Multi-degree-of-freedom robot and method for steel structure high-altitude roof climbing inspection

Abstract

The present application belongs to the field of robots, and aims to solve the problem that the current inspection robots cannot meet the demand of climbing and inspecting large steel structure high-altitude roof, and provides a multi-degree-of-freedom robot and method for climbing and inspecting steel structure high-altitude roof. The multi-degree-of-freedom robot for climbing and inspecting steel structure high-altitude roof comprises at least two mobile chassis, and a multi-degree-of-freedom passive spherical hinge is used to connect every two adjacent mobile chassis. Each mobile chassis comprises at least a control module, a motion assembly and a sensing assembly. The sensing assembly is used to sense the current steel structure diameter, the distance information of the steel structure and the robot pose information and transmit them to the control module. The control module is used to adaptively fit the steel structure according to the current steel structure diameter and the multi-degree-of-freedom of the motion assembly, and control the motion assembly to perform the climbing and inspection task according to the distance information of the steel structure and the robot pose information. The robot can meet the intelligent inspection demand of large steel structure high-altitude roof.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of robots, and particularly relates to a multi-degree-of-freedom robot and method for climbing and inspecting a high-altitude roof of a steel structure. BACKGROUND

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

[0003] A large high-altitude roof of a steel structure refers to a high-altitude load-bearing roof structure constructed by welding, bolting, etc. of steel pipes and other steel materials, commonly found in public buildings such as stadiums, exhibition centers, and large factories, and characterized by large span, variable curvature, dense components, and high-altitude suspended working environment. The long-term safe operation of such structures relies on regular inspection, but their own characteristics pose many challenges to the inspection work: first, the roof height usually exceeds ten meters, and some large venues even reach several dozen meters, requiring the construction of high-altitude working platforms for manual inspection, which is high-risk and low-efficiency; second, the roof surface is mostly inclined, curved, or grid-shaped, with complex connections between components, making it difficult for traditional ground inspection equipment to reach; third, the high-altitude environment is significantly affected by wind and temperature changes, requiring high stability and anti-interference capability of the inspection equipment. Current inspection robots cannot meet the needs of climbing and inspecting large high-altitude roofs of steel structures in terms of adaptability to steel structures of different curvatures and stable movement on complex steel structure surfaces. SUMMARY

[0004] To solve at least one of the technical problems in the background art, the present application provides a multi-degree-of-freedom robot and method for climbing and inspecting a high-altitude roof of a steel structure, which can meet the intelligent inspection needs of large high-altitude roofs of steel structures.

[0005] To achieve the above-mentioned purposes, the present application adopts the following technical solutions:

[0006] The first aspect of the present application provides a multi-degree-of-freedom robot for climbing and inspecting a high-altitude roof of a steel structure.

[0007] A multi-degree-of-freedom robot for climbing and inspecting a high-altitude roof of a steel structure, comprising: at least two mobile chassis, each two adjacent mobile chassis being connected by a multi-degree-of-freedom passive spherical hinge; each mobile chassis comprising at least a control module, a motion component, and a perception component;

[0008] The motion component and the perception component are in communication connection with the control module; the perception component is used for perceiving the current steel structure diameter, distance information of the steel structure and robot pose information and transmitting to the control module; the control module is used for adapting to the steel structure according to the current steel structure diameter and the multi-degree-of-freedom of the motion component, and controlling the motion component to perform the climbing inspection task according to the distance information of the steel structure and the robot pose information.

[0009] As an implementation form, the motion component comprises at least two multi-degree-of-freedom mechanical arm claws, and the multi-degree-of-freedom mechanical arm claws are symmetrically arranged on two sides of the mobile chassis; the end of each multi-degree-of-freedom mechanical arm claw is connected with a permanent magnetic adsorption wheel through one side of a hand wheel frame, and the other side of the hand wheel frame is sequentially connected with an elbow mechanism, an arm mechanism and a shoulder mechanism, and the shoulder mechanism is installed on the mobile chassis.

[0010] As an implementation form, the shoulder mechanism, the arm mechanism, the elbow mechanism and the hand wheel frame all have independent degrees of freedom, and together constitute four degrees of freedom of the mechanical arm claw.

[0011] As an implementation form, one end of the shoulder mechanism is connected with the mobile chassis, and the other end is connected with the arm mechanism, and the shoulder mechanism can realize the rotational degree of freedom around the side of the mobile chassis.

[0012] The end of the arm mechanism away from the shoulder mechanism is connected with the elbow mechanism, and the arm mechanism can realize the telescopic degree of freedom along the axial direction of the arm mechanism.

[0013] The elbow mechanism is connected with the hand wheel frame, the elbow mechanism can realize the bending degree of freedom perpendicular to the direction of the arm mechanism, and the hand wheel frame can realize the rotational degree of freedom relative to the elbow mechanism.

[0014] As an implementation form, the mobile chassis is further provided with a passive omni-directional wheel, and the passive omni-directional wheel and the permanent magnetic adsorption wheels on two sides thereof constitute a three-point support structure.

[0015] As an implementation form, the permanent magnetic adsorption wheel comprises a wheel body shell, a permanent magnet group, a wheel shaft and a wear-resistant wheel surface; the wheel shaft penetrates through the center of the wheel body shell and is rotationally connected with the hand wheel frame at two ends; the wear-resistant wheel surface is wrapped outside the wheel body shell, and the surface is provided with anti-skid lines; the permanent magnet group is used for generating a continuous magnetic field, so that the adsorption wheel is stably adsorbed on the steel structure surface.

[0016] The second aspect of the application provides a control method of a multi-degree-of-freedom robot for steel structure high-altitude roof climbing inspection.

[0017] The control method of the multi-degree-of-freedom robot for steel structure high-altitude roof climbing inspection comprises the following steps.

[0018] Based on the current diameter of the steel structure component and the multiple degrees of freedom of the robot arm, control the multi-degree-of-freedom robot to adaptively fit the steel structure component;

[0019] Based on the distance information of the steel structure components and the robot's pose information, the operating state of the multi-degree-of-freedom robot is predicted and controlled using a safety barrier tubular model based on adhesion margin, so as to control the motion safety and stability of the multi-degree-of-freedom robot in real time during inspection work.

[0020] As one implementation method, the tubular model of the safety barrier based on adhesion margin is as follows:

[0021] ;

[0022] in, Describes the minimum value function; Represents the 2-norm; subscript Let be the weight matrix corresponding to the 2-norm; for Control increment at any given moment; For the desired tangential velocity reference; This is the adhesion margin reward coefficient; for Adhesion safety margin at all times; For reference only; The radial error between the robot and the center of the steel pipe; It is the lateral tilt angle; for The tangential velocity along the tube at any given moment; This is the initial control variable; The tube center state at the end of the prediction domain for the nominal trajectory obtained by rolling prediction; The range of values ​​is ; It is a positive integer greater than or equal to 1.

[0023] As one implementation method, a non-magnetic vision-inertial fusion method constrained by a servo encoder is used to estimate robot pose information, thereby fully decoupling the robot pose information from the heading. The objective function of the non-magnetic vision-inertial fusion method constrained by the servo encoder is:

[0024] ;

[0025] in, Describes the minimum value function; Represents the L2 norm; This is a set of internal states and external parameters within the sliding window; For IMU pre-integrated residuals; The kinematic consistency residual is formed by the forward kinematics of the encoder angle and the fixed contact point / zero-velocity contact of the adsorbed end; subscript The covariance / weight corresponding to the L2 norm; For visual residuals; The normal alignment residual is obtained by cross-product of the normal of the terminal adsorption surface with the normal of the point cloud surface after the normal is rotated to the world system.

[0026] As one implementation method, during the process of controlling the multi-degree-of-freedom robot to adaptively fit the steel structure, the arm mechanism and elbow mechanism are driven to rotate by corresponding angles to ensure that the robot remains in the optimal fitting state; wherein, the required rotation angle of the arm mechanism... The required rotation angle of the elbow mechanism The calculation formula is:

[0027] ;

[0028] ;

[0029] ;

[0030] in, The radius of the steel structural component; The vertical distance from the shoulder pivot point to the bottom contact point of the omnidirectional passive wheel; The distance from the vertical centerline of the moving chassis to the shoulder axis point; The length of the elbow mechanism; The normal phase angle; For intermediate parameters; The height of the permanent magnet adsorption wheel.

[0031] The beneficial effects of this invention are:

[0032] This invention utilizes sensing components to perceive the diameter and distance information of the current steel structure component, as well as the robot's pose information. A control module, based on the diameter of the steel structure component and the multiple degrees of freedom of the robot's arm, controls the multi-degree-of-freedom robot to adaptively conform to the steel structure component. Furthermore, based on the distance information of the steel structure component and the robot's pose information, a tubular model of a safety barrier based on adhesion margin is used to predict and control the operating state of the multi-degree-of-freedom robot. This allows for real-time control of the robot's motion safety and stability during inspection work, achieving multi-degree-of-freedom adaptive conformation of the robot's arm to the steel structure component and meeting the intelligent inspection requirements of large steel structure high-altitude roofs.

[0033] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0034] The accompanying drawings, which form a part of this specification, are included to provide a further understanding of the application and are incorporated herein by reference. The embodiments of the present application, as described in the specification and annexed drawings, are illustrative of the application and are not meant to be an unwarranted limitation of the application.

[0035] Figure 1 Figure 1 is a schematic diagram of the overall structure of a multi-DOF robot for climbing and inspection of large steel structure high-altitude roof according to an embodiment of the present application;

[0036] Figure 2 Figure 2 is a top view of a multi-DOF robot for climbing and inspection of large steel structure high-altitude roof according to an embodiment of the present application;

[0037] Figure 3 Figure 3 is a bottom view of a multi-DOF robot for climbing and inspection of large steel structure high-altitude roof according to an embodiment of the present application;

[0038] Figure 4 Figure 4 is a front view of a multi-DOF robot for climbing and inspection of large steel structure high-altitude roof according to an embodiment of the present application;

[0039] Figure 5 Figure 5 is a schematic diagram of the radius detection principle according to an embodiment of the present application;

[0040] Figure 6 Figure 6 is a schematic diagram of the inverse kinematics model of the mechanical arm claw according to an embodiment of the present application;

[0041] Figure 7 Figure 7 is a flowchart of the non-magnetic vision-inertial fusion pose estimation based on the encoder constraint of the steering wheel according to an embodiment of the present application;

[0042] Figure 8 Figure 8 is a flowchart of the safety barrier tubular model predictive control based on the adhesion margin according to an embodiment of the present application.

[0043] Wherein, 1, permanent magnetic adsorption wheel; 2, passive omnidirectional wheel; 3, perception component; 301, central ranging module; 302, two-wing ranging module; 4, power module moving chassis; 5, multi-DOF passive spherical hinge; 6, control module moving chassis; 7, four-DOF mechanical arm claw; 701, shoulder mechanism; 702, arm mechanism; 703, elbow mechanism; 704, hand wheel carrier. DETAILED DESCRIPTION

[0044] The application will be further described below in conjunction with the drawings and embodiments.

[0045] It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the application. Unless otherwise indicated, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0046] It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments in accordance with the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, steps, operations, elements, components, and / or groups thereof, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups thereof.

[0047] In the present application, the terms such as "upper", "lower", "left", "right", "front", "back", "vertical", "horizontal", "side", "bottom" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, which is only a relationship word determined for the purpose of describing the structural relationship of the components or elements of the present application, and cannot be understood as a limitation of the present application.

[0048] In the present application, the terms such as "fixedly connected", "connected", "connected" and the like should be understood in a broad sense, which means that it can be fixedly connected, integrally connected or detachably connected, and it can be directly connected or indirectly connected through an intermediate medium. For relevant scientific or technical personnel in the art, the specific meaning of the above terms in the present application can be determined according to the specific circumstances, and cannot be understood as a limitation of the present application.

[0049] In one or more embodiments, a multi-degree-of-freedom robot for high-altitude roof climbing and inspection of a steel structure includes: at least two mobile chassis, each two adjacent mobile chassis being connected by a multi-degree-of-freedom passive spherical hinge; each mobile chassis at least including a control module, a motion assembly and a perception assembly;

[0050] The motion assembly and the perception assembly are respectively in communication connection with the control module; the perception assembly is used for perceiving current steel structure diameter, distance information of the steel structure and robot pose information and transmitting to the control module; the control module is used for adapting to the steel structure according to the current steel structure diameter and the multi-degree-of-freedom of the motion assembly, and controlling the motion assembly to perform the climbing and inspection task according to the distance information of the steel structure and the robot pose information.

[0051] In the following, taking two mobile chassis as an example, Figure 1 is a whole structure schematic diagram of a multi-degree-of-freedom robot for high-altitude roof climbing and inspection of a large steel structure according to an embodiment of the present application, which includes two mobile chassis, namely a power module mobile chassis 4 carrying a power module and a control module mobile chassis 6 carrying a control module, and the two mobile chassis are connected by a multi-degree-of-freedom passive spherical hinge 5.

[0052] Combined with Figure 1 , Figure 2 , Figure 3 and Figure 4As shown, each mobile chassis includes two four-degree-of-freedom mechanical arm claws 7 arranged on both sides of the mobile chassis; the four-degree-of-freedom mechanical arm claw 7 includes a shoulder mechanism 701, an arm mechanism 702, an elbow mechanism 703 and a hand wheel frame 704; one side of the shoulder mechanism 701 is connected with the mobile chassis, and the other side is connected with the arm mechanism 702; the other side of the arm mechanism 702 is connected with the elbow mechanism 703; the other side of the elbow mechanism 703 is connected with the hand wheel frame 704; the other side of the hand wheel frame 704 is connected with the permanent magnetic adsorption wheel 1. Each mobile chassis is provided with a passive omnidirectional wheel 2 and a sensing assembly 3; the passive omnidirectional wheel 2 is arranged in two at the bottom of the mobile chassis along the head-tail direction of the robot; the sensing assembly 3 includes a central ranging module 301 and two wing ranging modules 302, and the two wing ranging modules 302 are arranged at the outer side of the robot on the upper surface of the mobile chassis.

[0053] The movement process of the multi-degree-of-freedom robot of the embodiment of the application mainly includes three kinds: axial movement when the hand wheel frame moves forward, circumferential movement when the hand wheel frame is rotated to make the magnetic wheel move in a direction perpendicular to the axial movement direction when the robot needs to be adjusted in the circumferential position on the steel structure surface, and lifting and moving the mechanical arm claw to the surface of the target steel structure by rotating the shoulder joint and other joints when the robot needs to move to other steel structure parts.

[0054] The following will be described in combination with Figure 1 The multi-degree-of-freedom robot shown in the figure is used for climbing and inspecting a large steel structure high-altitude roof, and the control process is given as follows:

[0055] Step 1: according to the diameter of the current steel structure part and the multi-degree-of-freedom of the robot arm claw, the multi-degree-of-freedom robot is controlled to adaptively fit the steel structure part.

[0056] In the process of controlling the multi-degree-of-freedom robot to adaptively fit the steel structure part, the radius and diameter information of the steel structure part are obtained by the difference between the average value obtained by the two wing ranging modules of the sensing assembly and the ranging value of the central ranging module.

[0057] Figure 5 is the radius detection principle diagram of the embodiment of the application, the radius of the steel structure part is detected by the sensing assembly; the measurement value of the central ranging module is denoted as , the average value of the measurement values of the two wing ranging modules is denoted as , the distance between the central ranging module and the two wing ranging modules is denoted as , and the difference between is denoted as ; the radius of the steel structure part can be calculated by the following equation:

[0058] ;

[0059] Wherein, R is the radius of the steel structure.

[0060] According to the rotation angle of each mechanism of the arm claw analyzed by the inverse kinematics model of the mechanical arm claw, the forward motion control is carried out through the driving rudder, and the magnetic wheel is in the best fitting posture and is fitted on the surface of the steel structure.

[0061] Figure 6 It is the inverse kinematics model diagram of the mechanical arm claw of the embodiment of the application, in the embodiment, the four-degree-of-freedom mechanical arm claw is suitable for the inverse kinematics problem of the mechanical arm claw of the multi-curvature steel structure, and the inverse kinematics problem has constraint conditions:

[0062] For the fitting position, the P point is located on the cylindrical surface of the steel pipe, and has:

[0063] ;

[0064] Wherein, O point is the center point, P point is the wheel bottom contact point, and R is the radius of the steel structure.

[0065] For the best fitting position, that is, the bottom of the magnetic wheel is tangent to the steel pipe, in the geometric model, the short side of the magnetic wheel is tangent to the tangent line of the steel pipe circle, and the center line of the forearm is along the radius direction:

[0066] ; ;

[0067] Wherein, B point is the elbow shaft, C point is the center of the hand wheel frame end, is a unit tangent vector, is a unit radial vector; Indicates parallel; Indicates vertical.

[0068] Let be the normal angle, then:

[0069] .

[0070] Through the unit radial vector and the unit tangent vector under the set best fitting state, A, B, C, P four points and their vectors are analyzed, wherein, A point is the shoulder shaft point.

[0071] Specifically, the coordinates of the shoulder shaft point (i.e. A point) are obtained by Figure 6 It can be known that:

[0072] ;

[0073] Wherein, is the distance from the vertical center line of the moving chassis to the shoulder shaft point (i.e. A point), and H is the vertical distance from the shoulder shaft point (i.e. A point) to the bottom contact point of the omnidirectional passive wheel.

[0074] According to the constraint condition, the wheel bottom contact point P and the hand wheel frame end center C are expressed as:

[0075] ;

[0076] For the elbow axis point B, it can be analyzed by the shoulder axis point A and the hand wheel frame end center C, respectively:

[0077] ;

[0078] wherein, is the length of the arm mechanism, is the length of the elbow mechanism, is the angle between the moving chassis horizontal center line and the arm mechanism to be solved.

[0079] According to , the intermediate parameter , is the height of the permanent magnet adsorption wheel; the shoulder axis length equation can be obtained:

[0080] ;

[0081] That is, ;

[0082] wherein, the distance between the shoulder axis point A, the elbow axis point B and the hand wheel frame end center C and the center point O is respectively represented as A(R), B(R), C(R), and the corresponding expression is:

[0083] ;

[0084] The reachable criterion is obtained: ; wherein, is the radius value of all reachable steel structural members;

[0085] In the process of controlling the multi-degree-of-freedom robot to adaptively fit the steel structural member, the arm mechanism and the elbow mechanism are driven to rotate corresponding angles to ensure that the robot remains in the best fitting state; wherein the angle required for the arm mechanism to rotate and the angle required for the elbow mechanism to rotate The calculation formula is:

[0086] ;

[0087] ;

[0088] wherein, is the radius of the steel structural member; is the vertical distance from the shoulder axis point to the bottom contact point of the omnidirectional passive wheel; Distance from the vertical centerline of the mobile chassis to the shoulder axis point; Length of the elbow mechanism; Law phase angle; Intermediate parameter.

[0089] Step 2: According to the distance information of the steel structure and the robot pose information, the running state of the multi-degree-of-freedom robot is predicted by using the safety barrier tube model based on the adhesion margin, so as to control the motion safety and stability of the multi-degree-of-freedom robot in the inspection work in real time.

[0090] In the embodiment of the application, the robot pose information is estimated according to the rudder encoder constrained non-magnetic vision-inertial fusion method, so that the robot pose information is fully decoupled from the heading.

[0091] The magnetic adsorption wheel forms a strong and time-varying magnetic field near the robot, resulting in large measurement deviation of the geomagnetic meter and serious distortion of the direction jump, which cannot be used for heading observation. If only gyroscopes / accelerometers are used, the heading will drift over time. In the control module, in addition to the IMU (inertial measurement unit), the geometric constraints of the depth camera and the kinematics / touch constraints of the rudder magnetic encoder are introduced, the body pose is stably estimated without using the geomagnetic meter, and the robot pose is decoupled from the motion of the four claws.

[0092] Figure 7 It is a rudder encoder constrained non-magnetic vision-inertial fusion pose estimation flowchart of the embodiment of the application. The process of the rudder encoder constrained non-magnetic vision-inertial fusion pose estimation is as follows:

[0093] Firstly, the system state at time k in the sliding window is defined: ;

[0094] Wherein, is the rotation matrix from the world to the robot, is the velocity, is the position information, and are the gyro and accelerometer zero biases respectively.

[0095] Without using the geomagnetic, firstly, the time domain propagation is carried out, and the prior of at time k is obtained by the IMU (inertial measurement unit):

[0096] ;

[0097] Wherein, and are IMU measurements, is the gravity acceleration, is the Lie group exponential mapping, is the sampling interval.

[0098] Since the magnetic attraction wheel forms a strong and time-varying magnetic field near the robot, the inertial IMU alone will cause measurement drift, so the relative pose of adjacent frames provided by the depth camera is introduced to form a visual residual to constrain the prior:

[0099] ;

[0100] where, is the camera inter-relative pose derived from the current state, denotes the three-dimensional rigid body motion Lie group, that is, the complete 3D pose set, is the robot pose, is the camera-IMU extrinsic parameter, is the relative pose estimated by the depth camera, is the logarithmic mapping from the Lie group to the Lie algebra.

[0101] The kinematics / contact constraints are constructed using the magnetic encoder angle of each arm claw to decouple the body attitude and joint motion:

[0102] When the end is adsorbed, the position should remain unchanged in the world system:

[0103] ;

[0104] where, is the fixed contact point of the th adsorbed end effector in the world system, is the contact set, is obtained by the forward kinematics equation.

[0105] The zero-velocity contact constraint is used to enhance robustness:

[0106] ;

[0107] where, is the robot angular velocity, is the end Jacobian; adsorption .

[0108] The robot pose and heading are fitted from the depth map to the surface normal / axis, and are aligned with the end adsorption surface normal, so the normal alignment residual obtained by the cross product of the end adsorption surface normal rotated to the world system and the point cloud surface normal is:

[0109] ;

[0110] where, is the surface normal fitted from the point cloud, is the representation of the end adsorption surface normal in the machine system;​ Convert the vector into a cross product matrix for normal alignment.

[0111] By incorporating all the above information as factors into the same sliding window optimization, the objective function of the non-magnetic vision-inertial fusion method constrained by the servo encoder is obtained as follows:

[0112] ;

[0113] in, Describes the minimum value function; Represents the L2 norm; This is a set of internal states and external parameters within the sliding window; For IMU pre-integrated residuals; The kinematic consistency residual is formed by the forward kinematics of the encoder angle and the fixed contact point / zero-velocity contact of the adsorbed end; subscript The covariance / weight corresponding to the L2 norm; For visual residuals; The normal alignment residual is obtained by cross-product of the normal of the terminal adsorption surface with the normal of the point cloud surface after the normal is rotated to the world system. It is the kinematic consistency residual formed by the forward kinematics of the encoder angle and the "fixed contact point / zero velocity contact at the adsorbed end". The corresponding covariance / weights are set according to the sensor / model noise.

[0114] Figure 8 This is a flowchart of the predictive control of a tubular model for a safety barrier based on adhesion margin, according to an embodiment of the present invention. In this embodiment, to achieve unified optimization of the robot's "radial distance—attitude—turning," the system state and control variables are first vectorized and modeled; this modeling facilitates subsequent real-time solution using quadratic programming.

[0115] ;

[0116] in, for The state vector at any given time; The radial error between the robot and the center of the steel pipe; This represents the rate of change of radial error. It is the lateral tilt angle; The tilt angle and angular velocity; The heading angle is along the tangent of the pipe; It is the set of rotation angles of all steering wheels; For control vectors; The tangential velocity is the combined velocity along the pipe. This represents the angle increment for each steering wheel; This sets the amount for adhesion / normal force. The subscript... k express kTime instant; denotes the transpose of a vector.

[0117] The robot dynamics and geometric constraints are linearized at the task point and sampled at a period discretized as follows:

[0118] ;

[0119] where, is the state at the next sampling instant; is the system state matrix; is the input matrix; is the disturbance channel matrix; is the total equivalent disturbance.

[0120] To combat uncertainty, the control law is decomposed into a nominal control and a feedback correction, and the nominal trajectory is predicted as:

[0121] ;

[0122] where, and are the nominal state and nominal control, respectively; is the stabilizing state feedback gain; is the deviation state between actual and nominal; where the subscript k denotes k the time instant; denotes the time instant.

[0123] In this embodiment, radial error and attitude stabilization are the primary objectives, and control smoothness and economy terms are introduced; meanwhile, an incentive is given to the adhesion safety margin, resulting in a safety barrier tube model based on the adhesion margin as:

[0124] is the nominal trajectory obtained by rolling prediction at the end of the prediction domain; the current estimated state is recursively obtained from the discrete linearized model using the nominal control sequence to obtain the entire nominal trajectory:

[0125] ;

[0126] where, denotes the minimum function; denotes the two-norm; the subscript is the weight matrix of the corresponding two-norm; is the control increment at the time instant; is the desired tangential velocity reference; is the adhesion margin reward coefficient; for Adhesion safety margin at all times; For reference only; The radial error between the robot and the center of the steel pipe; It is the lateral tilt angle; for The tangential velocity along the tube at any given moment; This is the initial control variable; The tube center state at the end of the prediction domain for the nominal trajectory obtained by rolling prediction; The range of values ​​is ; It is a positive integer greater than or equal to 1.

[0127] In this embodiment, the adhesion safety margin is defined as the difference between the available anti-overturning moment and the required anti-overturning moment, used to measure the risk of overturning. Adhesion safety margin at all times The expression is:

[0128] ;

[0129] in, The anti-overturning moment is provided by the adhesive device; This refers to the required torque caused by gravity, centrifugal effect, and contact shear. They represent The state vector and control vector at each moment; for The amount of adhesion / normal force set at any given moment.

[0130] To facilitate real-time optimization, the relationship between the adhesive torque and the set value is locally approximated using a linear approach:

[0131]

[0132] in, Baseline adhesion torque; The gain coefficient of the adhesion setting on the torque.

[0133] The first-order sensitivity of the demand torque to the state and control variables is expressed using a linear representation:

[0134] ;

[0135] in, For constant terms; This is the state sensitivity vector; for Transpose of; To control the sensitivity vector; for transpose of

[0136] In this embodiment, the "radius band, inclination upper bound, adhesion margin lower bound" are constructed as control barrier functions, to guarantee the forward invariance of the safe set:

[0137] ;

[0138] where, is the barrier function; is the maximum allowed radial deviation; is the maximum allowed inclination; is the adhesion margin lower bound.

[0139] The control barrier function is incorporated into a quadratic program, and a discrete first-order approximation is used to construct linear inequalities within the sampling interval:

[0140] ;

[0141] where, is the barrier function gradient; is the class function coefficient, used to guarantee the forward invariance; the terms in brackets are the discrete approximation of one-step state change.

[0142] In this embodiment, the Coulomb friction cone is approximated in a polyhedral manner, restricting the tangential contact force not to exceed the available friction force:

[0143] ;

[0144] where, is the norm of the tangential contact force; is the friction coefficient; is the normal force; the right-hand side is its equivalent linear inequality form, is the coefficient matrix after approximation and the upper bound vector.

[0145] For tubular robust control and barrier constraints, this embodiment constructs a disturbance observer and performs online identification of key parameters:

[0146] ;

[0147] where, is the disturbance estimate; is the observer gain matrix; the terms in brackets are the residual between model prediction and measurement. Preferably, the friction coefficient and the normal force model are estimated online using recursive least squares or Kalman filtering, and their conservative lower bounds are taken in the optimization.

[0148] When the ranging increases and there is a tendency to tilt, the resultant velocity pointing to the center of the steel pipe and the steering adjustment are automatically generated to stabilize the radial error and make it converge; through the safety barrier function constraint, the rigid satisfaction of the radius, the maximum inclination angle and the adhesion margin is realized to avoid pipe departure and overturning; based on the tubular robust structure and online parameter identification, the influence of the friction, the curvature and the load uncertainty on the control performance is effectively inhibited.

[0149] The above only is the preferred embodiment of the present application, and is not used to limit the present application, for the person skilled in the art, the present application can have various changes and changes. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A control method of a multi-degree-of-freedom robot for steel structure high-altitude roof climbing inspection, the multi-degree-of-freedom robot for steel structure high-altitude roof climbing inspection comprising: At least two mobile chassis, each two adjacent mobile chassis are connected by multi-degree-of-freedom passive spherical hinge; Each mobile chassis comprises at least a control module, a motion assembly and a perception assembly; The motion assembly and the perception assembly are respectively in communication connection with the control module; the perception assembly is used for perceiving current steel structure diameter, distance information of the steel structure and robot pose information and transmitting to the control module; the control module is used for adapting to the steel structure according to the current steel structure diameter and the multi-degree-of-freedom of the motion assembly, and controlling the motion assembly to perform the climbing inspection task according to the distance information of the steel structure and the robot pose information, The control method of the multi-degree-of-freedom robot for the steel structure high-altitude roof climbing inspection comprises the following steps: Controlling the multi-degree-of-freedom robot to adapt to the steel structure according to the current steel structure diameter and the multi-degree-of-freedom of the robot arm claw; According to the distance information of the steel structure and the robot pose information, the running state of the multi-degree-of-freedom robot is predicted and controlled by using a safety barrier tubular model based on adhesion margin, so as to control the motion safety and stability of the multi-degree-of-freedom robot in the inspection work in real time; The safety barrier tubular model based on adhesion margin is as follows: ； wherein, denotes a minimum function; denotes a two-norm; subscript is a weight matrix of the corresponding two-norm; is a control increment at time is a desired tangential velocity reference; is a stickiness margin reward coefficient; is a stickiness safety margin at time is a reference state; is a radial error of the robot from the center of the pipe; is a lateral tilt angle; is a tangential resultant velocity along the pipe at time is an initial control; is a pipe center state at the end of the prediction horizon of the nominal trajectory resulting from the rolling prediction; the value of ; is a positive integer greater than or equal to 1; In the process of controlling the multi-degree-of-freedom robot to adaptively fit the steel structure, the arm mechanism and the elbow mechanism are rotated by corresponding angles to ensure that the robot is kept in the best fitting state; wherein the calculation formula of the angle required for the rotation of the arm mechanism and the angle required for the rotation of the elbow mechanism are as follows: ； ； ； wherein, R is the radius of the steel structure; H is the vertical distance from the shoulder axis point to the bottom contact point of the omni-directional passive wheel; L is the distance from the vertical centerline of the moving chassis to the shoulder axis point; L is the length of the elbow mechanism; is the normal phase angle; is the intermediate parameter; is the height of the permanent magnetic adsorption wheel.

2. The control method of the multi-degree-of-freedom robot for climbing and inspecting a steel structure high-rise roof according to claim 1, wherein The motion assembly comprises at least two multi-degree-of-freedom mechanical arm claws, which are symmetrically arranged on the two sides of the mobile chassis; the end of each multi-degree-of-freedom mechanical arm claw is connected with a permanent magnetic adsorption wheel through one side of a hand wheel frame, and the other side of the hand wheel frame is sequentially connected with an elbow mechanism, an arm mechanism and a shoulder mechanism, and the shoulder mechanism is installed on the mobile chassis.

3. The control method of the multi-degree-of-freedom robot for climbing and inspecting a steel structure high-rise roof according to claim 2, wherein The shoulder mechanism, the arm mechanism, the elbow mechanism and the hand wheel frame all have independent degrees of freedom, and together constitute four degrees of freedom of the mechanical arm claw.

4. The control method of the multi-degree-of-freedom robot for climbing and inspecting a steel structure high-rise roof according to claim 3, wherein One end of the shoulder mechanism is connected with the mobile chassis, and the other end is connected with the arm mechanism, and the shoulder mechanism can realize the rotation degree of freedom around the lateral direction of the mobile chassis; One end of the arm mechanism away from the shoulder mechanism is connected with the elbow mechanism, and the arm mechanism can realize the extension and contraction degree of freedom along the axial direction of the arm mechanism; The elbow mechanism is connected with the hand wheel frame, the elbow mechanism can realize the bending degree of freedom perpendicular to the direction of the arm mechanism, and the hand wheel frame can realize the rotation degree of freedom relative to the elbow mechanism.

5. The control method of the multi-degree-of-freedom robot for climbing and inspecting a steel structure high-rise roof according to claim 2, wherein The bottom of the mobile chassis is further provided with a passive omni-directional wheel, and the passive omni-directional wheel and the permanent magnetic adsorption wheels on the two sides thereof constitute a three-point support structure.

6. The control method of the multi-degree-of-freedom robot for climbing and inspecting a steel structure high-rise roof according to claim 2, wherein The permanent magnetic adsorption wheel comprises a wheel body shell, a permanent magnet group, a wheel shaft and a wear-resistant wheel surface; the wheel shaft penetrates through the center of the wheel body shell and is rotatably connected with the hand wheel frame at both ends; the wear-resistant wheel surface is wrapped outside the wheel body shell, and the surface is provided with anti-skid lines; the permanent magnet group is used for generating a continuous magnetic field, so that the adsorption wheel is stably adsorbed on the surface of the steel structure.

7. The control method of the multi-degree-of-freedom robot for climbing and inspecting a high-rise roof of a steel structure according to claim 1, wherein The robot pose information is estimated by using a rudder encoder constrained non-magnetic vision-inertial fusion method, so that the robot pose information is fully decoupled from the heading; the objective function of the rudder encoder constrained non-magnetic vision-inertial fusion method is as follows: ； where, denotes the minimum function; denotes the two-norm; is the set of states and extrinsics within the sliding window; is the IMU pre-integration residual; is the kinematic consistency residual consisting of forward kinematics from encoder angles and the absorbed end-fixed contact / zero-velocity contact; subscript is the covariance / weight corresponding to the two-norm; is the visual residual; is the normal alignment residual from the cross product of the normal of the end effector suction surface rotated to the world frame and the surface normal of the point cloud.

Images