform's terrain-crossing capabilities, the robotic arm's flexibility and effective working range, and its ability to automatically identify target workpieces.

[0003] Existing mobile assembly robots typically consist of a mobile chassis and a working mechanism mounted on the chassis. The types of mobile chassis mainly include wheeled chassis, tracked chassis, rocker-arm obstacle-crossing chassis, and articulated-legged chassis. The types of working mechanisms mainly include gantry-type actuators, SCARA robotic arms, and articulated multi-degree-of-freedom robotic arms. However, wheeled chassis struggle to traverse high steps and require smooth road surfaces; tracked chassis have low motion efficiency and high steering wear; rocker-arm chassis have complex structures and are heavy, reducing mobility on flat ground; and articulated-legged chassis have multiple joints and linkage mechanisms in each leg, resulting in a large number of joint motors, complex control, and high cost, making stable control difficult for robots equipped with working mechanisms. Moreover, the size and weight of the gantry-type actuator are difficult to reduce, which is not conducive to the overall miniaturization and lightweighting of the robot; the SCARA robotic arm adopts a cantilever beam structure, which has weak vertical stiffness and is prone to longitudinal sway at the end. In addition, the center of gravity of the whole machine is relatively high, which can easily cause the whole machine to sway and the center of gravity to shift when the mobile platform turns or the robotic arm moves. Summary of the Invention

[0004] The present invention aims to provide a composite mobile assembly robot and a workpiece gripping method to solve the above-mentioned technical problems, and to provide a mobile assembly robot with good terrain mobility and stable and reliable operation.

[0005] To address the aforementioned technical problems, this invention provides a composite mobile assembly robot, comprising: a chassis frame, four sets of single-degree-of-freedom wheel legs symmetrically mounted on both sides of the chassis frame, a six-degree-of-freedom robotic arm mounted on the chassis frame, and a control module. The single-degree-of-freedom wheel leg includes a leg joint motor, a leg body, a rudder system structure, and a wheel assembly; the leg joint motor is fixedly connected to the chassis frame; the drive shaft of the leg joint motor is fixedly connected to the leg body, and is used to drive the leg body to swing around a horizontal axis; the rudder system structure is disposed on the leg body and is fixedly connected to the wheel assembly. The six-degree-of-freedom robotic arm, from bottom to end, comprises: a yaw gimbal, a first pitch joint, a first robotic arm, a second pitch joint, a second robotic arm, an end effector assembly, and a gripper structure. The yaw gimbal is mounted on the chassis frame. The first end of the first robotic arm is hinged to the yaw gimbal via the first pitch joint. The second end of the first robotic arm is hinged to the first end of the second robotic arm via the second pitch joint. The second end of the second robotic arm is fixedly connected to the end effector assembly. The gripper structure is located at the end of the end effector assembly. The control terminals of the leg joint motor, the rudder system, the wheel assembly, and the six-degree-of-freedom robotic arm are all electrically connected to the output terminal of the control module; the control module is used to control the composite mobile assembly robot to move and grasp workpieces.

[0006] Understandably, compared to existing technologies, the composite mobile assembly robot provided by this invention employs a chassis configuration with four sets of single-degree-of-freedom wheels symmetrically mounted on both sides of the chassis frame. Each set of wheels is driven by a leg joint motor to swing the leg body around a horizontal axis, cooperating with a rudder system and wheel set located at the end of the leg body to achieve steering and movement. Compared to multi-joint leg chassis, this significantly reduces the number of joint motors and linkage mechanisms, resulting in a simpler chassis structure, simpler control logic, and a significantly lighter overall weight. This allows it to retain good flat-ground mobility in unstructured terrain while also possessing active obstacle-crossing capabilities, thus achieving excellent terrain traversal ability. Simultaneously, six... The six-degree-of-freedom robotic arm adopts a series layout in which the yaw gimbal, first pitch joint, first robotic arm, second pitch joint, second robotic arm, end effector assembly, and gripper structure are hinged sequentially. This avoids the inherent problem of weak vertical stiffness in SCARA robotic arms, enabling the six-degree-of-freedom robotic arm to have sufficient motion flexibility and structural rigidity in both pitch and yaw directions. At the same time, the movement of each wheel leg and robotic arm is uniformly coordinated through the control module. When the chassis crosses steps, the body can be lifted and the attitude adjusted by the active swing of the leg body. The robotic arm can maintain end-effector stability through the coordinated movement of each joint when the chassis attitude changes, thereby ensuring terrain crossing ability and operational stability.

[0007] As a preferred embodiment, the rudder system structure includes: a rudder motor, a wheel assembly bracket, and a bearing mounting base; The rudder motor is mounted on the leg body; the drive shaft of the rudder motor is fixedly connected to the wheel assembly bracket; The wheel bracket and the wheel set are fixedly connected; The bearing mounting base is disposed at the end of the leg body, and the bearing mounting base is slidably connected to the wheel set bracket; The control terminal of the rudder motor serves as the control terminal of the rudder system structure, and it is electrically connected to the output terminal of the control module.

[0008] In the above scheme, the rudder system structure is further defined as consisting of a rudder motor, a wheel assembly bracket, and a bearing mounting base. The rudder motor is mounted on the leg body and fixedly connected to the wheel assembly bracket via a drive shaft. The wheel assembly bracket is fixedly connected to the wheel assembly. The bearing mounting base is located at the end of the leg body and slidably connected to the wheel assembly bracket. Since the drive shaft of the rudder motor directly drives the wheel assembly bracket to steer the wheel assembly, the rudder motor only needs to overcome the rolling resistance and steering resistance torque between the wheel assembly and the ground, without bearing the vertical load of the entire vehicle. The vertical load is borne by the bearing mounting base through the slidable connection surface between it and the wheel assembly bracket and transmitted to the leg body. This achieves the separation of steering drive and vertical load-bearing functions, allowing the rudder motor to drive the wheel assembly to steer flexibly without the need for a high-torque specification, reducing the difficulty and cost of motor selection. At the same time, the slidable connection between the bearing mounting base and the wheel assembly bracket allows for adjustment of coaxiality during assembly to eliminate accumulated errors in parts processing and assembly, ensuring the concentricity between the rudder motor drive shaft and the bearing mounting base, reducing additional resistance and wear during steering, and further improving the steering flexibility of the wheel assembly and the overall motion reliability of the machine.

[0009] As a preferred embodiment, the wheelset includes a wheel motor and a support wheel; the wheel motor is fixedly connected to the wheelset bracket; the drive shaft of the wheel motor is fixedly connected to the support wheel; the control terminal of the wheel motor serves as the control terminal of the wheelset and is electrically connected to the output terminal of the control module.

[0010] In the above scheme, by setting a wheel motor and a support wheel in the wheel set, the wheel motor is fixedly connected to the wheel set bracket, and the drive shaft of the wheel motor is fixedly connected to the support wheel. This allows the wheel motor to directly drive the support wheel to rotate, thereby enabling the robot to move forward, backward, and turn in place. Since the wheel motor and the support wheel it drives are integrated into the wheel set bracket, each wheel set has independent driving and steering capabilities. This allows the driving power to be converted into forward thrust more efficiently. Even under power-limited conditions, it can still maintain a high moving speed and obstacle-crossing driving force, thus improving the overall motion efficiency and terrain adaptability of the robot while ensuring omnidirectional mobility.

[0011] As a preferred embodiment, the end effector assembly includes: a first roll joint, a third pitch joint, and a second roll joint; The first tumbling joint is located between the second end of the second robotic arm and the third pitch joint; The third pitch joint is disposed between the first roll joint and the second roll joint, and is used to drive the second roll joint to pitch and rotate. The second roll joint is disposed between the third pitch joint and the gripper structure; The second rolling joint is fixedly connected to the gripper structure and is used to drive the gripper structure to rotate about its own axis.

[0012] In the above solution, a first roll joint, a third pitch joint, and a second roll joint are incorporated into the end effector assembly. The first roll joint is positioned between the second robotic arm and the third pitch joint, and the third pitch joint is positioned between the first roll joint and the second roll joint. The third pitch joint drives the second roll joint to pitch and rotate. The second roll joint is positioned between the third pitch joint and the gripper structure, and the second roll joint drives the gripper structure to rotate around its own axis. Because the first roll joint enables continuous rotation around the axis of the second robotic arm, the end effector of the robotic arm has no blind spots in the workspace. The third pitch joint provides the gripper structure with freedom for pitch and oscillation. The first pitch joint allows the gripper structure to approach and dock with workpieces at different pitch angles, adapting to assembly requirements of different heights and postures. The second roll joint allows the gripper structure to rotate around its own axis, adjusting the gripper orientation to match the workpiece's interface direction when gripping workpieces with specific posture requirements. These three joints give the robotic arm's end effector three independent degrees of freedom for posture adjustment. Combined with the position adjustment capabilities provided by the yaw gimbal, the first pitch joint, and the second pitch joint, they together constitute a complete six-degree-of-freedom operation capability, enabling the end effector gripper to approach the target workpiece at any spatial position and any spatial posture, meeting the precise control requirements of the end effector posture for high-level assembly stations.

[0013] As a preferred embodiment, the gripper structure includes: a gripper drive motor, a crank-slider structure, and grippers; The gripper drive motor is located at the end of the second rolling joint; The drive shaft of the gripper drive motor is fixedly connected to the first end of the crank-slider structure; The second end of the crank-slider structure is fixedly connected to the gripper. The control terminal of the gripper drive motor serves as the control terminal of the six-degree-of-freedom robotic arm, and it is electrically connected to the output terminal of the control module.

[0014] In the above solution, the gripper structure is configured to consist of a gripper drive motor, a crank-slider structure, and grippers. The gripper drive motor is located at the end of the second tumbling joint and drives the crank-slider structure to move via a transmission shaft. The crank-slider structure then transmits the motion to the grippers to achieve opening and closing actions. The crank-slider structure converts the rotational motion of the motor output shaft into the linear reciprocating motion of the grippers. Furthermore, the cooperation between the crank and the slider in the crank-slider structure effectively amplifies the driving force in the gripping state, thus obtaining sufficient clamping force to firmly hold the workpiece without the need for a high-power motor. Compared to pneumatic grippers or vacuum suction cups, the gripper structure does not require an external air source or vacuum source, avoiding the problem of air pipes and cables accumulating and tangling at the end of the robotic arm, which is beneficial for maintaining a clear end view and tidy cable management. Compared to worm gear or lead screw grippers, the crank-slider structure ensures sufficient clamping force while having a faster response speed and a larger opening and closing stroke, and can adapt to workpieces of different sizes and specifications, thus taking into account multiple operational requirements such as clamping force, response speed, and workpiece adaptability.

[0015] As a preferred embodiment, the composite mobile assembly robot further includes: a positioning label lifting mechanism; the positioning label lifting mechanism is mounted on the chassis frame; The positioning tag lifting mechanism includes: a positioning tag lifting platform, a lifting mechanism drive motor, a first link, and a second link; The lifting mechanism drive motor is mounted on the chassis frame; The drive shaft of the lifting mechanism's drive motor is connected to the first connecting rod via a gear transmission mechanism. The drive shaft of the lifting mechanism drive motor is connected to the second connecting rod through a gear transmission mechanism; Both the first link and the second link are mounted on the positioning tag lifting platform; The control terminal of the lifting mechanism drive motor is electrically connected to the output terminal of the control module.

[0016] In the above scheme, the positioning tag lifting mechanism includes a positioning tag lifting platform, a lifting mechanism drive motor, a first link, and a second link. The lifting mechanism drive motor drives the first link and the second link simultaneously through a transmission shaft, causing the positioning tag lifting platform to rise and fall relative to the chassis frame. Since the first link and the second link move synchronously under the drive of the lifting mechanism drive motor, the positioning tag lifting platform maintains a horizontal posture throughout the lifting process, preventing the positioning tag from tilting as the platform rises and falls. This ensures that the positioning tag can be accurately identified and measured by the external positioning system at different height positions, realizing automatic deployment and retraction of the positioning tag and precise positioning height control, and improving the robot's positioning adaptability and autonomous operation capability in multiple workstations and scenarios.

[0017] As a preferred embodiment, guide wheels are provided on both protrusions of the chassis frame.

[0018] In the above solution, guide wheels are installed on the protruding parts on both sides of the chassis frame. When the robot crosses a step, the guide wheels located on both sides of the front end of the chassis frame can make contact with the edge of the step first. Compared with the rubber-coated wheel surface of the wheel set directly colliding with the edge of the step, the guide wheels, with their smaller diameter and smoother shape, can roll more smoothly along the edge of the step, guiding the front of the vehicle body to smoothly place on the surface of the step, reducing the impact load and resistance during obstacle crossing. At the same time, when the robot passes through undulating roads or slopes, the guide wheels can serve as auxiliary support points, providing temporary support and guidance when the wheel set is suspended or slipping, preventing the chassis frame from directly scraping and colliding with the ground. During the robot's movement, the guide wheels can also play a role in buffering and collision prevention. When the robot accidentally touches an obstacle, the guide wheels first absorb the impact and guide the vehicle body away from the direction of the obstacle through rolling guidance, reducing the damage caused by direct impact to the chassis frame and internal components, thereby improving the safety and reliability of the robot driving in complex terrain.

[0019] Accordingly, embodiments of the present invention provide a workpiece gripping method for a composite mobile assembly robot, applicable to a composite mobile assembly robot as described above, comprising: Image data of the target workpiece is acquired, and the image data is preprocessed to extract several candidate target point clouds; the several candidate target point clouds are clustered to obtain several candidate point cloud clusters. Calculate the cross-sectional radius distribution of each candidate point cloud cluster to construct the cross-sectional radius distribution curve of each candidate point cloud cluster; Based on the cross-sectional radius distribution curve, the candidate point cloud clusters are geometrically filtered to determine the pose information of the target workpiece. Based on the pose information of the target workpiece, the composite mobile assembly robot is controlled to grasp the target workpiece.

[0020] Understandably, compared to existing technologies, this invention acquires image data of the target workpiece and preprocesses it to extract candidate target point clouds. These candidate point clouds are then clustered to obtain candidate point cloud clusters. The cross-sectional radius distribution of each cluster is calculated, and a cross-sectional radius distribution curve is constructed. Based on this curve, the clusters are geometrically filtered to determine the pose information of the target workpiece. Finally, the robot is controlled to perform grasping based on the pose information. Since this workpiece grasping method does not rely on the appearance features of the target workpiece, such as color, texture, or surface pattern, but rather on the cross-sectional radius distribution features of the workpiece's three-dimensional geometric shape for identification and judgment, it does not require pre-collecting a large number of training samples for model training, nor does it require adjusting recognition parameters for different lighting conditions. It has a natural resistance to interference from workpiece surface conditions and changes in ambient lighting. Furthermore, by characterizing the shape of the target workpiece through the compact geometric feature of the cross-sectional radius distribution curve, it features low computational load and fast processing speed, meeting the response speed requirements of real-time grasping operations. Moreover, the geometric shape filtering method allows for flexible adjustment based on the geometric characteristics of different workpiece models, providing good workpiece model adaptability and convenient on-site deployment.

[0021] As a preferred embodiment, calculating the cross-sectional radius distribution of each candidate point cloud cluster to construct the cross-sectional radius distribution curve of each candidate point cloud cluster includes: For each candidate point cloud cluster, the covariance matrix of the candidate point cloud cluster is calculated according to a preset principal component analysis algorithm; Eigenvalue decomposition is performed on the covariance matrix to determine the principal axis of the candidate point cloud clusters; Project each of the candidate target point clouds in the candidate point cloud cluster onto the main axis to obtain the projection position of each of the candidate target point clouds on the main axis; The projection position is divided into slices to obtain several slices, and the average distance from all candidate target point clouds in each slice to the principal axis is calculated to obtain the average distance value of each slice. Based on the order of each slice on the main axis, the average distance values ​​of each slice are sorted to form the cross-sectional radius distribution curve of the candidate point cloud cluster.

[0022] In the above scheme, principal component analysis (PCA) is used to calculate the covariance matrix and perform eigenvalue decomposition for each candidate point cloud cluster to determine the principal axis of the cluster. Then, each candidate target point cloud in the cluster is projected onto the principal axis to obtain the projected position. The projected position is then sliced, and the average distance from all candidate target point clouds within each slice to the principal axis is calculated. Finally, the average distance values ​​are arranged in order of slices on the principal axis to form a cross-sectional radius distribution curve. The longest extension direction of the candidate point cloud cluster, which PCA can accurately extract, is used as the principal axis, ensuring that each slice is perpendicular to the principal axis direction during subsequent slicing, thus guaranteeing... The cross-sectional radius distribution curve can accurately reflect the cross-sectional contour variation of the target workpiece along its length. By dividing the projected position into slices and calculating the average radial distance within each slice, irregular point cloud data can be transformed into a regular one-dimensional radius distribution curve. This preserves the geometric contour information of the target workpiece along the main axis while significantly reducing the data dimensionality. This eliminates the need for complex 3D registration or feature matching of the original high-dimensional point cloud data in subsequent geometric shape screening. As a result, while ensuring recognition accuracy, computational complexity and processing time are significantly reduced, making it suitable for online grasping operations with high real-time requirements.

[0023] As a preferred embodiment, the step of performing geometric morphology screening on the candidate point cloud clusters based on the cross-sectional radius distribution curve to determine the pose information of the target workpiece includes: Identify the maximum radius value, peak distribution, and trough distribution in the cross-sectional radius distribution curve; The expansion ratio of the candidate point cloud clusters is determined based on the peak distribution and the maximum radius value. Based on the expansion ratio, peak distribution, and trough distribution, the candidate point cloud clusters are subjected to peak number condition detection, trough condition detection, expansion ratio condition detection, peak symmetry condition detection, and size constraint detection, respectively. If the candidate point cloud clusters simultaneously satisfy the peak number condition, trough condition, expansion ratio condition, peak symmetry condition, and size constraint, then the candidate point cloud clusters are taken as the point cloud clusters of the target workpiece. The point cloud clusters of the target workpiece are transformed to obtain the pose information of the target workpiece.

[0024] In the above scheme, the peaks and troughs of the cross-sectional radius distribution curve are identified and statistically analyzed. Specifically, this includes detecting the maximum radius value, peak distribution, and trough distribution. Based on the peak distribution and maximum radius value, the expansion ratio is calculated. Then, the candidate point cloud clusters are sequentially subjected to peak number condition detection, trough condition detection, expansion ratio condition detection, peak symmetry condition detection, and size constraint detection. Only when all conditions are met is it determined to be a target workpiece. The peak number condition ensures that the target workpiece has a characteristic shape with two protruding ends. The trough condition ensures that there is a concave feature in the middle of the workpiece. The expansion ratio condition quantitatively describes the significance of the protrusion relative to the concavity. The symmetry condition utilizes the geometric characteristics of the two ends of the target workpiece being symmetrical about the center. The size constraint excludes non-target objects from the absolute scale, effectively eliminating the interference of non-target objects and greatly reducing the false detection rate and false detection rate. Thus, it achieves real-time accurate identification and pose determination of target workpieces that are robust to changes in illumination and surface state without training data. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the structure of a composite mobile assembly robot provided in an embodiment of the present invention; Figure 2 A front view of a single-degree-of-freedom wheel leg provided in an embodiment of the present invention; Figure 3 A side view of a single-degree-of-freedom wheel leg provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of a six-degree-of-freedom robotic arm provided in an embodiment of the present invention; Figure 5 A schematic diagram of a composite mobile assembly robot traversing steps provided in an embodiment of the present invention; Figure 5 include Figure 5 (a) Figure 5 (b) Figure 5 (c) and Figure 5 (d); Figure 5 (a) A schematic diagram of a first type of composite mobile assembly robot crossing steps provided in an embodiment of the present invention; Figure 5 (b) A schematic diagram of a second type of composite mobile assembly robot crossing steps provided in an embodiment of the present invention; Figure 5 (c) A schematic diagram of a third type of composite mobile assembly robot crossing steps provided in an embodiment of the present invention; Figure 5 (d) is a schematic diagram of the fourth type of composite mobile assembly robot crossing steps provided in the embodiments of the present invention; Figure 6 This is a schematic diagram of a gripper structure provided in an embodiment of the present invention; Figure 7 This is a schematic diagram of the structure of a master-slave teleoperation controller provided in an embodiment of the present invention; Figure 8 This is a schematic diagram of a positioning tag lifting mechanism provided in an embodiment of the present invention; Figure 9 This is a schematic diagram of a chassis frame provided in an embodiment of the present invention; Figure 10 A flowchart illustrating the steps of a workpiece gripping method for a composite mobile assembly robot, as provided in an embodiment of the present invention. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] Example 1 Please refer to Figure 1 , Figure 1 This is a schematic diagram of the structure of a composite mobile assembly robot provided in an embodiment of the present invention. Figure 2 A front view of a single-degree-of-freedom wheel leg provided in an embodiment of the present invention; Figure 3 A side view of a single-degree-of-freedom wheel leg provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of a six-degree-of-freedom robotic arm provided in an embodiment of the present invention; The composite mobile assembly robot includes: a chassis frame 1, four sets of single-degree-of-freedom wheel legs 2 symmetrically installed on both sides of the chassis frame, a six-degree-of-freedom robotic arm 3 installed on the chassis frame, and a control module 4; The single-degree-of-freedom wheel leg 2 includes a leg joint motor 21, a leg body 22, a rudder system structure 23, and a wheel set 24; the leg joint motor 21 is fixedly connected to the chassis frame 1; the drive shaft of the leg joint motor 21 is fixedly connected to the leg body 22, and is used to drive the leg body 22 to swing around a horizontal axis; the rudder system structure 23 is disposed on the leg body 22, and is fixedly connected to the wheel set 24. The six-degree-of-freedom robotic arm 3, from bottom to end, comprises: a yaw gimbal 31, a first pitch joint 32, a first robotic arm 33, a second pitch joint 34, a second robotic arm 35, an end effector assembly 36, and a gripper structure 37. The yaw gimbal 31 is mounted on the chassis frame 1. The first end of the first robotic arm 33 is hinged to the yaw gimbal 31 via the first pitch joint 32. The second end of the first robotic arm 33 is hinged to the first end of the second robotic arm 35 via the second pitch joint 34. The second end of the second robotic arm 35 is fixedly connected to the end effector assembly 36. The gripper structure 37 is located at the end of the end effector assembly 36. The control terminals of the leg joint motor 21, the rudder system 23, the wheel set 24, and the six-degree-of-freedom robotic arm 3 are all electrically connected to the output terminal of the control module 4; the control module 4 is used to control the composite mobile assembly robot to move and grasp workpieces.

[0028] Understandably, compared to existing technologies, the composite mobile assembly robot provided by this invention employs a chassis configuration with four sets of single-degree-of-freedom wheels symmetrically mounted on both sides of the chassis frame. Each set of wheels is driven by a leg joint motor to swing the leg body around a horizontal axis, cooperating with a rudder system and wheel set located at the end of the leg body to achieve steering and movement. Compared to multi-joint leg chassis, this significantly reduces the number of joint motors and linkage mechanisms, resulting in a simpler chassis structure, simpler control logic, and a significantly lighter overall weight. This allows it to retain good flat-ground mobility in unstructured terrain while also possessing active obstacle-crossing capabilities, thus achieving excellent terrain traversal ability. Simultaneously, six... The six-degree-of-freedom robotic arm adopts a series layout in which the yaw gimbal, first pitch joint, first robotic arm, second pitch joint, second robotic arm, end effector assembly, and gripper structure are hinged sequentially. This avoids the inherent problem of weak vertical stiffness in SCARA robotic arms, enabling the six-degree-of-freedom robotic arm to have sufficient motion flexibility and structural rigidity in both pitch and yaw directions. At the same time, the movement of each wheel leg and robotic arm is uniformly coordinated through the control module. When the chassis crosses steps, the body can be lifted and the attitude adjusted by the active swing of the leg body. The robotic arm can maintain end-effector stability through the coordinated movement of each joint when the chassis attitude changes, thereby ensuring terrain crossing ability and operational stability.

[0029] In this embodiment, as Figure 2 and Figure 3 As shown, the rudder system structure 23 includes: a rudder motor 231, a wheel set bracket 232, and a bearing mounting base 233; The rudder motor 231 is mounted on the leg body 22; the drive shaft of the rudder motor 231 is fixedly connected to the wheel assembly bracket 232; The wheel bracket 232 and the wheel set 24 are fixedly connected; The bearing mounting seat 233 is disposed at the end of the leg body 22, and the bearing mounting seat 233 is slidably connected to the wheel set bracket 232; The control terminal of the rudder motor 231 serves as the control terminal of the rudder system structure 23, and is electrically connected to the output terminal of the control module 4.

[0030] In the above embodiments, the rudder system structure is further defined as consisting of a rudder motor, a wheel assembly bracket, and a bearing mounting seat. The rudder motor is mounted on the leg body and fixedly connected to the wheel assembly bracket via a drive shaft. The wheel assembly bracket is fixedly connected to the wheel assembly. The bearing mounting seat is located at the end of the leg body and slidably connected to the wheel assembly bracket. Since the drive shaft of the rudder motor directly drives the wheel assembly bracket to steer the wheel assembly, the rudder motor only needs to overcome the rolling resistance and steering resistance torque between the wheel assembly and the ground, and does not need to bear the vertical load of the entire vehicle. The vertical load is borne by the bearing mounting seat through the slidable connection surface between it and the wheel assembly bracket and transmitted to the leg body. This achieves the separation of steering drive and vertical load-bearing functions, allowing the rudder motor to drive the wheel assembly to steer flexibly without the need for a high torque specification, reducing the difficulty and cost of motor selection. At the same time, the slidable connection between the bearing mounting seat and the wheel assembly bracket allows for adjustment of coaxiality during assembly to eliminate accumulated errors in parts processing and assembly, ensuring the concentricity between the rudder motor drive shaft and the bearing mounting seat, reducing additional resistance and wear during steering, and further improving the steering flexibility of the wheel assembly and the overall motion reliability of the machine.

[0031] In this embodiment, as Figure 2 and Figure 3 As shown, the wheel set 24 includes a wheel motor 241 and a support wheel 242; the wheel motor 241 is fixedly connected to the wheel set bracket 232; the drive shaft of the wheel motor 241 is fixedly connected to the support wheel 242; the control end of the wheel motor 241 serves as the control end of the wheel set 24, and it is electrically connected to the output end of the control module 4.

[0032] In the above embodiments, by setting a wheel motor and a support wheel in the wheel set, with the wheel motor fixedly connected to the wheel set bracket and the drive shaft of the wheel motor fixedly connected to the support wheel, the wheel motor directly drives the support wheel to rotate, thereby realizing the robot's forward, backward, and turning functions. Since the wheel motor and its driven support wheel are integrated on the wheel set bracket, each wheel set has independent driving and steering capabilities, allowing the driving power to be converted into forward thrust more efficiently. Even under power-limited conditions, it can still maintain a high moving speed and obstacle-crossing driving force, thus improving the overall motion efficiency and terrain adaptability of the robot while ensuring omnidirectional mobility.

[0033] In one optional embodiment, the leg body 22 is approximately 200mm in length and is made of 6061-T6 aluminum alloy. It consists of a flange portion and a connecting rod portion, which are fixedly connected by cylindrical pins and bolts. Specifically, the leg body 22 is first topologically optimized to generate an optimal force transmission structure, and then its shape is modeled into a CNC-machineable flange portion and connecting rod portion. The two parts are connected by cylindrical pins with a diameter of 4mm and four M5 bolts. With the above optimized design, the weight of a single leg body 22 is only 254g, and it shows no deformation during impact testing.

[0034] The leg joint motor 21 is a DM6248 model, with a rated torque of 30 N·m, a peak torque of 97 N·m, and a weight of 628 g. This selection is based on the following: the leg body 22 is approximately 200 mm long, the motor center distance is approximately 284.5 mm from the wheel base, and with an estimated sprung mass of 25 kg, the single-motor lifting torque requirement is approximately 25 × 9.8 × 0.285 / 4 ≈ 17.45 N·m. Considering the robotic arm's forward load and redundancy, the leg joint motor 21 with a rated torque of 30 N·m, a peak torque of 97 N·m, and a weight of 628 g is selected to fully meet the torque requirements during lifting and obstacle-crossing operations. Simultaneously, the leg joint motor 21 underwent strength verification: The maximum overall weight was set at 32 kg, a drop height of 350 mm, a momentum conversion time of 0.1 s, a horizontal velocity of 3 m / s, a final landing velocity of approximately 7.48 m / s, and an impact force of approximately 2393.6 N. With a front-to-rear load ratio of 38:26, the maximum force on a single front leg was approximately 873.46 N. Simulations showed a vertical load of 1200 N and a lateral load of 600 N, which sufficiently met the torque requirements for standing and obstacle-crossing operations. Furthermore, to ensure the reliability of the cable connection of the leg joint motor 21, a notched printed outer shell was fixed to the tail of the motor. This printed outer shell, through a perforated fiberglass clamp, pressed the motor plug with a pre-tightening force to prevent the plug from loosening.

[0035] For the rudder system structure 23, a deep groove ball bearing is provided in the bearing mounting seat 233, and the wheel assembly bracket 232 is rotatably engaged with the bearing mounting seat 233 through the deep groove ball bearing. To achieve coaxiality adjustment during assembly, a 0.6mm gap is reserved between the wheel assembly bracket 232 and the bearing mounting seat 233, and the mounting hole is appropriately enlarged. During assembly, fine-tuning eliminates accumulated errors in parts machining and assembly, ensuring the concentricity between the rudder motor drive shaft and the bearing mounting seat 233. The rudder motor 231 is a DM6220 type motor.

[0036] For wheel assembly 24, wheel motor 241 is configured as an M3508 hub drive motor. The original dust cover of the gearbox of wheel motor 241 is replaced with a machined end cover, which, in conjunction with a deep groove ball bearing, forms a support end of the output shaft system of wheel motor 241, namely one end of support wheel 242. The other end of support wheel 242 can be supported by the output end of wheel motor 241, thus significantly simplifying the number of parts and the hub structure. Support wheel 242 is configured as a polyurethane-coated wheel with a coating thickness of 6mm and a hardness of A60, to buffer the ground impact caused by the active suspension response delay of leg joint motor 21. The selection criteria for support wheel 242 are as follows: Under the condition of a chassis drive power constraint of 120W, the power loss of the Mecanum wheel exceeds 50%. Based on a total machine weight of 30kg and a friction coefficient of 0.16, the estimated maximum speed is only approximately (50%×120) / (30×9.8×0.16)≈1.27m / s, and the prototype is prone to slipping when crossing steps during testing. Therefore, a steering wheel set with good grip and high motion efficiency is adopted. The wiring of wheel motor 241 is led out using a slip ring hanger. The wiring protective shell of the slip ring hanger also serves as the anti-torque support for the steering motor 231, and it is made of anti-aging ABS engineering plastic.

[0037] In one alternative embodiment, please refer to Figure 5 , Figure 5 A schematic diagram of a composite mobile assembly robot traversing steps provided in an embodiment of the present invention; Figure 5 include Figure 5 (a) Figure 5 (b) Figure 5 (c) and Figure 5 (d); Figure 5 (a) A schematic diagram of a first type of composite mobile assembly robot crossing steps provided in an embodiment of the present invention; Figure 5 (b) A schematic diagram of a second type of composite mobile assembly robot crossing steps provided in an embodiment of the present invention; Figure 5 (c) A schematic diagram of a third type of composite mobile assembly robot crossing steps provided in an embodiment of the present invention; Figure 5 (d) is a schematic diagram of the fourth type of composite mobile assembly robot crossing steps provided in the embodiments of the present invention; as shown. Figure 5 As shown in (a), the composite mobile assembly robot is in its initial state before the first step; then, as... Figure 5 As shown in (b), the two front single-degree-of-freedom wheel legs 2 of the composite mobile assembly robot first place themselves on the edge of the first step, the front legs swing to lift the front of the vehicle body, and then the entire vehicle moves forward, causing the two rear single-degree-of-freedom wheel legs 2 of the composite mobile assembly robot to come close to the first step. After this, as... Figure 5As shown in (c), the two single-degree-of-freedom wheels 2 on the rear side of the composite mobile assembly robot swing and lift the rear of the vehicle body, so that the entire composite mobile assembly robot is on the first step. Finally, as shown in... Figure 5 As shown in (d), the four single-degree-of-freedom wheel legs 2 of the composite mobile assembly robot are made to stand upright, thereby completing the crossing of the first step.

[0038] In this embodiment, as Figure 4 As shown, the end effector assembly 36 includes: a first roll joint 361, a third pitch joint 362, and a second roll joint 363. The first tumbling joint 361 is disposed between the second end of the second robotic arm 35 and the third pitch joint 362; The third pitch joint 362 is disposed between the first roll joint 361 and the second roll joint 363, and is used to drive the second roll joint 363 to pitch and rotate. The second roll joint 363 is disposed between the third pitch joint 362 and the gripper structure 37; The second tumbling joint 363 is fixedly connected to the gripper structure 37, and is used to drive the gripper structure 37 to rotate about its own axis.

[0039] In the above embodiments, by incorporating a first roll joint, a third pitch joint, and a second roll joint in the end effector assembly, wherein the first roll joint is positioned between the second robotic arm and the third pitch joint, and the third pitch joint is positioned between the first roll joint and the second roll joint, the third pitch joint drives the second roll joint to pitch and rotate, and the second roll joint is positioned between the third pitch joint and the gripper structure, driving the gripper structure to rotate around its own axis; since the first roll joint enables continuous rotation around the axis of the second robotic arm, the end effector of the robotic arm has no blind spots in the workspace; the third pitch joint provides the gripper structure with freedom of pitch and oscillation. The first pitch joint allows the gripper structure to approach and dock with workpieces at different pitch angles, adapting to assembly requirements of different heights and postures. The second roll joint allows the gripper structure to rotate around its own axis, adjusting the gripper orientation to match the workpiece's interface direction when gripping workpieces with specific posture requirements. These three joints give the robotic arm's end effector three independent degrees of freedom for posture adjustment. Combined with the position adjustment capabilities provided by the yaw gimbal, the first pitch joint, and the second pitch joint, they together constitute a complete six-degree-of-freedom operation capability, enabling the end effector gripper to approach the target workpiece at any spatial position and any spatial posture, meeting the precise control requirements of the end effector posture for high-level assembly stations.

[0040] In an optional embodiment, the yaw gimbal 31 is equipped with a first rudder motor, model DM6220. The control end of the first rudder motor serves as the control end of the yaw gimbal 31 and is electrically connected to the output end of the control module 4. The yaw gimbal 31 uses RAU6008 crossed roller bearings (rated dynamic load 5.68kN) to withstand the overturning moment. It adopts a 1:1 gear transmission and slip ring wiring to meet large rotation angles and facilitate wiring. The gear design features 32 teeth, a module of 2.5, and a pitch circle diameter of 80mm. The material used is 6061-T6 aluminum alloy (allowable bending stress 100MPa). Based on a load factor of 1.3 and the bending strength formula, the tooth width is ≥6.5mm. The drive wheel is 8mm and the mating wheel is 7mm. The transmission clearance is compensated by the motor parameters of the first rudder motor. Furthermore, the bottom mounting plate of the yaw gimbal 31 is designed as a one-piece machined part, achieving a 42% weight reduction without changing strength after structural optimization. The bottom mounting plate is connected to the chassis frame using four M5 bolts. This embodiment also verifies the yaw gimbal 31. Based on an arm mass of 3.003 kg, a gravity of approximately 29.43 N, a simplified torque of approximately 2.266 N·m, a friction coefficient of 0.2, and a reliability coefficient of 1.2, the preload is approximately 88.29 N under non-slip conditions. The total tensile force of the distal bolts under overturning conditions is approximately 123.54 N. Verifying the allowable stress of 540 MPa for 12.9 grade bolts, the minimum diameter is approximately 0.616 mm. Therefore, the M5 bolts of the yaw gimbal 31 have sufficient slack.

[0041] The first pitch joint 32 houses a first pitch drive motor and a second pitch drive motor, both of which are DM6248 type motors. They are mounted side-by-side in opposite directions on an L-shaped machined part and fixed below the yaw gimbal 31 (using four M5 bolts). This means the motor bodies are located below the yaw gimbal 31. This lowered motor placement significantly reduces the center of gravity and moment of inertia of the six-degree-of-freedom robotic arm 3, effectively suppressing overall swaying caused by platform turning or robotic arm movement. The control terminals of both the first and second pitch drive motors are electrically connected to the output terminal of the control module 4. Crossed roller bearings are installed inside the first and second pitch drive motors, directly connecting to the connecting rod of the first robotic arm under radial stress. Cylindrical pins are used at the four M5 bolt connections to enhance shear resistance.

[0042] The second pitch joint 34 houses a third pitch drive motor, a DM6248 type motor, whose control terminal is electrically connected to the output terminal of the control module 4. The third pitch drive motor is located at the lower end of the first robotic arm 33 (i.e., the end closest to the yaw gimbal 31), and its output terminal is hinged to the rotation axis of the second pitch joint 34 via a connecting rod of the second robotic arm 35. This allows the second robotic arm to be driven to pitch and swing around the axis of the second pitch joint via the connecting rod of the second robotic arm 35. The shaft system of the second pitch joint 34 uses a 7075 aluminum alloy hollow shaft with a nut-like nut and machined bolts of the same material to provide axial preload. Its internal shape is hollow, thus balancing strength and cable threading functionality. The connecting rod of the second robotic arm 35 has an irregular design to avoid shaft interference, allowing the robotic arm to be fully retracted. The second robotic arm 35 is a structure with a single-end machined support and 3D printed parts (using PA6-CF material) for auxiliary support, and its weight is much less than that of a carbon plate and aluminum strip stacked structure.

[0043] The first tumbling joint 361 incorporates a hollow harmonic motor and a slip ring, enabling continuous 360-degree rotation of the first tumbling joint 361 around the axis of the second robotic arm 35. It also provides a cable routing channel, preventing cable accumulation at the end from obstructing the camera's view. The hollow harmonic motor in the first tumbling joint 361 is a DMJH11 type motor, and its control terminal is electrically connected to the output terminal of the control module 4. The first tumbling joint 361 also features a thin-walled bearing to withstand radial forces and protect the harmonic motor; its output shaft is directly connected to the rotor via six M3 bolts and two pins.

[0044] The third pitch joint 362 is equipped with a fourth pitch drive motor, which is a DM4310 type motor. Its control terminal is electrically connected to the output terminal of the control module 4. The fourth pitch drive motor is driven by an S5M synchronous belt and is tensioned by an idler pulley. The fourth pitch drive motor is located behind the first roll joint 361 (i.e., on the side closer to the second robotic arm 35). Power is transmitted to the rotation shaft of the third pitch joint 362 through the synchronous belt, so as to realize the rear-mounted motor to balance the weight of the end effector gripper structure 37.

[0045] The second roll joint 363 is equipped with a roll drive motor, which is a DM4310P type motor. Its control end is electrically connected to the output end of the control module 4. The connecting part of the second roll joint 363 is an L-shaped sheet metal part, and it is equipped with a specially machined bearing seat to position the outer ring of the RAU3008 crossed roller bearing. The connection is equipped with a pin to achieve precise positioning. This design shortens the distance from the pitch axis to the center of the gripper structure 37, increasing the effective working length of the robotic arm by about 80mm, and significantly improving the success rate of high-position (above 650mm) assembly.

[0046] Specifically, each joint of the six-degree-of-freedom robotic arm 3 is equipped with an angle sensor (such as an encoder) to detect the actual rotation angle of each joint in real time and feed it back to the control module 4 to achieve closed-loop control. The yaw gimbal 31 rotates infinitely, the first pitch joint 32 has a rotation range of -80° to +30°, the second pitch joint 34 has a rotation range of -49° to +96°, the first roll joint 361 rotates infinitely, the third pitch joint 362 has a rotation range of ±102°, and the second roll joint 363 has a rotation range of ±100°.

[0047] In this embodiment, please refer to Figure 6 , Figure 6 This is a schematic diagram of a gripper structure provided in an embodiment of the present invention. Figure 6 As shown, the gripper structure 37 includes: a gripper drive motor 371, a crank-slider structure 372, and a gripper 373; The gripper drive motor 371 is located at the end of the second rolling joint 363; The drive shaft of the gripper drive motor 371 is fixedly connected to the first end of the crank-slider structure 372; The second end of the crank-slider structure 372 is fixedly connected to the gripper 373; The control terminal of the gripper drive motor 371 serves as the control terminal of the six-degree-of-freedom robotic arm 3, and is electrically connected to the output terminal of the control module 4.

[0048] In an optional embodiment, when the gripper structure 37 is in the gripping state, the rocker arm and connecting rod within the crank-slider structure 372 form a 90° angle to maximize force transmission efficiency, at which point the gripping force of the gripper 373 reaches 100N. The end of the gripper 373 is rounded to achieve automatic gripping positioning; the gripping surface is a curved surface that matches the shape of the workpiece, and the gripper 373 is printed using both PETG (rigid) and TPU (elastic) materials. Specifically, the gripper structure 37 also includes a vision recognition camera 374, which is mounted on the gripper drive motor 371 and is used to acquire environmental images of the gripper structure 37.

[0049] In the above embodiments, the gripper structure is configured to consist of a gripper drive motor, a crank-slider structure, and grippers. The gripper drive motor is located at the end of the second tumbling joint and drives the crank-slider structure to move via a transmission shaft. The crank-slider structure then transmits the motion to the grippers to achieve opening and closing actions. The crank-slider structure converts the rotational motion of the motor output shaft into the linear reciprocating motion of the grippers. The cooperation between the crank and the slider in the crank-slider structure effectively amplifies the driving force in the gripping state, thereby obtaining sufficient clamping force to firmly hold the workpiece without the need for a high-power motor. Compared to pneumatic grippers or vacuum suction cups, the gripper structure does not require an external air source or vacuum source, avoiding the problem of air pipes and cables accumulating and tangling at the end of the robotic arm, which is beneficial for maintaining a clear end view and tidy cable management. Compared to worm gear or lead screw grippers, the crank-slider structure ensures sufficient clamping force while having a faster response speed and a larger opening and closing stroke, and can adapt to workpieces of different sizes and specifications, thus taking into account multiple operational requirements such as clamping force, response speed, and workpiece adaptability.

[0050] In one alternative embodiment, please refer to Figure 7 , Figure 7 This is a schematic diagram of a master-slave teleoperation controller provided in an embodiment of the present invention. The composite mobile assembly robot also includes a master-slave teleoperation controller, which has the same configuration as the six-degree-of-freedom robotic arm 3, thereby realizing linear mapping of the entire joint stroke. The master-slave teleoperation controller includes a first motor 71, a second motor 72, a third motor 73, a fourth motor 74, a fifth motor 75, a sixth motor 76, and a gripper opening / closing button 77 connected in sequence. Among them, the first motor 71 is a DM4310, which corresponds to the yaw gimbal 31; the second motor 72 is a DM... Motor 4310 corresponds to the first pitch joint 32; the third motor 73 is model DM4310, corresponding to the second pitch joint 34; the fourth motor 74 is model DM4310, corresponding to the first roll joint 361; the fifth motor 75 is model DM4310, corresponding to the third pitch joint 362; the fifth motor 75 is model GM2006, corresponding to the second roll joint 363; torque compensation and force feedback to the operator are provided through the second motor 72, third motor 73, fourth motor 74, fifth motor 75, and sixth motor 76. The yaw shaft system of the first motor 71 uses a 61706 thin-walled bearing to bear the operating torque, thereby protecting the first motor 71. The gripper opening / closing button 77 is used to control the opening and closing of the gripper 373.

[0051] In this embodiment, please refer to Figure 8 , Figure 8 This is a schematic diagram of a positioning tag lifting mechanism provided in an embodiment of the present invention; as shown. Figure 8As shown, the composite mobile assembly robot also includes: a positioning label lifting mechanism 5; the positioning label lifting mechanism 5 is mounted on the chassis frame 1; The positioning tag lifting mechanism 5 includes: a positioning tag lifting platform 51, a lifting mechanism drive motor 52, a first connecting rod 53, and a second connecting rod 54; The lifting mechanism drive motor 52 is mounted on the chassis frame; The drive shaft of the lifting mechanism drive motor 52 is connected to the first connecting rod 53 through a gear transmission mechanism; The drive shaft of the lifting mechanism drive motor 52 is connected to the second connecting rod 54 through a gear transmission mechanism; The first link 53 and the second link 54 are both mounted on the positioning tag lifting platform 51; The control terminal of the lifting mechanism drive motor 52 is electrically connected to the output terminal of the control module 4.

[0052] In an optional embodiment, the stroke of the positioning label lifting platform 51 is 85mm; when the vehicle is upright, the label can be lowered to a working height of approximately 100mm above the ground, avoiding the need to lay the entire vehicle down for each positioning interaction. The first two links of the first link 53 and the second link 54 are machined to ensure lateral rigidity, and polytetrafluoroethylene gaskets are provided at the hinges of the first link 53 and the second link 54 for lubrication. A printing protective cover can also be provided on the outside of the positioning label lifting mechanism 5 to prevent wire tangling.

[0053] In the above embodiments, the positioning tag lifting mechanism includes a positioning tag lifting platform, a lifting mechanism drive motor, a first link, and a second link. The lifting mechanism drive motor simultaneously drives the first link and the second link through a transmission shaft, causing the positioning tag lifting platform to rise and fall relative to the chassis frame. Since the first link and the second link move synchronously under the drive of the lifting mechanism drive motor, the positioning tag lifting platform maintains a horizontal posture throughout the lifting process, preventing the positioning tag from tilting as the platform rises and falls. This ensures that the positioning tag can be accurately identified and measured by the external positioning system at different height positions, realizing automatic deployment and precise positioning height control of the positioning tag, and improving the robot's positioning adaptability and autonomous operation capability in multiple workstations and scenarios.

[0054] In this embodiment, please refer to Figure 9 , Figure 9 This is a schematic diagram of a chassis frame provided in an embodiment of the present invention. Guide wheels 6 are provided on both protruding parts of the chassis frame 1.

[0055] In an optional embodiment, the guide wheel 6 is a polyurethane-coated guide wheel with an A70 hardness.

[0056] In an optional embodiment, the chassis frame 1 is arranged in a shape of the Chinese character "mu", with its upper layer being a robotic arm mounting platform and its lower layer being an electrical compartment (for mounting the control module 4), and the overall dimensions are 594mm in length and 360mm in width. The main body of the chassis frame 1 is composed of two integrated machined parts and two longitudinal aluminum square tubes. The integrated machined parts integrate a cross beam, a leg joint motor mounting seat and a pipe plug structure, and are provided with threaded holes; the longitudinal aluminum square tubes are provided with longitudinal grooves, so that screws can be directly screwed into the inner side to fix various components, which avoids the problems that the traditional stacked aluminum square tube structure requires embedded reinforcing parts and local collapse caused by pre-tightening force. The robotic arm mounting platform is two 20mm×10mm×2mm aluminum square tubes vertically embedded in the frame, and 4mm carbon fiber plates are provided for reinforcement at the stress concentration positions of the notches. A TPU anti-collision plate 11 is arranged at the front end of the chassis frame 1.

[0057] In the above embodiment, guide wheels are arranged on the protruding parts on both sides of the chassis frame. When the robot crosses a step, the guide wheels arranged on both sides of the front end of the chassis frame can first contact the edge of the step. Compared with the direct collision contact between the coated wheel surface of the wheel set and the edge of the step, the guide wheels, with their small diameter and smooth profile, can roll more smoothly along the step edge, guide the front part of the vehicle body to smoothly get onto the step surface, and reduce the impact load and resistance during the obstacle-crossing process; at the same time, when the robot passes through an uneven road surface or a ramp, the guide wheels can serve as auxiliary support points to provide temporary support and guiding effect when the wheel set is suspended or slips, preventing the chassis frame from directly scraping and colliding with the ground; during the traveling process of the robot, the guide wheels can also play a role in buffering and anti-collision. When the robot accidentally touches an obstacle, the guide wheels bear the impact first and make the vehicle body deviate from the obstacle direction through rolling guidance, reducing the damage caused by direct impact to the chassis frame and internal components, thereby improving the safety and reliability of the robot traveling in complex terrain.

[0058] In an optional embodiment, the control module 4 is based on an STM32H723VET6 microcontroller, integrating three FDCAN channels, three USART channels, one I2C channel, and one virtual serial port. The overvoltage protection MOS of the control module 4 is an NCE40P20Q (14mΩ on-resistance), and the reverse connection protection MOS is a CJAB25N04A (6.2mΩ on-resistance). The control module 4 also includes an inertial measurement unit (IMU), which uses a BMI088 six-axis IMU chip, mounted on an FPC flexible board, connected to the control module 4 via a connector, and fixed to the surface of the control module 4 with insulating cotton at intervals. The IMU is powered by a separate LP5907MFX-3.3 low-noise LDO (10μVRMS output noise). Testing showed that the FPC-based solution could raise the temperature from room temperature to 59°C in just 7.7 seconds during temperature control initialization, significantly faster than the approximately 40-second rise time of the onboard solution. Furthermore, in case of damage, only the FPC component needs to be replaced, reducing maintenance costs. The control module 4 also includes a main control power distribution board, a tower structure with stacked terminals on top and bottom double-layer PCBs. The lower layer uses pin headers and sockets to directly connect to the STM32H723VET6 microcontroller, with vertical stacking compensating for horizontal area limitations. The control module 4 also includes a power supply board, which acts as the execution layer to precisely switch on / off and provide overcurrent protection for each power branch of the composite mobile assembly robot. The power switch used is a HYG045P03LQ1C2 (3.8mΩ on-resistance). The power supply board electrically isolates and independently controls the power supply circuit of the leg joint motor 21 from that of the six-degree-of-freedom robotic arm 3. The control module 4 also includes a soft-start module. To address the issue of battery overcurrent protection being triggered by surge current (I=C·dV / dt) when a large capacitive load is powered on, a 22pF capacitor is added between the gate and source of the power MOS to suppress occasional damage to the MOS caused by high power fluctuation spikes.

[0059] In summary, the composite mobile assembly robot provided in this embodiment has a total size of approximately 588mm×553mm×586mm when folded, a maximum extended size of approximately 1269mm×553mm×611mm, a total weight of approximately 25.1kg, a maximum moving speed of approximately 2.8m / s, an average acceleration of approximately 1.2m / s², a maximum step height that can be crossed of 210mm (two consecutive steps can be crossed in stages), and can temporarily store 3 target workpieces at the same time.

[0060] Example 2 Please refer to Figure 10 , Figure 10 The flowchart of a workpiece gripping method for a composite mobile assembly robot provided in an embodiment of the present invention includes steps S101 to S104.

[0061] Step S101: Acquire image data of the target workpiece, preprocess the image data to extract several candidate target point clouds; cluster the several candidate target point clouds to obtain several candidate point cloud clusters.

[0062] Step S102: Calculate the cross-sectional radius distribution of each candidate point cloud cluster to construct the cross-sectional radius distribution curve of each candidate point cloud cluster.

[0063] Step S103: Based on the cross-sectional radius distribution curve, perform geometric morphology screening on the candidate point cloud clusters to determine the pose information of the target workpiece.

[0064] Step S104: Control the composite mobile assembly robot to grasp the target workpiece based on the pose information of the target workpiece.

[0065] This embodiment acquires image data of the target workpiece and preprocesses it to extract candidate target point clouds. The candidate point clouds are then clustered to obtain candidate point cloud clusters. The cross-sectional radius distribution of each cluster is calculated, and a cross-sectional radius distribution curve is constructed. Based on this curve, the clusters are geometrically filtered to determine the pose information of the target workpiece. Finally, the robot is controlled to perform grasping based on the pose information. Since this workpiece grasping method does not rely on the appearance features of the target workpiece, such as color, texture, or surface pattern, but rather on the cross-sectional radius distribution features of the workpiece's three-dimensional geometry for identification and judgment, it does not require pre-collecting a large number of training samples for model training, nor does it require adjusting recognition parameters for different lighting conditions. It has a natural resistance to interference from workpiece surface conditions and changes in ambient lighting. Furthermore, by characterizing the shape of the target workpiece through the compact geometric feature of the cross-sectional radius distribution curve, it features low computational load and fast processing speed, meeting the response speed requirements of real-time grasping operations. Moreover, the geometric shape filtering method allows for flexible adjustment based on the geometric features of different workpiece models, providing good workpiece model adaptability and convenient on-site deployment.

[0066] In this embodiment, image data of the target workpiece is acquired, and the image data is preprocessed to extract several candidate target point clouds; the several candidate target point clouds are clustered to obtain several candidate point cloud clusters, including: Image data of the target workpiece is acquired using a depth camera, and the image data is decoded to obtain several initial 3D point clouds. The depth range of these initial 3D point clouds is then clipped, i.e., depth values ​​exceeding a preset effective range are removed. An initial 3D point cloud (the specific value of which can be set according to actual needs) is generated to remove background and noise points that are too far or too close; then, the initial 3D point cloud after depth range clipping is subjected to voxel grid downsampling to reduce the point cloud density and reduce the amount of computation; next, the downsampled point cloud data is subjected to the Random Sample Consensus (RANSAC) algorithm for planar iterative culling, which is repeated multiple times (3 times in this embodiment) to remove interference from supporting planes such as the ground and workbench, and obtain several candidate target point clouds; finally, the several candidate target point clouds are clustered by Euclidean distance to obtain several candidate point cloud clusters.

[0067] In this embodiment, calculating the cross-sectional radius distribution of each candidate point cloud cluster to construct the cross-sectional radius distribution curve of each candidate point cloud cluster includes: For each candidate point cloud cluster, the covariance matrix of the candidate point cloud cluster is calculated according to a preset principal component analysis algorithm; The covariance matrix is ​​decomposed into eigenvalues ​​to determine the principal axes of the candidate point cloud clusters; Project each of the candidate target point clouds in the candidate point cloud cluster onto the main axis to obtain the projection position of each of the candidate target point clouds on the main axis; The projection position is divided into slices to obtain several slices, and the average distance from all candidate target point clouds in each slice to the principal axis is calculated to obtain the average distance value of each slice. Based on the order of each slice on the main axis, the average distance values ​​of each slice are sorted to form the cross-sectional radius distribution curve of the candidate point cloud cluster.

[0068] In an optional embodiment, for each candidate point cloud cluster, the covariance matrix of the candidate point cloud cluster is calculated according to a preset principal component analysis (PCA) algorithm. Eigenvalue decomposition is performed on the covariance matrix, and the direction of the eigenvector corresponding to the largest eigenvalue is taken as the principal axis direction of the candidate point cloud cluster. Each point in the candidate point cloud cluster is projected onto the principal axis direction to obtain the projection position of each candidate target point cloud on the principal axis. The projection positions are then sliced, i.e., all projection values ​​are arranged from smallest to largest, and the entire projection interval is uniformly divided into K slices (K is a preset constant, K=50 in this embodiment). The average radial distance from all candidate target point clouds in each slice to the principal axis is calculated to obtain the average distance value of each slice. Based on the order of each slice on the principal axis, the average distance values ​​of each slice are sorted to form the cross-sectional radius distribution curve of the candidate point cloud cluster.

[0069] In the above embodiments, principal component analysis (PCA) is used to calculate the covariance matrix and perform eigenvalue decomposition for each candidate point cloud cluster to determine the principal axis of the cluster. Then, each candidate target point cloud in the cluster is projected onto the principal axis to obtain the projected position. The projected position is then sliced, and the average distance from all candidate target point clouds within each slice to the principal axis is calculated. Finally, the average distance values ​​are arranged sequentially along the principal axis to form a cross-sectional radius distribution curve. The longest extension direction of the candidate point cloud cluster, which PCA can accurately extract, is used as the principal axis, ensuring that each slice is perpendicular to the principal axis direction during subsequent slicing. The cross-sectional radius distribution curve can accurately reflect the cross-sectional contour variation of the target workpiece along its length. By dividing the projected position into slices and calculating the average radial distance within each slice, irregular point cloud data can be transformed into a regular one-dimensional radius distribution curve. This preserves the geometric contour information of the target workpiece along the main axis while significantly reducing the data dimensionality. This eliminates the need for complex 3D registration or feature matching of the original high-dimensional point cloud data in subsequent geometric shape screening. As a result, while ensuring recognition accuracy, computational complexity and processing time are significantly reduced, making it suitable for online grasping operations with high real-time requirements.

[0070] In this embodiment, the step of performing geometric morphological screening on the candidate point cloud clusters based on the cross-sectional radius distribution curve to determine the pose information of the target workpiece includes: Identify the maximum radius value, peak distribution, and trough distribution in the cross-sectional radius distribution curve; The expansion ratio of the candidate point cloud clusters is determined based on the peak distribution and the maximum radius value. Based on the expansion ratio, peak distribution, and trough distribution, the candidate point cloud clusters are subjected to peak number condition detection, trough condition detection, expansion ratio condition detection, peak symmetry condition detection, and size constraint detection, respectively. If the candidate point cloud clusters simultaneously satisfy the peak number condition, trough condition, expansion ratio condition, peak symmetry condition, and size constraint, then the candidate point cloud clusters are taken as the point cloud clusters of the target workpiece. The point cloud clusters of the target workpiece are transformed to obtain the pose information of the target workpiece.

[0071] In an optional embodiment, the distribution of peaks and valleys in the cross-sectional radius distribution curve is identified, and the number of peaks and valleys is counted; the valley distribution includes the peak position; the peak distribution includes the peak position; the peak values ​​of each peak in the peak distribution are added together, and then divided by the maximum radius value to obtain the expansion ratio of the candidate point cloud cluster.

[0072] The peak count condition is defined as follows: the number of peaks is equal to 2, and the peak value is greater than 0.6 times the maximum radius. The trough condition is defined as follows: there is only one trough between two peaks. The expansion ratio condition is defined as follows: the expansion ratio is greater than 1.3. The peak symmetry condition is defined as follows: the two peaks are symmetrical about the center point of the candidate point cloud cluster. The size constraint is defined as follows: the height value of the candidate point cloud cluster along the main axis and the maximum width value perpendicular to the main axis fall within the preset target workpiece size range.

[0073] After determining that the peak number condition, trough condition, expansion ratio condition, peak symmetry condition, and size constraint are simultaneously satisfied, the candidate point cloud cluster is taken as the point cloud cluster of the target workpiece. The center point of this candidate point cloud cluster is used as the three-dimensional position of the target workpiece in the camera coordinate system, and the principal axis direction is used as the principal orientation direction of the target workpiece. Six-dimensional pose information containing the three-dimensional position and orientation direction is output.

[0074] In the above embodiments, the peaks and troughs of the cross-sectional radius distribution curve are identified and statistically analyzed. Specifically, this includes detecting the maximum radius value, peak distribution, and trough distribution. The expansion ratio is calculated based on the peak distribution and the maximum radius value. Then, the candidate point cloud clusters are sequentially subjected to peak number condition detection, trough condition detection, expansion ratio condition detection, peak symmetry condition detection, and size constraint detection. Only when all conditions are met is it determined to be a target workpiece. The peak number condition ensures that the target workpiece has a characteristic shape with two protruding ends. The trough condition ensures that there is a concave feature in the middle of the workpiece. The expansion ratio condition quantitatively describes the significance of the protrusion relative to the concavity. The symmetry condition utilizes the geometric characteristics of the two ends of the target workpiece being symmetrical about the center. The size constraint excludes non-target objects from the absolute scale, effectively eliminating the interference of non-target objects and greatly reducing the false detection rate and false detection rate. Thus, it achieves real-time accurate identification and pose determination of the target workpiece without training data and robust to changes in illumination and surface state.

[0075] In one optional embodiment, the composite mobile assembly robot is controlled to grasp the target workpiece based on its pose information. This includes: converting the pose information into a robot base coordinate system (which can be obtained through calibration of hand-eye extrinsic parameters; the base-to-end-effector transformation is calculated in real-time by forward kinematics and broadcast via TF); then, using the OMPL library's RRTConnect algorithm to solve for a collision-free path in joint space, with the target pose set to maintain the current end-effector posture and only replacing position components to constrain the inverse solution space; specifically, to address the problem of planning interruption due to strict matching of joint state timestamps (failure occurs with a time difference of 3-4ms), this embodiment employs a custom joint state caching mechanism, i.e., independently subscribing to joint state topics, caching the latest joint angles in callbacks, and manually constructing the robot state before planning, bypassing the internal timestamp matching logic. The above process of controlling the composite mobile assembly robot to grasp the target workpiece can be adjusted according to actual needs, and this embodiment does not impose too many limitations on it.

[0076] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A composite mobile assembly robot, characterized in that, include: The chassis frame, four sets of single-degree-of-freedom wheel legs symmetrically mounted on both sides of the chassis frame, a six-degree-of-freedom robotic arm mounted on the chassis frame, and a control module; The single-degree-of-freedom wheel leg includes a leg joint motor, a leg body, a rudder system structure, and a wheel assembly; the leg joint motor is fixedly connected to the chassis frame; the drive shaft of the leg joint motor is fixedly connected to the leg body, and is used to drive the leg body to swing around a horizontal axis; the rudder system structure is disposed on the leg body and is fixedly connected to the wheel assembly. The six-degree-of-freedom robotic arm, from bottom to end, comprises: a yaw gimbal, a first pitch joint, a first robotic arm, a second pitch joint, a second robotic arm, an end effector assembly, and a gripper structure. The yaw gimbal is mounted on the chassis frame; the first end of the first robotic arm is hinged to the yaw gimbal via the first pitch joint; the second end of the first robotic arm is hinged to the first end of the second robotic arm via the second pitch joint; the second end of the second robotic arm is fixedly connected to the end effector assembly; the gripper structure is located at the end of the end effector assembly. The control terminals of the leg joint motor, the rudder system, the wheel assembly, and the six-degree-of-freedom robotic arm are all electrically connected to the output terminal of the control module; the control module is used to control the composite mobile assembly robot to move and grasp workpieces.

2. The composite mobile assembly robot as described in claim 1, characterized in that, The rudder system structure includes: a rudder motor, a wheel assembly bracket, and a bearing mounting base; The rudder motor is mounted on the leg body; the drive shaft of the rudder motor is fixedly connected to the wheel assembly bracket; The wheel bracket and the wheel set are fixedly connected; The bearing mounting base is disposed at the end of the leg body, and the bearing mounting base is slidably connected to the wheel set bracket; The control terminal of the rudder motor serves as the control terminal of the rudder system structure, and it is electrically connected to the output terminal of the control module.

3. The composite mobile assembly robot as described in claim 2, characterized in that, The wheelset includes a wheel motor and a support wheel; the wheel motor is fixedly connected to the wheelset bracket; the drive shaft of the wheel motor is fixedly connected to the support wheel; the control terminal of the wheel motor serves as the control terminal of the wheelset and is electrically connected to the output terminal of the control module.

4. The composite mobile assembly robot as described in claim 1, characterized in that, The end effector assembly includes: a first roll joint, a third pitch joint, and a second roll joint; The first tumbling joint is located between the second end of the second robotic arm and the third pitch joint; The third pitch joint is disposed between the first roll joint and the second roll joint, and is used to drive the second roll joint to pitch and rotate. The second roll joint is disposed between the third pitch joint and the gripper structure; The second rolling joint is fixedly connected to the gripper structure and is used to drive the gripper structure to rotate about its own axis.

5. A composite mobile assembly robot as described in claim 4, characterized in that, The gripper structure includes: a gripper drive motor, a crank-slider structure, and grippers; The gripper drive motor is located at the end of the second rolling joint; The drive shaft of the gripper drive motor is fixedly connected to the first end of the crank-slider structure; The second end of the crank-slider structure is fixedly connected to the gripper. The control terminal of the gripper drive motor serves as the control terminal of the six-degree-of-freedom robotic arm, and it is electrically connected to the output terminal of the control module.

6. A composite mobile assembly robot as described in any one of claims 1 to 5, characterized in that, Also includes: Positioning tag lifting mechanism; The positioning tag lifting mechanism is mounted on the chassis frame; The positioning tag lifting mechanism includes: a positioning tag lifting platform, a lifting mechanism drive motor, a first link, and a second link; The lifting mechanism drive motor is mounted on the chassis frame; The drive shaft of the lifting mechanism's drive motor is connected to the first connecting rod via a gear transmission mechanism. The drive shaft of the lifting mechanism's drive motor is connected to the second connecting rod via a gear transmission mechanism; Both the first link and the second link are mounted on the positioning tag lifting platform; The control terminal of the lifting mechanism drive motor is electrically connected to the output terminal of the control module.

7. A composite mobile assembly robot as described in any one of claims 1 to 5, characterized in that, Guide wheels are provided on both sides of the protruding parts of the chassis frame.

8. A workpiece gripping method for a composite mobile assembly robot, characterized in that, A composite mobile assembly robot as described in any one of claims 1 to 7, comprising: Image data of the target workpiece is acquired, and the image data is preprocessed to extract several candidate target point clouds; the several candidate target point clouds are clustered to obtain several candidate point cloud clusters. Calculate the cross-sectional radius distribution of each candidate point cloud cluster to construct the cross-sectional radius distribution curve of each candidate point cloud cluster; Based on the cross-sectional radius distribution curve, the candidate point cloud clusters are geometrically filtered to determine the pose information of the target workpiece. Based on the pose information of the target workpiece, the composite mobile assembly robot is controlled to grasp the target workpiece.

9. The workpiece gripping method of a composite mobile assembly robot as described in claim 8, characterized in that, The step of calculating the cross-sectional radius distribution of each candidate point cloud cluster to construct the cross-sectional radius distribution curve of each candidate point cloud cluster includes: For each candidate point cloud cluster, the covariance matrix of the candidate point cloud cluster is calculated according to a preset principal component analysis algorithm; The covariance matrix is ​​decomposed into eigenvalues ​​to determine the principal axes of the candidate point cloud clusters; Project each of the candidate target point clouds in the candidate point cloud cluster onto the main axis to obtain the projection position of each of the candidate target point clouds on the main axis; The projection position is divided into slices to obtain several slices, and the average distance from all candidate target point clouds in each slice to the principal axis is calculated to obtain the average distance value of each slice. Based on the order of each slice on the main axis, the average distance values ​​of each slice are sorted to form the cross-sectional radius distribution curve of the candidate point cloud cluster.

10. The workpiece gripping method of a composite mobile assembly robot as described in claim 9, characterized in that, The step of performing geometric morphological screening of the candidate point cloud clusters based on the cross-sectional radius distribution curve to determine the pose information of the target workpiece includes: Identify the maximum radius value, peak distribution, and trough distribution in the cross-sectional radius distribution curve; The expansion ratio of the candidate point cloud clusters is determined based on the peak distribution and the maximum radius value. Based on the expansion ratio, peak distribution, and trough distribution, the candidate point cloud clusters are subjected to peak number condition detection, trough condition detection, expansion ratio condition detection, peak symmetry condition detection, and size constraint detection, respectively. If the candidate point cloud clusters simultaneously satisfy the peak number condition, trough condition, expansion ratio condition, peak symmetry condition, and size constraint, then the candidate point cloud clusters are taken as the point cloud clusters of the target workpiece. The point cloud clusters of the target workpiece are transformed to obtain the pose information of the target workpiece.

Images