Multi-scenario obstacle-crossing adaptive climbing inspection robot and control method thereof

Abstract

The present invention belongs to the field of robots. In order to solve the problem that the current inspection robots are not suitable for high-rise steel structures, a multi-scenario obstacle-crossing adaptive climbing inspection robot and a control method thereof are provided. Among them, the multi-scenario obstacle-crossing adaptive climbing inspection robot includes at least three sections of mobile mechanisms, and any two adjacent sections of the mobile mechanisms are connected by a node connecting rod; the middle section of the node connecting rod is provided with a ball-jointed multi-degree-of-freedom servo active joint; each section of the mobile mechanism includes at least a control part, a support part, a sensing part and a climbing part; the climbing part includes at least a pair of side arm connecting rods, which are symmetrically arranged on both sides of the support part; the end of each side arm connecting rod is connected to a dynamic magnetic wheel through a wrist-like joint; an adaptive passive wheel is also provided at the bottom of the support part, and the adaptive passive wheel and the dynamic magnetic wheels on both sides thereof form a three-point clamping structure. It can meet the visual full coverage path planning and spatial precise positioning requirements of high-rise steel structures.

Description

Technical Field

[0001] The present invention belongs to the field of robots, and in particular relates to a multi-scenario obstacle-crossing adaptive climbing inspection robot and a control method thereof. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] High-rise steel structure generally refers to a structure with more than six floors (or more than 30 meters), which is mainly made of steel and steel plates connected or welded into components, and then connected and welded. High-rise steel structure often uses steel frame structure, which is an important load-bearing structural system and is widely used in public buildings such as residential, office, and entertainment.

[0004] High-rise steel structures have the characteristics of large number of components, variable sizes, complex connection forms and mutual obstruction, which brings challenges to the inspection of high-rise steel structures. The current inspection robots cannot meet the requirements of visual full coverage path planning and spatial precise positioning of high-rise steel structures in terms of load capacity, obstacle crossing ability and environmental adaptability. Summary of the invention

[0005] In order to solve the technical problems existing in the above-mentioned background technology, the present invention provides a multi-scenario obstacle-crossing adaptive climbing inspection robot and a control method thereof, which can meet the requirements of visual full-coverage path planning and spatial precise positioning of high-rise steel structures.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] A first aspect of the present invention provides a multi-scenario obstacle-crossing adaptive climbing inspection robot.

[0008] A multi-scenario obstacle-crossing adaptive climbing inspection robot comprises: at least three moving mechanisms, any two adjacent moving mechanisms are connected by a node connecting rod; the middle section of the node connecting rod is provided with a ball-jointed multi-degree-of-freedom steering gear active joint;

[0009] Each section of the mobile mechanism at least includes a control part, a support part, a sensing part and a climbing part; the control part is arranged in the support part, and the sensing part and the climbing part are both connected to the control part in communication;

[0010] The climbing part comprises at least a pair of side arm connecting rods, which are symmetrically arranged on both sides of the supporting part; the end of each side arm connecting rod is connected to the dynamic magnetic attraction wheel through a wrist-like joint;

[0011] An adaptive passive wheel is also provided at the bottom of the support portion, and the adaptive passive wheel and the dynamic magnetic wheels on both sides thereof form a three-point clamping structure;

[0012] The sensing unit at least includes an obstacle sensing element and a distance measuring element, which are used to sense obstacle information and detect the distance of the obstacle respectively and transmit both to the control unit.

[0013] As an embodiment, the adaptive passive wheel includes: an adaptive spring, a wheel frame, a bearing, a magnetic passive wheel and a connecting plate; one end of the adaptive spring is connected to the connecting plate, and the other end is connected to the top of the wheel frame; the connecting plate is fixed to the bottom of the support; the bearing is fixed to the end of the wheel frame, and the magnetic passive wheel is installed at both ends of the bearing. As an embodiment, the dynamic magnetic wheel includes a flexible wheel surface, and the flexible wheel surface is wrapped with a Halbach array electromagnet; the Halbach array electromagnet is wrapped with a coil, and the coil is connected to a power module; the Halbach array electromagnet is used to generate a magnetic field strength when the coil is energized, so that the dynamic magnetic wheel is adsorbed on the steel structure.

[0014] As an implementation mode, the adsorption force of the dynamic magnetic wheel is determined by the strength of the magnetic field, and the strength of the magnetic field is controlled by the output current of the power module; wherein, the output current of the power module is controlled by dynamically adjusting the duty cycle of the PWM signal; the duty cycle of the PWM signal is dynamically adjusted according to the current adsorption force, the set target adsorption force and the external load change using a fuzzy rule algorithm.

[0015] As an implementation method, the expression for dynamically adjusting the duty cycle of the PWM signal using the fuzzy rule algorithm is:

[0016] ;

[0017] ;

[0018] ; ;

[0019] ;

[0020] ;

[0021] in, represents the possible value range of the output variable, Represents the membership function of multi-rule fuzzy reasoning; and are adsorption force and load change, respectively; Indicates that the climbing inspection robot is in the Fuzzy reasoning membership function under different postures; Indicates the number of posture types of the climbing inspection robot; and represent the membership functions of adsorption force and load variation, respectively; It is the corresponding variable parameter of the climbing inspection robot in different postures.

[0022] As an implementation mode, the ball-jointed multi-degree-of-freedom servo active joint includes a flexible shell, two active joint servos and two rigid joint connecting rods; all the active joint servos are arranged in the flexible shell, the active joint servos correspond to the rigid joint connecting rods one by one, and the corresponding active joint servos are coupled and hinged with the rigid joint connecting rods; one end of the rigid joint connecting rod also extends out of the flexible shell and is connected to the node connecting rod.

[0023] As an implementation mode, the obstacle sensing element is fixed to a support portion setting position of the head segment moving mechanism and the tail segment moving mechanism of the multi-scenario obstacle-crossing adaptive climbing inspection robot; and the distance measuring element is installed on each dynamic magnetic wheel.

[0024] A second aspect of the present invention provides a control method for a multi-scenario obstacle-crossing adaptive climbing inspection robot.

[0025] A control method for a multi-scenario obstacle-crossing adaptive climbing inspection robot, comprising:

[0026] Determine the next obstacle avoidance operation to be performed based on the currently perceived obstacle information and the detected obstacle distance;

[0027] Based on the current posture of the climbing inspection robot and the obstacle avoidance operation to be performed next, the posture of the climbing inspection robot is controlled so that the climbing inspection robot can successfully avoid obstacles;

[0028] The process of obstacle avoidance operation is as follows:

[0029] Control the active joint of the ball-jointed multi-degree-of-freedom servo to drive the connecting rod and the head section moving mechanism to move, so that the head section moving mechanism is lifted, and at the same time control the dynamic magnetic wheel of the head section moving mechanism to lift up to avoid obstacles, control the climbing inspection robot to move forward, so that the head section moving mechanism crosses the obstacle, and control the middle section moving mechanism to prepare for lifting;

[0030] The dynamic magnetic attraction wheel of the head section moving mechanism is controlled to restore to its original state to fit the adsorption steel structure, and the ball-jointed multi-degree-of-freedom servo active joint is controlled to drive the connecting rod and the head section moving mechanism to move, so that the head section moving mechanism is clamped at three points, and at the same time, the middle section moving mechanism is controlled to be lifted, and the dynamic magnetic attraction wheel of the middle section moving mechanism is controlled to be lifted;

[0031] Control the climbing inspection robot to move forward, make the middle section moving mechanism pass over the obstacle, control the tail moving mechanism to prepare to lift up; then control the middle section moving mechanism to restore the original moving state and the tail moving mechanism to lift up and pass over the obstacle; finally control the tail moving mechanism to restore the original moving state.

[0032] As an implementation method, before controlling the posture of the climbing inspection robot, a deep reinforcement learning method is used to determine the posture adjustment action based on the currently perceived obstacle information and obstacle distance; wherein the reward function expression of the deep reinforcement learning method is:

[0033] ;

[0034] in, is the normalized value of the current dynamic magnetic wheel contact area, and its value range is , used to measure the fit between the dynamic magnetic wheel and the steel structure; is the theoretical maximum contact area, used for normalization; weight coefficient Balance the weights between contact area optimization, posture stability, distance adaptability and energy consumption; For the The real-time distance measurement value of each distance measurement element, Indicates the number of distance measuring elements; is the ideal distance measurement value, is the smoothing parameter of the ranging deviation; attitude angle Indicates the pitch and roll angles of the current attitude; are the pitch angle and roll angle corresponding to the ideal stable posture respectively; is the current power consumption, is the maximum power consumption threshold, used to normalize the power consumption data.

[0035] As an implementation method, the wrist-like joint adjustment angle is calculated according to the distance between the wheel surface of the dynamic magnetic wheel and the surface of the steel structure, so that the dynamic magnetic wheel actively fits the steel structure; the wrist-like joint adjustment angle The calculation formula is:

[0036] ;in, is the distance between the wheel surface of the dynamic magnetic wheel and the surface of the steel structure; It is the vertical distance from the wrist joint axis to the surface of the steel structure; It is the length from the wrist joint to the wheel surface of the dynamic magnetic wheel.

[0037] The beneficial effects of the present invention are:

[0038] (1) In order to solve the problem that the movement of the inspection robot is hindered by the complex geometric connections and various obstacles of the steel structure, the present invention proposes an electromagnetic adsorption climbing structure based on multi-joint drive and deformable mechanism, and uses a ball-jointed multi-degree-of-freedom servo drive connecting rod combined with a dynamic magnetic adsorption component to improve the maneuverability and adaptability of the robot in complex truss and grid structures, and can achieve a peak stable movement of 4m / min.

[0039] (2) The present invention determines the obstacle avoidance operation to be performed next based on the currently perceived obstacle information and the detected obstacle distance, and then controls the next posture of the climbing inspection robot based on the current posture of the climbing inspection robot, so that the climbing inspection robot can successfully avoid obstacles. In addition, in each step, at least two sections of the vehicle body are on the steel structure, which ensures the stability and safety of the climbing inspection robot during the steel structure inspection process.

[0040] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0042] Figure 1 2 is a schematic diagram of the structure of a multi-scenario obstacle-crossing adaptive climbing inspection robot according to an embodiment of the present invention;

[0043] Figure 2 is a schematic diagram of a three-point clamping structure according to an embodiment of the present invention;

[0044] Figure 3 2 is a schematic structural diagram of a ball-jointed multi-degree-of-freedom steering gear active joint according to an embodiment of the present invention;

[0045] Figure 4 is a schematic diagram of a state of a wrist-like joint according to an embodiment of the present invention;

[0046] Figure 5 is a schematic diagram of another state of the wrist-like joint according to an embodiment of the present invention;

[0047] Figure 6 is a schematic structural diagram of an adaptive passive wheel according to an embodiment of the present invention;

[0048] Figure 7 It is a schematic structural diagram of a dynamic magnetic attraction wheel according to an embodiment of the present invention.

[0049] Among them, 1. distance measuring element; 2. dynamic magnetic wheel; 3. wrist-like joint; 4. side arm connecting rod; 5. node connecting rod; 6. vehicle body connecting arm; 7. depth camera; 8. gimbal; 9. positioning system; 10. ball-jointed multi-degree-of-freedom servo active joint; 11. LED lighting system; 12. obstacle sensing element; 13. adaptive passive wheel; 1301. adaptive spring; 1302. wheel frame; 1303. bearing; 1304. magnetic passive wheel; 1305. connecting plate; 1001. active joint servo; 1002. rigid joint connecting rod; 1003. flexible shell; 1004. coupling hinge; 301. rigid connecting rod; 302. servo; 303. rigid shell; 3041, first bevel gear; 3042, second bevel gear; 3043, third bevel gear; 14, support portion; 15, steel structure; 201, flexible wheel surface; 202, Halbach array electromagnet. DETAILED DESCRIPTION

[0050] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0051] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0052] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0053] Figure 1 Schematic diagram of the structure of the multi-scenario obstacle-crossing adaptive climbing inspection robot according to an embodiment of the present invention. Figure 1 As shown, a multi-scenario obstacle-crossing adaptive climbing inspection robot according to an embodiment of the present invention comprises: a multi-scenario obstacle-crossing adaptive climbing inspection robot comprises: at least three sections of mobile mechanisms, any two adjacent sections of the mobile mechanisms are connected by a node connecting rod 5; the middle section of the node connecting rod 5 is provided with a ball-jointed multi-degree-of-freedom steering gear active joint 10;

[0054] Each section of the mobile mechanism at least includes a control unit, a support unit 14, a sensing unit and a climbing unit; the control unit is arranged in the support unit 14, and the sensing unit and the climbing unit are both connected to the control unit for communication;

[0055] The climbing part includes at least a pair of side arm connecting rods 4, which are symmetrically arranged on both sides of the support part 14; the end of each side arm connecting rod 4 is connected to the dynamic magnetic attraction wheel 2 through a wrist-like joint 3;

[0056] The bottom of the support portion 14 is also provided with an adaptive passive wheel 13, and the adaptive passive wheel 13 and the dynamic magnetic wheels 2 on both sides thereof clamp the steel structure 15 together to form a three-point clamping structure. Figure 2 As shown;

[0057] The sensing unit at least includes an obstacle sensing element 12 and a distance measuring element 1, which are used to sense obstacle information and detect the distance of the obstacle respectively and transmit both to the control unit.

[0058] In a specific implementation process, the side arm connecting rod 4 is also hingedly connected to both sides of the supporting part through the vehicle body connecting arm 6.

[0059] exist Figure 1 In the embodiment, the obstacle sensing element 12 is fixed to the support part setting position of the head segment moving mechanism and the tail segment moving mechanism of the multi-scenario obstacle-crossing adaptive climbing inspection robot; the distance measuring element 1 is installed on each dynamic magnetic wheel 2. An LED lighting system 11 is also arranged around the obstacle sensing element 12 to provide corresponding brightness to the obstacle sensing element 12.

[0060] In some other embodiments, the sensing unit further includes a depth camera 7, which is used to collect images during the inspection process and transmit them to the control unit. The depth camera 7 is mounted on the moving mechanism of the middle section through a pan-tilt platform 8.

[0061] In some embodiments, the moving mechanism of the middle section is also equipped with a positioning system 9 for locating the position information of the current climbing inspection robot.

[0062] In this embodiment, a flexible synaptic silicone wheel surface can be provided on the outer side of the dynamic magnetic wheel. The flexible synaptic silicone wheel surface and the adaptive passive wheel form a stable three-point clamping configuration, which can increase the magnetic adsorption area and increase stability.

[0063] In this embodiment, if Figure 6 As shown, the adaptive passive wheel 13 includes: an adaptive spring 1301, a wheel frame 1302, a bearing 1303, a magnetic passive wheel 1304 and a connecting plate 1305; one end of the adaptive spring 1301 is connected to the connecting plate 1305, and the other end is connected to the top of the wheel frame 1302; the connecting plate 1305 is fixed to the bottom of the support part 14; the bearing 1303 is fixed to the end of the wheel frame 1302, and the magnetic passive wheel 1304 is installed at both ends of the bearing 1303. Among them, the adaptive spring can enable the robot to directly adaptively pass through smaller obstacles encountered during the inspection process.

[0064] Specifically, the dynamic magnetic wheel 2 includes a flexible wheel surface 201, and the flexible wheel surface 201 is wrapped with a Halbach array electromagnet 202. Figure 7 As shown; a coil is wound around the outside of the Halbach array electromagnet, and the coil is connected to a power module; the Halbach array electromagnet is used to generate a magnetic field strength when the coil is energized, so that the magnetic adsorption wheel is adsorbed on the steel structure 15.

[0065] During the specific implementation process, the adsorption force of the dynamic magnetic wheel is determined by the magnetic field strength, and the magnetic field strength is controlled by the output current of the power module; wherein, the output current of the power module is controlled by dynamically adjusting the duty cycle of the PWM signal; the duty cycle of the PWM signal is dynamically adjusted according to the current adsorption force, the set target adsorption force and the external load change using a fuzzy rule algorithm.

[0066] Among them, the fuzzy rule algorithm is used to dynamically adjust the PWM signal duty cycle The expression is:

[0067] ;

[0068] ;

[0069] ; ;

[0070] ;

[0071] ;

[0072] in, represents the possible value range of the output variable, Represents the membership function of multi-rule fuzzy reasoning; and are adsorption force and load change, respectively; Indicates that the climbing inspection robot is in the Fuzzy reasoning membership function under different postures; Indicates the number of posture types of the climbing inspection robot; and represent the membership functions of adsorption force and load variation, respectively; It is the corresponding variable parameter of the climbing inspection robot in different postures.

[0073] The postures of the climbing inspection robot include: normal moving state, obstacle crossing state, set angle component switching state, set obstacle passing state and emergency docking state.

[0074] For example, during normal travel, the adsorption force required is low and stable, and the load variation range is small. Take:

[0075] ;

[0076] In the obstacle crossing state, the adsorption force requirement increases and the load changes greatly, so take:

[0077] ;

[0078] When setting the angle component switching, the adsorption force requirement is high and the load changes dramatically. Take:

[0079] ;

[0080] When setting obstacles to pass (such as vertical climbing or horizontal sharp turns), the adsorption force requirement is extremely high and the load may have extreme values. Take: ;

[0081] In emergency situations (such as extreme tilt), the suction force needs to be maximized and the load fluctuation range is larger. , Represents the robot's own gravity.

[0082] It should be noted here that These four parameters can be set according to actual conditions.

[0083] In this embodiment, the adsorption force can be divided into three ranges, namely low adsorption force, medium adsorption force and high adsorption force; these three ranges are used These two parameters determine that when the adsorption force is less than or equal to When the adsorption force is greater than and less than or equal to When the adsorption force is greater than When the load changes, it is considered to have high adsorption capacity. These two parameters are determined; when the load change amplitude is greater than or equal to and less than When the load changes more than , it is a case of load increase.

[0084] For example, if the suction force is a "low suction force" load and is "increasing", the PWM signal should be adjusted high;

[0085] If the adsorption force is "medium adsorption force" load and is "stable", the PWM signal is maintained at a medium level;

[0086] If the suction force is "high suction force", the PWM signal is adjusted low.

[0087] Then, the fuzzy value of the input variable is converted into the fuzzy value of the PWM signal by using the fuzzy inference system, and the results of multiple rules are combined for reasoning. Finally, through defuzzification, the result of fuzzy inference is converted into the actual PWM duty cycle, so as to adjust the current of the electromagnet and realize the precise control of the magnetic field strength, thereby optimizing the performance and stability of electromagnetic adsorption.

[0088] In this embodiment, the actual duty cycle Adjusting the PWM signal , the formula is: .in, is the PWM signal period, and finally adjusts the duty cycle Control the current of the electromagnet to accurately optimize the magnetic field strength and the stability of electromagnetic adsorption.

[0089] Specifically, adjust the duty cycle The process of controlling the current of the electromagnet is:

[0090] Known duty cycle Maximum current at , combined with the duty cycle ,according to , the current of the electromagnet can be obtained .

[0091] like Figure 3 As shown, the ball-jointed multi-DOF servo active joint 10 includes a flexible housing 1003, two active joint servos 1001 and two rigid joint links 1002; all active joint servos 1001 are arranged in the flexible housing 1003, the active joint servos 1001 are connected to the rigid joint links 1002 through coupling hinges 1004, and the corresponding active joint servos 1001 are coupled and hinged with the rigid joint links 1002; one end of the rigid joint link 1002 also extends out of the flexible housing 1003 and is connected to the node link. In this way, the robot can deform and adjust its posture to realize the inspection path through various complex steel structures.

[0092] In this embodiment, if Figure 4 and Figure 5As shown, the wrist-like joint 3 includes a rigid housing 303 and bevel gears, wherein the bevel gears include a first bevel gear 3041, a second bevel gear 3042 and a third bevel gear 3043, wherein the first bevel gear 3041 and the third bevel gear 3043 are both meshed with the second bevel gear 3042; and the first bevel gear 3041 and the third bevel gear 3043 are respectively connected to a steering gear 302, and each steering gear 302 is connected to a rigid connecting rod 301; the steering gear is used to drive the corresponding bevel gear to mesh and rotate with the second bevel gear. In this way, the steering gear drives the bevel gear to rotate, so as to adjust the dynamic magnetic wheel posture to fit the steel structure.

[0093] In other embodiments, a permanent magnet protection mechanism is further provided inside the support part; the permanent magnet protection mechanism is in a non-working state; the permanent magnet protection mechanism is connected to a control module; the control module is used to obtain the magnetic field strength of the Halbach array electromagnet and compare it with a preset magnetic field strength threshold to control the permanent magnet protection mechanism to switch to a working state, so that the climbing inspection robot is adsorbed on the steel structure.

[0094] Among them, when the climbing inspection robot is inspecting normally, the permanent magnet protection mechanism is in a non-working state; when the climbing inspection robot is powered off due to insufficient power or an emergency, the permanent magnet protection mechanism switches to a working state.

[0095] In some embodiments, the permanent magnet protection mechanism includes a connecting disk, a permanent magnet and a driving module; the upper end of the connecting disk is fixedly connected to the inside of the climbing robot, the lower end of the connecting disk is connected to the driving module, the driving module is connected to the permanent magnet and the control module, and the driving module is used to drive the magnet to rotate under the action of the control module; when the permanent magnet protection mechanism is in a non-working state, the permanent magnet is on the side away from the steel structure; when the permanent magnet protection mechanism is switched to a working state, the driving module is used to drive the permanent magnet to rotate to the side close to the steel structure so as to be adsorbed on the steel structure.

[0096] The connection disk, permanent magnet and drive module here can all be implemented using existing structures.

[0097] In other embodiments, the structure of the permanent magnet protection mechanism may also be implemented using other structures, which will not be described in detail here.

[0098] This embodiment utilizes a permanent magnet protection mechanism to prevent the weakening of the adsorption force and the climbing inspection robot from falling due to a weakening or sudden drop in magnetic field strength when the climbing inspection robot is short of power or a sudden power outage, thereby improving the operating stability of the climbing inspection robot.

[0099] The control method of the multi-scenario obstacle-crossing adaptive climbing inspection robot described above includes:

[0100] Step 1: Determine the next obstacle avoidance operation to be performed based on the currently perceived obstacle information and the detected obstacle distance;

[0101] Step 2: Based on the current posture of the climbing inspection robot and the obstacle avoidance operation to be performed in the next step, the posture of the climbing inspection robot is controlled so that the climbing inspection robot can successfully avoid obstacles;

[0102] The process of obstacle avoidance operation is as follows:

[0103] Control the active joint of the ball-jointed multi-degree-of-freedom servo to drive the connecting rod and the head section moving mechanism to move, so that the head section moving mechanism is lifted, and at the same time control the dynamic magnetic wheel of the head section moving mechanism to lift up to avoid obstacles, control the climbing inspection robot to move forward, so that the head section moving mechanism crosses the obstacle, and control the middle section moving mechanism to prepare for lifting;

[0104] The dynamic magnetic attraction wheel of the head section moving mechanism is controlled to restore to its original state to fit the adsorption steel structure, and the ball-jointed multi-degree-of-freedom servo active joint is controlled to drive the connecting rod and the head section moving mechanism to move, so that the head section moving mechanism is clamped at three points, and at the same time, the middle section moving mechanism is controlled to be lifted, and the dynamic magnetic attraction wheel of the middle section moving mechanism is controlled to be lifted;

[0105] Control the climbing inspection robot to move forward, make the middle section moving mechanism pass over the obstacle, control the tail moving mechanism to prepare to lift up; then control the middle section moving mechanism to restore the original moving state and the tail moving mechanism to lift up and pass over the obstacle; finally control the tail moving mechanism to restore the original moving state.

[0106] Specifically, before controlling the posture of the climbing inspection robot, the posture adjustment action is determined based on the currently perceived obstacle information and obstacle distance using a deep reinforcement learning method; wherein the reward function expression of the deep reinforcement learning method is:

[0107] ;

[0108] in, is the normalized value of the current dynamic magnetic wheel contact area, and its value range is , used to measure the fit between the dynamic magnetic wheel and the steel structure; is the theoretical maximum contact area, used for normalization; weight coefficient Balance the weights between contact area optimization, posture stability, distance adaptability and energy consumption; For the The real-time distance measurement value of each distance measurement element, Indicates the number of distance measuring elements; is the ideal distance measurement value, is the smoothing parameter of the ranging deviation; attitude angle Indicates the pitch and roll angles of the current attitude; are the pitch angle and roll angle corresponding to the ideal stable posture respectively; is the current power consumption, is the maximum power consumption threshold, used to normalize the power consumption data.

[0109] This embodiment uses a deep reinforcement learning method to accurately and quickly determine posture adjustment actions, thereby improving the inspection efficiency of the climbing inspection robot.

[0110] In some embodiments, an active and passive adaptive fitting method is also used to determine the adjustment angle of the wrist-like joint;

[0111] Specifically, the active and passive adaptive fitting method is as follows: according to the distance between the wheel surface of the dynamic magnetic wheel and the surface of the steel structure, the wrist-like joint adjustment angle is calculated, so that the dynamic magnetic wheel actively fits the steel structure; the wrist-like joint adjustment angle The calculation formula is:

[0112] ;in, is the distance between the wheel surface of the dynamic magnetic wheel and the surface of the steel structure; It is the vertical distance from the wrist joint axis to the surface of the steel structure; It is the length from the wrist joint to the wheel surface of the dynamic magnetic wheel.

[0113] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A control method for a multi-scenario obstacle-crossing adaptive climbing inspection robot, characterized in that: The multi-scenario obstacle-crossing adaptive climbing inspection robot comprises: at least three sections of mobile mechanisms, any two adjacent sections of the mobile mechanisms are connected by a node connecting rod; the middle section of the node connecting rod is provided with a ball-jointed multi-degree-of-freedom steering gear active joint; Each section of the mobile mechanism at least includes a control part, a support part, a sensing part and a climbing part; the control part is arranged in the support part, and the sensing part and the climbing part are both connected to the control part in communication; The climbing part comprises at least a pair of side arm connecting rods, which are symmetrically arranged on both sides of the supporting part; the end of each side arm connecting rod is connected to the dynamic magnetic attraction wheel through a wrist-like joint; An adaptive passive wheel is also provided at the bottom of the support portion, and the adaptive passive wheel and the dynamic magnetic wheels on both sides thereof form a three-point clamping structure; The sensing unit at least includes an obstacle sensing element and a distance measuring element, which are used to sense obstacle information and detect the distance of the obstacle respectively and transmit both to the control unit; The control method of the multi-scenario obstacle-crossing adaptive climbing inspection robot includes: Determine the next obstacle avoidance operation to be performed based on the currently perceived obstacle information and the detected obstacle distance; Based on the current posture of the climbing inspection robot and the obstacle avoidance operation to be performed next, the posture of the climbing inspection robot is controlled so that the climbing inspection robot can successfully avoid obstacles; The process of obstacle avoidance operation is as follows: Control the active joint of the ball-jointed multi-degree-of-freedom servo to drive the connecting rod and the head section moving mechanism to move, so that the head section moving mechanism is lifted, and at the same time control the dynamic magnetic wheel of the head section moving mechanism to lift up to avoid obstacles, control the climbing inspection robot to move forward, so that the head section moving mechanism crosses the obstacle, and control the middle section moving mechanism to prepare for lifting; The dynamic magnetic attraction wheel of the head section moving mechanism is controlled to restore to its original state to fit the adsorption steel structure, and the ball-jointed multi-degree-of-freedom servo active joint is controlled to drive the connecting rod and the head section moving mechanism to move, so that the head section moving mechanism is clamped at three points, and at the same time, the middle section moving mechanism is controlled to be lifted, and the dynamic magnetic attraction wheel of the middle section moving mechanism is controlled to be lifted; Control the climbing inspection robot to move forward, so that the middle section moving mechanism can cross the obstacle, and control the tail moving mechanism to prepare to lift up; then control the middle section moving mechanism to restore the original moving state and the tail moving mechanism to lift up and cross the obstacle; finally control the tail moving mechanism to restore the original moving state; Before controlling the posture of the climbing inspection robot, the deep reinforcement learning method is used to determine the posture adjustment action based on the currently perceived obstacle information and obstacle distance; the reward function expression of the deep reinforcement learning method is: ； in, is the normalized value of the current dynamic magnetic wheel contact area, and its value range is , used to measure the fit between the dynamic magnetic wheel and the steel structure; is the theoretical maximum contact area, used for normalization; weight coefficient Balance the weights between contact area optimization, posture stability, distance adaptability and energy consumption; For the The real-time distance measurement value of each distance measurement element, Indicates the number of distance measuring elements; is the ideal distance measurement value, is the smoothing parameter of the ranging deviation; attitude angle Indicates the pitch and roll angles of the current attitude; are the pitch angle and roll angle corresponding to the ideal stable posture respectively; is the current power consumption, is the maximum power consumption threshold, used to normalize the power consumption data.

2. The control method of the multi-scenario obstacle-crossing adaptive climbing inspection robot according to claim 1 is characterized in that: According to the distance between the wheel surface of the dynamic magnetic wheel and the surface of the steel structure, the wrist-like joint adjustment angle is calculated to make the dynamic magnetic wheel actively fit the steel structure; the wrist-like joint adjustment angle The calculation formula is: ;in, is the distance between the wheel surface of the dynamic magnetic wheel and the surface of the steel structure; It is the vertical distance from the wrist joint axis to the surface of the steel structure; It is the length from the wrist joint to the wheel surface of the dynamic magnetic wheel.

3. The control method of the multi-scenario obstacle-crossing adaptive climbing inspection robot according to claim 1 is characterized in that: The adaptive passive wheel includes: an adaptive spring, a wheel frame, a bearing, a magnetic passive wheel and a connecting plate; one end of the adaptive spring is connected to the connecting plate, and the other end is connected to the top of the wheel frame; the connecting plate is fixed to the bottom of the support part; the bearing is fixed to the end of the wheel frame, and the magnetic passive wheel is installed at both ends of the bearing.

4. The control method of the multi-scenario obstacle-crossing adaptive climbing inspection robot according to claim 1 is characterized in that: The dynamic magnetic attraction wheel includes a flexible wheel surface, in which a Halbach array electromagnet is wrapped; a coil is wound around the Halbach array electromagnet, and the coil is connected to a power module; the Halbach array electromagnet is used to generate a magnetic field strength when the coil is energized, so that the dynamic magnetic attraction wheel is adsorbed on the steel structure.

5. The control method of the multi-scenario obstacle-crossing adaptive climbing inspection robot according to claim 1 or 4, characterized in that: The adsorption force of the dynamic magnetic wheel is determined by the magnetic field strength, and the magnetic field strength is controlled by the output current of the power module; wherein the output current of the power module is controlled by dynamically adjusting the duty cycle of the PWM signal; the duty cycle of the PWM signal is dynamically adjusted according to the current adsorption force, the set target adsorption force and the external load change using a fuzzy rule algorithm.

6. The control method of the multi-scenario obstacle-crossing adaptive climbing inspection robot according to claim 5 is characterized in that: Dynamically adjust the duty cycle of PWM signal using fuzzy rule algorithm The expression is: ； ； ； ； ； ； in, represents the possible value range of the output variable, Represents the membership function of multi-rule fuzzy reasoning; and are adsorption force and load change, respectively; Indicates that the climbing inspection robot is in the Fuzzy reasoning membership function under different postures; Indicates the number of posture types of the climbing inspection robot; and represent the membership functions of adsorption force and load variation, respectively; It is the corresponding variable parameter of the climbing inspection robot in different postures.

7. The control method of the multi-scenario obstacle-crossing adaptive climbing inspection robot according to claim 1, characterized in that: The ball-jointed multi-degree-of-freedom steering gear active joint comprises a flexible housing, two active joint steering gears and two rigid joint connecting rods; All active joint servos are arranged in a flexible shell, the active joint servos correspond to the rigid joint connecting rods one by one, and the corresponding active joint servos are coupled and hinged with the rigid joint connecting rods; one end of the rigid joint connecting rod also extends out of the flexible shell and is connected to the node connecting rod.

8. The control method of the multi-scenario obstacle-crossing adaptive climbing inspection robot according to claim 1, characterized in that: The obstacle sensing element is fixed at the support part setting position of the head section moving mechanism and the tail section moving mechanism of the multi-scenario obstacle-crossing adaptive climbing inspection robot; the distance measuring element is installed on each dynamic magnetic wheel.

Images