Continuous body robot and body motion posture real-time sensing method thereof

Abstract

The invention belongs to the field of robot motion posture sensing, and particularly discloses a continuum robot and a body motion posture real-time sensing method thereof.The continuum robot comprises a corrugated pipe backbone, a framework, a cable and a camera system, the inner wall and the outer wall of the corrugated pipe backbone are each provided with a corrugated structure, and a base is arranged at the bottom of the corrugated pipe backbone; the plurality of frameworks are axially and uniformly sleeved and fixed on the outer side of the corrugated pipe backbone; the upper end of the cable is fixed on the framework at the top end and sequentially penetrates through the frameworks on the lower side, and the lower end of the cable penetrates out of the base; the camera system is installed on the base and located in the corrugated pipe backbone. Movement is achieved by pulling the cable, meanwhile, images are obtained through the camera system, and then the position and direction of any point on the robot body are output based on the neural network. According to the invention, sensing of the whole-body motion attitude of the continuum robot is realized, and the method can be applied to motion control and interaction of the continuum robot.

Description

Technical Field

[0001] The present invention belongs to the field of robot motion posture perception, and more specifically, relates to a continuum robot and a real-time perception method of its body motion posture. Background Art

[0002] The proprioceptive perception of continuum robots is very important for their motion control and safe interaction. The posture perception of existing continuum robots focuses on the position information of the end points in space, and the overall perception of the whole body information relies on multiple distributed discrete sensors, which brings challenges to the structural design and wiring of the robot.

[0003] Therefore, there is an urgent need for a new continuum robot that can conveniently, stably and accurately perceive the overall motion posture information of its body in real time. Summary of the invention

[0004] In view of the above defects or improvement needs of the prior art, the present invention provides a continuum robot and a real-time perception method of its body motion posture, the purpose of which is to achieve real-time perception of the robot's whole body posture information.

[0005] To achieve the above object, according to one aspect of the present invention, a continuum robot is provided, comprising a bellows backbone, a skeleton, a cable and a camera system, wherein:

[0006] The inner and outer walls of the bellows trunk both have a corrugated structure, and a base is provided at the bottom;

[0007] A plurality of the skeletons are axially and evenly sleeved and fixed on the outside of the bellows backbone;

[0008] The upper end of the cable is fixed on the top frame and passes through the lower frames in sequence, and the lower end of the cable passes through the base;

[0009] The camera system is mounted on the base and is located within the bellows backbone.

[0010] As a further preference, the skeleton comprises a skeleton sheet and a clamping plate, wherein the inner side of the skeleton sheet is nested in the corrugated structure of the corrugated tube backbone, and the outer side is provided with a hole for the cable to pass through; two clamping plates fix the skeleton sheet from the upper and lower sides.

[0011] As a further preference, a skeleton comprises N skeleton plates, which are evenly arranged circumferentially on the corrugated tube backbone; correspondingly, N cables are provided, and each cable connects all skeleton plates at the same circumferential position.

[0012] As further preferred, N=3.

[0013] As a further preference, a conduit is embedded in the hole on the outer side of the skeleton piece, and the cable passes through the conduit and can slide freely in the conduit; a curved pipe is provided on the lower side of the cable, and the cable passes through the curved pipe; the upper end of the curved pipe is fixed in the hole of the bottom skeleton piece, and the lower end is fixed on the base.

[0014] As a further preference, the cable is made of PE wire, and the conduit and the elbow are both made of Teflon material.

[0015] As a further preference, the device further comprises a lighting system, which is arranged around the camera system.

[0016] As a further preferred embodiment, the lighting system includes a PCB board with a hollow center and LED lamp beads, the PCB board is installed on a base, and the lens of the camera system is arranged at the hollow center of the PCB board; a plurality of the LED lamp beads are arranged on the PCB board.

[0017] According to another aspect of the present invention, a method for real-time perception of the body motion posture of the above-mentioned continuum robot is provided, comprising the following steps:

[0018] A series of target points are taken in the axial direction of the bellows backbone, and the relative positions of the target points on the bellows backbone are determined based on the distances of the target points from the bottom of the bellows backbone in the axial direction;

[0019] The bellows backbone image acquired by the camera system and the relative position of the target point on the bellows backbone are input into the trained convolutional neural network model to obtain the spatial position coordinates of the target point on the bellows backbone;

[0020] The spatial position coordinates of a series of target points are obtained to determine the posture of the continuum robot.

[0021] As a further preferred embodiment, after the bellows backbone image and the relative position of the target point on the bellows backbone are input into the trained convolutional neural network model: the image is flattened into a 512-dimensional vector, and the vector is multiplied with the relative position of the target point on the bellows backbone to obtain a new vector; the new vector is merged with the relative position of the target point on the bellows backbone to obtain a 513-dimensional feature vector; and the spatial position coordinates of the target point are output according to the feature vector.

[0022] In general, the above technical solution conceived by the present invention has the following technical advantages compared with the prior art:

[0023] 1. The bellows backbone designed in the present invention takes into account both structural stability and the ability of flexible movement, and provides installation space and information for the built-in camera system; combined with the visual feedback arranged inside the body, a single camera can achieve real-time perception of the robot's whole body posture information, and has broad application prospects.

[0024] 2. The driving form of the continuum robot of the present invention is universal and is not limited to linear drive, pneumatic or other driving forms; and the structure is simple and compact.

[0025] 3. The real-time perception method of the continuum robot's body motion posture proposed in the present invention combines the camera image and the relative position of the target point, and can select the target point by itself, so as to realize the spatial motion posture perception of any point on the whole body of the continuum robot, rather than being limited to the end point or a specific limited point. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a schematic diagram of the overall structure of a continuum robot according to an embodiment of the present invention;

[0027] Figure 2 This is a schematic diagram of the assembly structure of the bellows backbone and the frame according to an embodiment of the present invention;

[0028] Figure 3 Schematic diagram of the skeleton structure of an embodiment of the present invention, wherein (a) is an exploded view of the skeleton; (b) is a schematic diagram of a cross-section of the skeleton;

[0029] Figure 4 Schematic diagram of the base structure of an embodiment of the present invention, wherein (a) is a schematic diagram of the base; (b) is an exploded diagram of the base;

[0030] Figure 5 This is an exploded view of the continuum robot structure according to an embodiment of the present invention.

[0031] In all the drawings, the same figure numbers are used to represent the same elements or structures, among which: 1-corrugated tube backbone, 2-skeleton, 21-skeleton sheet, 22-splint, 3-base, 31-base frame, 32-base cover, 4-cable assembly, 41-cable, 42-conduit, 43-bend pipe, 5-lighting system, 51-PCB board, 52-LED lamp beads, 6-camera system, 61-lens, 62-control board, 7-top cover. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0033] A continuum robot provided by an embodiment of the present invention is as follows: Figure 1 and Figure 5As shown, it includes a bellows backbone 1, a skeleton 2, a base 3, a cable assembly 4, a lighting system 5, a camera system 6 and a top cover 7, wherein:

[0034] The inner and outer walls of the bellows trunk 1 have obvious corrugated structures. When the bellows trunk 1 is deformed by movement, the corrugations on the inner and outer walls of the bellows trunk will also be deformed. Preferably, the bellows trunk 1 is made of white material as a whole, and black paint is sprayed on the outside to reduce the influence of the intensity of the external environment light when the built-in camera system is working. The top cover 7 is fixed to the top of the bellows trunk 1 with threaded fasteners, and the base 3 is set at the bottom of the bellows trunk 1.

[0035] In the axial direction of the bellows backbone 1, a plurality of skeletons 2 are evenly and equidistantly arranged on the bellows backbone 1, such as Figure 2 The skeleton 2 includes a skeleton sheet 21 and a splint 22. Figure 3 As shown, a group of skeletons 2 includes N skeleton sheets 21, which are evenly arranged on the corrugated pipe backbone 1 in the circumferential direction, and the inner side of the skeleton sheets 21 is nested in the corrugated structure of the corrugated pipe backbone 1; in this embodiment, N=3, that is, 3 skeleton sheets 21 are arranged at a spacing of 120° on the circumference; the 3 skeleton sheets 21 are sandwiched by 2 plywood 22 and clamped by bolts to fix a group of skeletons 2 on the corrugated pipe backbone 1.

[0036] The base 3 includes a base frame 31 and a base cover 32 thereon. The base frame 31 reserves a plurality of holes for the built-in camera system 6. The wiring of the lighting system 5 is used for power supply and signal transmission. The base 3 is preferably made of 3D printing material that can withstand high temperatures of 200°C.

[0037] The cable assembly 4 includes a cable 41, a conduit 42 and a bend 43. The number of cables 41 corresponds to the number of skeleton sheets 21 in a group of skeletons 2, that is, each cable 41 is connected to all skeleton sheets 21 at the same circumferential position, and the cable 41 is preferably made of PE wire; the conduit 42 is fixed by glue in the round holes of all skeleton sheets 21 except the first skeleton section at the bottom; one end of the bend 43 is fixed by glue in the round hole of the first skeleton section 21 at the bottom, and the other end of the bend 43 passes through the round hole of the base frame 31. The upper side of the cable 41 is fixed on the skeleton sheet 2 at the top, and passes through the conduit 42 in each skeleton section 21 downward in turn, and finally passes through the bend 43, and the lower side of the cable 41 extends out from the bend 43, and the cable 41 can slide freely in the conduit 42 and the bend 43. The conduit 42 and the bend 43 are preferably made of Teflon material, which can greatly reduce the friction with the cable 41. By driving the cable 41, the free bending movement of the corrugated tube backbone 1 in space can be realized.

[0038] The camera system 6 includes a lens 61 and a control panel 62. Figure 4As shown, the lens 61 is glued to the base cover 32 and is located inside the bellows backbone 1. It can capture the deformation of the corrugated structure of the inner wall of the bellows in real time during the movement of the continuum robot; the lens 61 is connected to the control board 62 through a flexible cable, and there are four protruding cones in the base frame 31. The structural dimensions of the cones match the mounting holes on the control board 62. The control board 62 is fixed on the cones in the base frame 31, and the wiring of the control board 62 is led out from the reserved holes on the base frame 31.

[0039] The lighting system 5 includes a PCB board 51 with a hollow center and LED lamp beads 52. The PCB board 51 is fixed on the base cover 32. The lens 61 is located at the hollow center of the PCB board 51. Four LED lamp beads 52 are welded on the PCB board 51. The lighting system 5 is connected to wires that are led out from the reserved holes on the base frame 31. When the lighting system 5 is powered on, it can illuminate the internal environment of the corrugated tube backbone 1.

[0040] The embodiment of the present invention provides a method for real-time perception of the motion posture of a continuum robot, comprising the following steps:

[0041] Take a series of target points (selected as needed) on the axial direction of the bellows backbone, and use the length of the target point from the bottom along the axial direction of the bellows as the relative position of the target point on the bellows backbone; specifically, the length can be normalized to a real number ranging from 0 to 1, that is, a relative position of 0 represents a target point at the bottom, and a relative position of 1 represents a target point at the end;

[0042] When the continuum robot moves, the camera system collects the internal image information of the bellows backbone in real time;

[0043] The internal image information and the relative position of the target point on the bellows backbone are input into the trained convolutional neural network model, and the spatial position coordinates of the target point are output;

[0044] Specifically, after the relative positions of the internal image and the target point on the bellows backbone are input into the trained convolutional neural network model, the image passes through the convolutional network and is flattened into a 512-dimensional vector, which is multiplied by the relative position of the target point on the bellows backbone to obtain a new vector; the relative position of the target point on the bellows backbone is merged into the new vector to obtain a 513-dimensional feature vector; the feature vector passes through the fully connected layer to output the spatial position coordinates of the target point.

[0045] By inputting a series of relative position sequences of an unlimited number of target points on the bellows backbone into the convolutional neural network model, the convolutional neural network model can simultaneously output a series of spatial coordinates of an unlimited number of target points. Ultimately, the real-time perception of the body motion posture of the continuum robot based on visual information feedback is achieved.

[0046] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A continuum robot, characterized in that: It comprises a bellows backbone (1), a frame (2), a cable (41) and a camera system (6), wherein: The inner and outer walls of the bellows trunk (1) both have a corrugated structure, and a base (3) is provided at the bottom; A plurality of the skeletons (2) are axially evenly sleeved and fixed on the outside of the bellows backbone (1); The upper end of the cable (41) is fixed on the top frame (2) and passes through the lower frames (2) in sequence, and the lower end of the cable (41) passes through the base (3); The camera system (6) is mounted on the base (3) and is located inside the bellows backbone (1).

2. The continuum robot according to claim 1, characterized in that: The frame (2) comprises a frame piece (21) and a clamping plate (22); the frame piece (21) is nested inside the corrugated structure of the corrugated tube backbone (1), and is provided with a hole on the outside for the cable (41) to pass through; two clamping plates (22) fix the frame piece (21) from the upper and lower sides.

3. The continuum robot according to claim 2, characterized in that: A skeleton (2) comprises N skeleton sheets (21), which are evenly arranged circumferentially on the corrugated tube backbone (1); and N cables (41) are correspondingly provided, each cable (41) connecting all skeleton sheets (21) at the same circumferential position.

4. The continuum robot according to claim 3, characterized in that: N＝3。 5. The continuum robot according to claim 3, characterized in that: A conduit (42) is embedded in the hole on the outside of the skeleton sheet (21), and the cable (41) passes through the conduit (42) and can slide freely in the conduit (42); a curved tube (43) is provided on the lower side of the cable (41), and the cable (41) passes through the curved tube (43); the upper end of the curved tube (43) is fixed in the hole of the bottom skeleton sheet (21), and the lower end is fixed on the base (3).

6. The continuum robot according to claim 5, characterized in that: The cable (41) is made of PE wire, and the conduit (42) and the elbow (43) are both made of Teflon material.

7. The continuum robot according to any one of claims 1 to 6, characterized in that: It also includes a lighting system (5), which is arranged around the camera system (6).

8. The continuum robot according to claim 7, characterized in that: The lighting system (5) comprises a PCB board (51) with a hollowed-out middle portion and LED lamp beads (52); the PCB board (51) is mounted on a base (3); a lens of the camera system (6) is arranged at the hollowed-out middle portion of the PCB board (51); and a plurality of the LED lamp beads (52) are arranged on the PCB board (51).

9. A method for real-time perception of the body motion posture of a continuum robot as claimed in any one of claims 1 to 8, characterized in that: The steps include: A series of target points are taken in the axial direction of the bellows backbone (1), and the relative positions of the target points on the bellows backbone (1) are determined based on the distances of the target points in the axial direction from the bottom of the bellows backbone (1); The image of the bellows trunk (1) acquired by the camera system (6) and the relative position of the target point on the bellows trunk (1) are input into the trained convolutional neural network model to obtain the spatial position coordinates of the target point on the bellows trunk (1); The spatial position coordinates of a series of target points are obtained to determine the posture of the continuum robot.

10. The method for real-time perception of the body motion posture of a continuum robot according to claim 9, characterized in that: After the bellows backbone (1) image and the relative position of the target point on the bellows backbone (1) are input into the trained convolutional neural network model: the image is flattened into a 512-dimensional vector, and the vector is multiplied with the relative position of the target point on the bellows backbone (1) to obtain a new vector; the new vector is combined with the relative position of the target point on the bellows backbone (1) to obtain a 513-dimensional feature vector; and the spatial position coordinates of the target point are output based on the feature vector.

Images