Adaptive control method and system for inspection robot based on multi-mode motion switching

Abstract

The invention relates to the technical field of inspection robot control, in particular to an inspection robot self-adaptive control method and system based on multi-mode motion switching, and the method comprises the steps: collecting environment and motion state information, judging a current applicable motion mode, and triggering mode switching when a switching criterion and a minimum residence time condition are met; generating a reference trajectory in the mode, and establishing a linear parameterized dynamic regression equation for online adaptive estimation; constructing a tracking error and a composite error, calculating a control command by adopting a control law, carrying out online estimation on a parameter vector in the dynamic regression equation in combination with a recursive least square (RLS) and a gradient adaptive law, and applying projection and leakage terms to an estimated value so as to enable the estimated value to be bounded; during mode switching, parameter memory is used to migrate an estimated parameter of a previous mode and a parameter of a new mode, and interpolation is carried out between a previous control law and a new control law with a smooth weighting coefficient in a switching duration.

Description

(19) State Intellectual Property Office (12) Invention Patent Application (10) Application Publication Number (43) Application Publication Date (21) Application Number 202511286461.4 (22) Application Date 2025.09.10 (71) Applicant Super Vision Technology Co., Ltd. Address 075000 No. 28, Zhanqian East Street, Qiaodong District, Zhangjiakou City, Hebei Province, 4th Floor, Building 10, Hebei Guokong Northern Silicon Valley High-tech New City (72) Inventors Yan Hao, Wang Yongfei, Zhang Yuang (51) Int.Cl. G05B 13 / 04 (2006.01) (54) Invention Title: Adaptive Control Method and System for Inspection Robot Based on Multi-Mode Motion Switching (57) Abstract: This invention relates to the field of inspection robot control technology, and particularly to an adaptive control method and system for inspection robot based on multi-mode motion switching. The method includes: collecting environmental and motion state information, determining the currently applicable motion mode, and triggering mode switching when the switching criterion and minimum dwell time condition are met; generating a reference trajectory in the mode, establishing a linear parameterized dynamic regression equation for online adaptive estimation; constructing tracking error and composite error, using a control law to calculate control commands, combining recursive least squares (RLS) and gradient adaptive law to perform online estimation of the parameter vector in the dynamic regression equation, and applying projection and leakage terms to the estimated values ​​to make the estimated values ​​bounded; during mode switching, using parameter memory to transfer the estimated parameters of the previous mode to the parameters of the new mode, and interpolating between the previous control law and the new control law with a smooth weighting coefficient during the switching duration. Claims 2 pages, Description 9 pages, Drawings 2 pages, CN 121115490 A 2025.12.12 CN 1 21 11 54 90 A 1. An adaptive control method for an inspection robot based on multi-mode motion switching, characterized by the following steps: Collecting environmental and motion state information, obtaining the robot's pose and velocity through state estimation; determining the currently applicable motion mode based on the environmental and motion state information, triggering mode switching when the switching criterion and minimum dwell time condition are met; generating a reference trajectory under the current motion mode through motion planning based on the current motion mode, establishing a linear parameterized dynamic regression equation for the current mode for online adaptive estimation; constructing tracking error and composite error, calculating control commands using a control law, combining recursive least squares (RLS) and gradient adaptive law to perform online estimation of the parameter vector in the dynamic regression equation, and applying projection and leakage terms to the estimated values ​​to make the estimated values ​​bounded; During mode switching, parameter memory is used to transfer the estimated parameters of the previous mode to the parameters of the new mode, and during the switching duration, a smooth weighting coefficient is used to interpolate between the previous control law and the new control law. 2. The adaptive control method for inspection robots based on multi-mode motion switching according to claim 1, characterized in that...The switching criteria include logical judgments based on terrain features, robot speed, contact force information, and minimum dwell time, and the time interval must be greater than 0 before the switching is triggered. 3. The adaptive control method for an inspection robot based on multi-mode motion switching according to claim 1, characterized in that the online estimation of the parameter vector terms in the dynamic regression equation by combining recursive least squares (RLS) and gradient adaptive law includes: recursively updating the linear regression composed of the regression vector and the measurement quantity in the dynamic regression equation using recursive least squares (RLS), and RLS includes a forgetting factor; slowly adjusting the parameter vector using gradient adaptive law, and fusing the result of RLS in a weighted or mixed manner as the final parameter estimate. 4. The adaptive control method for an inspection robot based on multi-mode motion switching according to claim 3, characterized in that the application of projection and leakage terms to the estimate to make the estimate bounded includes: applying a projection operator to the estimate to restrict it to a pre-defined bounded set, and adding a leakage term during the projection process. 5. The adaptive control method for an inspection robot based on multi-mode motion switching according to claim 3, characterized in that the RLS operates at a high rate only when the regression vector satisfies the persistent excitation condition; otherwise, it operates at a reduced frequency or only with the gradient adaptive law. 6. The adaptive control method for an inspection robot based on multi-mode motion switching according to claim 1, characterized in that, in the interpolation between the previous control law and the new control law using a smooth weighting coefficient during the switching duration, the smooth weighting coefficient is 1 at the start of the switching and continuously and monotonically decreases to 0 during the transition duration T, and the transition duration T is configurable. 7. An adaptive control system for an inspection robot based on multi-mode motion switching, characterized in that it uses the method described in any one of claims 1 to 6, the system comprising: a mode judgment unit, used to collect environmental and motion state information, obtain robot pose and velocity through state estimation; determine the currently applicable motion mode based on the environmental and motion state information, and trigger mode switching when the switching criterion and the minimum dwell time condition are met; a regression equation construction unit, used to generate a reference trajectory in the current motion mode based on motion planning, and establish a linear parameterized dynamic regression equation for the current mode for online adaptive estimation; a parameter estimation unit, used to construct tracking error and composite error, calculate control commands using control laws, and perform online estimation of the parameter vector in the dynamic regression equation by combining recursive least squares (RLS) and gradient adaptive laws, and apply projection and leakage terms to the estimated values ​​to make the estimated values ​​bounded; and a mode switching unit, used to use parameter memory to combine the estimated parameters of the previous mode with the parameters of the new mode during mode switching.The numbers are transferred and interpolated between the previous control law and the new control law with a smooth weighting coefficient during the switching duration. 8. A computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that when the processor executes the computer program, it implements the method as described in any one of claims 1-6. 9. A storage medium storing a computer program thereon, characterized in that when the computer program is executed by a processor, it implements the method as described in any one of claims 1-6. Claims 2 / 2 Page 3 CN 121115490 A Adaptive Control Method and System for Inspection Robots Based on Multi-Mode Motion Switching Technical Field

[0001] This invention relates to the field of inspection robot control technology, and in particular to an adaptive control method and system for inspection robots based on multi-mode motion switching. Background Art

[0002] Inspection robots are widely used for inspection operations in complex environments such as power, petrochemical, and rail. Existing technologies mostly adopt a single motion mode or a simple switching strategy based on preset rules, and the controller usually relies on precise modeling or fixed-gain PID / robust control. When inspection robots encounter sudden changes in terrain (such as steps, obstacles, narrow passages), friction, or load changes, the tracking error increases significantly, and parameter uncertainty and external interference are difficult to compensate for in a timely manner. In addition, traditional switching is often instantaneous, which can easily cause control shock, actuator stress peaks, and task failure.

[0003] Existing online identification methods converge slowly when computing power is limited, and lack a unified mode transfer, transition smoothing, and safety monitoring mechanism, making it difficult to meet the real-time and robustness requirements on site.

[0004] The information disclosed in this background section is only intended to deepen the understanding of the overall background technology of the present invention, and should not be regarded as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] The present invention provides an adaptive control method and system for inspection robots based on multi-mode motion switching, thereby effectively solving the problems in the background technology.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is: an adaptive control method for an inspection robot based on multi-mode motion switching, comprising the following steps:

[0007] Collecting environmental and motion state information, and obtaining the robot's pose and velocity through state estimation; determining the currently applicable motion mode based on the environmental and motion state information, and triggering mode switching when the switching criterion and the minimum dwell time condition are met;

[0008] Generating a reference trajectory in the current motion mode by motion planning based on the current motion mode, and establishing a linear parameterized dynamic regression equation for the current mode for online adaptive estimation;

[0009] Constructing tracking error and composite error, calculating control commands using control laws, and combining recursive least squares (RLS) with...The gradient adaptive law estimates the parameter vector in the dynamic regression equation online, and applies projection and leakage terms to the estimated value to make the estimated value bounded;

[0010] During mode switching, parameter memory is used to transfer the estimated parameters of the previous mode to the parameters of the new mode, and interpolation is performed between the previous control law and the new control law with smooth weighting coefficients during the switching duration.

[0011] Further, the switching criterion includes logical judgment based on terrain features, robot speed, contact force information, and minimum dwell time, and the time interval must be greater than 0 before the switching is triggered.

[0012] Further, the online estimation of the parameter vector terms in the dynamic regression equation by combining recursive least squares (RLS) and gradient adaptive law includes:

[0013] Recursively updating the linear regression consisting of the regression vector and the measurement quantity in the dynamic regression equation using recursive least squares (RLS), and RLS includes a forgetting factor; Specification 1 / 9 page 4 CN 121115490 A

[0014] Slowly adjusting the parameter vector using gradient adaptive law, and fusing the result of RLS in a weighted or mixed manner as the final parameter estimate.

[0015] Further, the application of projection and leakage terms to the estimate to make the estimate bounded includes:

[0016] Applying a projection operator to the estimate to restrict it to a pre-defined bounded set, and adding a leakage term during the projection process.

[0017] Further, the RLS operates at a high rate only when the regression vector satisfies the persistent excitation condition; otherwise, it operates at a lower frequency or only with the gradient adaptive law.

[0018] Further, in the interpolation between the previous control law and the new control law with a smooth weighted coefficient during the switching duration, the smooth weighted coefficient is 1 at the start of the switching and continuously and monotonically decreases to 0 during the transition duration T, and the transition duration T is configurable.

[0019] The present invention also includes an adaptive control system for an inspection robot based on multi-mode motion switching, using the method described above. The system includes:

[0020] a mode judgment unit, used to collect environmental and motion state information, obtain the robot's pose and velocity through state estimation; determine the currently applicable motion mode based on the environmental and motion state information, and trigger mode switching when the switching criterion and the minimum dwell time condition are met;

[0021] a regression equation construction unit, used to generate a reference trajectory in the current motion mode from motion planning, and establish a linear parameterized dynamic regression equation for the current mode for online adaptive estimation;

[0022] a parameter estimation unit, used to construct tracking error and composite error, calculate control commands using control laws, and perform online estimation of the parameter vector in the dynamic regression equation by combining recursive least squares (RLS) and gradient adaptive law, and perform online estimation of...The estimated value is projected and a leakage term is applied to make the estimated value bounded;

[0023] A mode switching unit is used to migrate the estimated parameters of the previous mode to the parameters of the new mode using parameter memory during mode switching, and to interpolate between the previous control law and the new control law with a smooth weighting coefficient during the switching duration.

[0024] The present invention also includes a computer device, including a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the method as described above.

[0025] The present invention also includes a storage medium storing a computer program thereon, wherein when the computer program is executed by a processor, it implements the method as described above.

[0026] The beneficial effects of the present invention are as follows: through linear parameterization and online parameter estimation, the online parameters, combined with recursive least squares (RLS) and gradient adaptive law, have a strong compensation capability for unknown mass, friction, and terrain disturbances, ensuring that a small tracking error can still be maintained when the load changes or friction changes abruptly; by adopting a smooth transition strategy of minimum dwell time, parameter memory, and control law interpolation, the impact and oscillation caused by switching transients can be reduced, the peak force on the actuator and the detected object can be reduced, the hardware life can be extended, and the task success rate can be improved. Brief Description of the Drawings

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 is a flowchart of the method of the present invention; Specification 2 / 9 pages 5 CN 121115490 A

[0029] Figure 2 is a structural schematic diagram of the system of the present invention;

[0030] Figure 3 is a structural schematic diagram of the computer device of the present invention. Detailed Description

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0032] Embodiment 1:

[0033] As shown in Figure 1: An adaptive control method for an inspection robot based on multi-mode motion switching includes the following steps:

[0034] Collect environmental and motion state information, and obtain the robot pose and speed through state estimation; determine the currently applicable motion mode based on the environmental and motion state information, and trigger mode switching when the switching criterion and the minimum dwell time condition are met;

[0035] Generate a reference trajectory in the current motion mode by motion planning based on the current motion mode, and establish a linear parameterized dynamic regression equation for the current mode for online adaptive estimation;

[0036] The tracking error and composite error are constructed, and the control command is calculated using the control law. The parameter vector in the dynamic regression equation is estimated online by combining the recursive least squares (RLS) method and the gradient adaptive law. The estimated value is then projected and a leakage term is applied to make the estimated value bounded.

[0037] When switching modes, parameter memory is used to transfer the estimated parameters of the previous mode to the parameters of the new mode. During the switching duration, a smooth weighted coefficient is used to interpolate between the previous control law and the new control law.

[0038] Through linear parameterization and online parameter estimation, the online parameters, combined with the recursive least squares (RLS) method and the gradient adaptive law, have a strong compensation ability for unknown mass, friction and terrain disturbances, ensuring that a small tracking error can still be maintained when the load changes or friction changes suddenly. The smooth transition strategy of minimum dwell time, parameter memory and control law interpolation can reduce the impact and oscillation caused by switching transients, reduce the peak force on the actuator and the detected object, extend the hardware life and improve the mission success rate.

[0039] In this embodiment, the switching criterion includes logical judgment based on terrain features, robot speed, contact force information, and minimum dwell time, and the time interval must be greater than 0 before the switching is triggered.

[0040] The parameter vector term in the dynamic regression equation is estimated online by combining recursive least squares (RLS) and gradient adaptive law, including:

[0041] Recursively updating the linear regression consisting of the regression vector and the measurement quantity in the dynamic regression equation using recursive least squares (RLS), and RLS includes a forgetting factor;

[0042] Slowly adjusting the parameter vector using gradient adaptive law, and fusing the result of RLS in a weighted or mixed manner as the final parameter estimate.

[0043] Wherein, applying projection and leakage terms to the estimate to make the estimate bounded includes:

[0044] Applying a projection operator to the estimate to restrict it to a pre-defined bounded set, and adding a leakage term during the projection process.

[0045] In this embodiment, RLS only runs at a high rate when the regression vector satisfies the persistent excitation condition; otherwise, it runs at a reduced frequency or only runs with the gradient adaptive law.

[0046] During the switching duration, a smooth weighting coefficient is used to interpolate between the previous control law and the new control law. The smooth weighting coefficient is 1 at the beginning of the switching and continuously and monotonically decreases to 0 during the transition duration T. The transition duration T can be configured as per the specification page 3 / 9, 6 CN 121115490 A.

[0047] As shown in Figure 2, this embodiment also includes an adaptive control system for an inspection robot based on multi-mode motion switching. Using the method described above, the system includes:

[0048] A mode judgment unit, used to collect environmental and motion state information, obtain the robot pose and speed through state estimation; determine the currently applicable motion mode based on the environmental and motion state information, and when the switching criterion is met and the minimum dwell time is satisfied.The time condition triggers mode switching;

[0049] The regression equation construction unit is used to generate the reference trajectory in the current motion mode by motion planning based on the current motion mode, and establish a linear parameterized dynamic regression equation for the current mode for online adaptive estimation;

[0050] The parameter estimation unit is used to construct the tracking error and composite error, calculate the control command using the control law, combine the recursive least squares (RLS) method and the gradient adaptive law to perform online estimation of the parameter vector in the dynamic regression equation, and apply projection and leakage terms to the estimated value to make the estimated value bounded;

[0051] The mode switching unit is used to migrate the estimated parameters of the previous mode and the parameters of the new mode using parameter memory during mode switching, and interpolate between the previous control law and the new control law with a smooth weighting coefficient during the switching duration.

[0052] Example 2:

[0053] The implementation method, steps, main parameter settings, implementation details and the role and beneficial effects of the present invention are described below with specific embodiments, but the present invention is not limited to the following embodiments.

[0054] I. System Hardware and Software Composition;

[0055] Hardware Composition:

[0056] 1. Inspection robot platform (differential chassis or wheel-leg hybrid chassis);

[0057] 2. Sensing units: inertial measurement unit (IMU), wheel / joint encoder, lidar or depth camera (for terrain / obstacle detection), force / contact sensor (for contact / stair climbing detection), camera (optional);

[0058] 3. Execution unit: wheel motor driver, multi-degree-of-freedom servo motor or joint actuator;

[0059] 4. Computing unit: embedded controller (e.g., ARM or x86-based industrial main controller), running a real-time operating system (e.g., real-time Linux / RTOS);

[0060] 5. Communication and storage: CAN / RS-485 / Ethernet bus, non-volatile memory for storing regression templates, parameter boundaries and initial values.

[0061] Software Modules:

[0062] 1. Perception Module (point cloud / depth map processing, step and slope detection, obstacle distance detection);

[0063] 2. State Estimation Module (fusion of IMU and encoder, output pose and velocity);

[0064] 3. Supervisor: implements pattern recognition, switching criteria, minimum dwell time logic, and parameter migration strategy;

[0065] 4. Trajectory Generator for Each Mode: generates reference trajectories for different modes (flat ground, obstacle crossing, steps, narrow passage);

[0066] 5. Adaptive Control Module: includes regression modeling unit, RLS sub-unit, gradient estimation sub-unit, projection and hybrid fusion unit, and control law calculation unit;

[0067] 6. Smooth Transition Unit: executes control law / parameter weight interpolation;

[0068] 7. Safety Monitoring Unit: slippage, impact, and overload monitoring and emergency logic.

[0069] II. Dynamic Modeling and Regression;

[0070] 1. Simplified Dynamic Model; Manual 4 / 9 Page 7 CN 121115490 A

[0071] Taking a two-dimensional differential chassis (center of mass coordinates (x, y) and heading angle ψ) as an example, the control input is the left and right wheel drive torques T_L, T_R. Its center of mass dynamics can be approximately expressed in Cartesian form as:

[0072]

[0073] Where: v is the forward velocity, ω is the angular velocity; uv, uω are the equivalent force / torque obtained by mapping the wheel drive input; m, Iz are the equivalent mass and moment of inertia; b, bω are the viscous friction or equivalent damping; dv, dω are unmodeled disturbances (e.g., uneven terrain, slope component, etc.).

[0074] The unknown quantities are linearly parameterized to obtain the regression form:

[0075]

[0076] Example of regression vector: The parameter vector is θv=[m, b]T.

[0077] For multi-joint / multi-degree-of-freedom platforms, the regression matrix Y_m can be obtained by standard robot dynamics linearization tools (such as

[0078] Newton-Euler regression form).

[0079] III. Specific implementation of control law and parameter estimation algorithm;

[0080] 1. Definition of reference trajectory and error;

[0081] In mode m, set the reference trajectory qd(t), and define the error:

[0082] e = q - qd, Λ = diag(λ1,...,λn) > 0;

[0083] Recommended example parameters (engineering values): sampling period Ts = 0.01s, I = diag(10, 10, 10) (adjusted according to degree of freedom).

[0084] 2. Discrete control law;

[0085] Calculate at sampling time k:

[0086]

[0087] Where the damping gain matrix K is a diagonal positive definite matrix, such as K = diag(20, 20, 20). During execution, T is saturated to satisfy the actuator constraints.

[0088] 3. RLS Discrete Update (for fast identification);

[0089] Let the regression vector φk and the measurement quantity yk (e.g., the target quantity obtained by mapping the left end of the dynamics to the control input) satisfy a linear relationship

[0090] RLS update formula:

[0091]

[0092] Example: Initial covariance P0 = 103I, forgetting factor λ = 0.995. When it is less than the preset threshold (indicating insufficient excitation), the RLS high-speed update is paused.

[0093] 4. Gradient-type adaptive law (slow);

[0094] Discrete gradient law (Euler approximation): Specification 5 / 9 page 8 CN 121115490 A

[0095]

[0096] Where the learning rate Γ is a diagonal positive definite matrix (example is diag(0.1.)), and the leakage term σ = 1×10-3.

[0097] 5. Projection and fusion strategy;

[0098] The two estimation results are fused using a weighted method:

[0099]

[0100] The weight η∈[0,1] can be dynamically adjusted according to the RLS excitation (η→1 when the excitation is strong, η→0 when the excitation is weak). The projection operator is used to ensure that it is within the boundary set Ω (e.g., Ω=[θmin,θmax]).

[0101] 6. Key points of Lyapunov stability explanation;

[0102] Selecting the Lyapunov function:

[0103]

[0104] Under the idealized assumptions (parameter linearization holds, the estimation law uses projection and the perturbation is bounded), it can be proved that, given the perturbation bound and the selection of a suitable gain, both the tracking error and the parameter estimation are bounded and tend to a small domain.

[0105] IV. Specific Implementation of Multi-Mode Recognition and Switching Strategy;

[0106] 1. Define the mode set;

[0107]

[0108] 2. Switching Criteria (Example Logic);

[0109] Step mode trigger: When the depth camera or point cloud detects a continuous height difference h ∈ [hmin, hmax] (e.g., 0.03m ~ 0.2m), and the approach distance is less than the threshold, and the current speed is within the climbable step speed range;

[0110] Obstacle crossing mode trigger: The width and height of the detected obstacle are within the traversable range; Narrow obstacle avoidance trigger: The channel width is less than the threshold and precise attitude control is required; Switching must meet the minimum dwell time (Dwell-time) condition: Tdwell = 0.5s (example value) to prevent jittery switching.

[0111] 3. Smooth transition;

[0112] When the switch is triggered, the Supervisor starts the transition timer and smoothly transitions the control law from the old mode to the new mode within the transition period Ttrans (0.2 to 0.5 s in the example) according to the interpolation coefficient α(t) = 1-t / Ttrans. At the same time, the old mode parameter estimate is transferred to the new mode initial estimate with weights:

[0113]

[0114] In the example, β = 0.8 is selected (higher if the scene is similar).

[0115] V. Safety Monitoring and Emergency Strategies;

[0116] 1. Real-time monitoring of wheel-to-ground adhesion indicators (such as slip ratio based on wheel speed and vehicle acceleration). When the slip ratio exceeds the threshold (example 0.15), immediately enter safety mode, reduce the target speed, and replan the trajectory;

[0117] 2. If the contact force exceeds the threshold (such as abnormally large contact force detected when climbing stairs), trigger retreat and backtrack to a safe position;

[0118] 3. In the event of sensor malfunction or estimation divergence, Supervisor can switch to conservative mode (low speed, large resistance) and report the fault log.

[0119] VI. Numerical Examples and Simulation Verification;

[0120] 1. Simulation Platform;

[0121] Use ROS+Gazebo or PyBullet as the simulation environment, or establish a multi-mode simulation in MATLAB / Simulink.The dynamic model was tested in a closed loop.

[0122] 2. Verification scenarios;

[0123] Scenario A (flat ground): speed tracking under different friction coefficients (0.5, 0.8, 1.0); Scenario B (steps): detection and switching to step mode to verify attitude and force peak value during transition; Scenario C (load change): sudden addition of 5kg load to verify parameter estimation convergence; Scenario D (obstacle avoidance): passing through a narrow passage with low error.

[0124] 3. Expected indicators;

[0125] The tracking RMS error is significantly reduced (compared to the unadapted PID control, the expected RMS error is reduced by 30% to 60%);

[0126] The switching transient peak value is controlled (the acceleration peak value decreases by 40% during the transition period);

[0127] The parameter estimation converges to within 5% of the true value under excitation conditions.

[0128] The role and beneficial effects of this implementation scheme;

[0129] 1. Enhanced robustness: Through linear parameterization and online parameter estimation (RLS + gradient), it has a strong compensation capability for unknown mass, friction and terrain disturbances, ensuring that it can still maintain a small tracking error when the load changes or friction changes abruptly.

[0130] 2. Smooth mode switching: By adopting a smooth transition strategy of minimum dwell time, parameter memory and control law interpolation, it can reduce the impact and oscillation caused by switching transients, reduce the peak force on the actuator and the detected object, extend the hardware life and improve the task success rate.

[0131] 3. Real-time performance and engineering feasibility: By setting RLS as conditional trigger and combining it with low-overhead gradient estimation, it takes into account the estimation convergence speed and computational resource constraints, and is suitable for typical embedded controller platforms.

[0132] 4. Safety: By integrating slip, contact force and control input detection, it can quickly trigger emergency strategies and protect the robot and surrounding facilities in dangerous situations.

[0133] 5. Universality and scalability: This method is based on a modular design and is applicable to differential chassis, wheel-leg hybrid platforms, and inspection vehicles with robotic arms. The regression template Y_m, parameter set m, and initial parameters are configurable and can be updated remotely, facilitating on-site calibration and model expansion.

[0134] 6. Applicable engineering parameters: By providing a series of engineering parameters (sampling period, dwell time, RLS forgetting factor, gain magnitude, etc.), it is easy to quickly carry out simulation verification and on-site debugging.

[0135] Please refer to Figure 3, which shows a schematic diagram of the structure of the computer device provided in the embodiment of this application. The computer device 400 provided in the embodiment of this application includes: a processor 410 and a memory 420. The memory 420 stores a computer program executable by the processor 410. When the computer program is executed by the processor 410, it performs the above-mentioned method.

[0136] The embodiment of this application also provides a storage medium 430, on which a computer program is stored. When the computer program is run by the processor 410, it performs the above-mentioned method.

[0137] The storage medium 430 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. Specification page 7 / 9 10 CN 121115490 A

[0138] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. The meaning of "multiple" is two or more, unless otherwise explicitly specified.

[0139] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal connection of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in this invention can be understood according to the specific circumstances.

[0140] In the description of this specification, the description with reference to the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., means that the specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any suitable manner in one or more embodiments or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described herein, as well as the features of the different embodiments or examples.

[0141] Any process or method described in the flowcharts or otherwise herein can be understood to represent includingA module, segment, or portion of code containing one or more executable instructions for implementing a particular logical function or process, and the scope of preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order according to the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0142] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, a “computer-readable medium” can be any means that can contain, store, communicate, propagate, or transmit a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples of computer-readable media (a non-exhaustive list) include the following: electrical connections (electronic devices) having one or more wires, portable computer disks (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Additionally, computer-readable media can even be paper or other suitable media on which the program is printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0143] It should be understood that various parts of the invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logic functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0144] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware, and the program can be stored in a computer-readable storage medium.In this process, the program, when executed, includes one or a combination of the steps of the method embodiments.

[0145] The storage medium mentioned above may be a read-only memory, a disk, or an optical disk, etc. Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention. Instruction sheet 9 / 9 page 12 CN 121115490 A Figure 1 Figure 2 Instruction drawing 1 / 2 page 13 CN 121115490 A Figure 3 Instruction sheet drawing 2 / 2 page 14 CN 121115490 A Abstract Abdominal ultrasound examination method, system and device Abstract -------------------------------------------------------------------------------------------------------------- Adaptive Control Method and System for Inspection Robot Based on Multi-Mode Motion Switching The present invention relates to the technical field of inspection robot control, and in particular, to an adaptive control method and system for an inspection robot based on multi-mode motion switching. The method includes: collecting environment and motion state information, determining a currently applicable motion mode, and triggering a mode switch when a switching criterion is satisfied and a minimum dwell time condition ismet; generating a reference trajectory for said mode, and establishing a linearly parameterized dynamic regression equation for online adaptive estimation; constructing a tracking error and a composite error, calculating a control command using a control law, performing online estimation on a parameter vector in the dynamic regression equation by combining Recursive Least Squares (RLS) and a gradient adaptive law, and applying projection and leakage terms to the estimated values to ensure boundedness of the estimated values; and during mode switching, employing parameter memory to migrate estimated parameters from a previous mode to parameters of a new mode, and performing interpolation between a previous control law and a new control law with a smooth weighting coefficient during a switching duration.

Claims

1. An adaptive control method for an inspection robot based on multi-mode motion switching, characterized in that, Includes the following steps: Collect environmental and motion state information, and obtain the robot's pose and velocity through state estimation; determine the currently applicable motion mode based on the environmental and motion state information, and trigger mode switching when the switching criteria and minimum dwell time conditions are met; Based on the current motion pattern, a reference trajectory under this pattern is generated by motion planning, and a linear parameterized dynamic regression equation is established for the current pattern for online adaptive estimation. The tracking error and composite error are constructed, the control command is calculated using the control law, and the parameter vector in the dynamic regression equation is estimated online by combining the recursive least squares (RLS) method and the gradient adaptive law. The estimated values ​​are then subjected to projection and leakage terms to make the estimated values ​​bounded. During mode switching, parameter memory is used to transfer the estimated parameters of the previous mode to the parameters of the new mode, and during the switching duration, a smooth weighted coefficient is used to interpolate between the previous control law and the new control law.

2. The adaptive control method for an inspection robot based on multi-mode motion switching according to claim 1, characterized in that, The switching criteria include logical judgments based on terrain features, robot speed, contact force information, and minimum dwell time, and the time interval must be greater than 0 before the switching is triggered.

3. The adaptive control method for an inspection robot based on multi-mode motion switching according to claim 1, characterized in that, The online estimation of the parameter vector terms in the dynamic regression equation by combining recursive least squares (RLS) and gradient adaptive law includes: The recursive least squares (RLS) method is used to recursively update the linear regression consisting of the regression vector and the measured quantity in the dynamic regression equation, and the RLS includes a forgetting factor. The parameter vector is slowly adjusted using a gradient adaptive law, and the results of the RLS are fused in a weighted or mixed manner as the final parameter estimates.

4. The adaptive control method for an inspection robot based on multi-mode motion switching according to claim 3, characterized in that, The step of applying a projection and a leakage term to the estimated value to make the estimated value bounded includes: The estimated value is subjected to a projection operator to restrict it to a predefined bounded set, and a leakage term is added during the projection process.

5. The adaptive control method for an inspection robot based on multi-mode motion switching according to claim 3, characterized in that, The RLS operates at a high rate only when the regression vector satisfies the persistent excitation condition; otherwise, it operates at a lower rate or only with the gradient adaptive law.

6. The adaptive control method for an inspection robot based on multi-mode motion switching according to claim 1, characterized in that, In the process of interpolating between the previous control law and the new control law with a smooth weighting coefficient during the switching duration, the smooth weighting coefficient is 1 at the start of the switching and decreases continuously and monotonically to 0 during the transition duration T, and the transition duration T is configurable.

7. An adaptive control system for an inspection robot based on multi-mode motion switching, characterized in that, Using the method of any one of claims 1 to 6, the system comprises: The mode determination unit is used to collect environmental and motion state information, obtain the robot's pose and velocity through state estimation, determine the currently applicable motion mode based on the environmental and motion state information, and trigger mode switching when the switching criteria and minimum dwell time conditions are met. The regression equation construction unit is used to generate a reference trajectory under the current motion mode by motion planning, and to establish a linear parameterized dynamic regression equation for the current mode for online adaptive estimation. The parameter estimation unit is used to construct the tracking error and the composite error, calculate the control command using the control law, and perform online estimation of the parameter vector in the dynamic regression equation by combining the recursive least squares (RLS) method and the gradient adaptive law. The estimated values ​​are then subjected to projection and leakage terms to make the estimated values ​​bounded. The mode switching unit is used to transfer the estimated parameters of the previous mode to the parameters of the new mode by means of parameter memory during mode switching, and to interpolate between the previous control law and the new control law by means of smooth weighting coefficients during the switching duration.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1-6.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-6.

Images