Search

CN-121979255-A - Human-shaped robot walking stable control method and system based on nonlinear model predictive control and readable medium

CN121979255ACN 121979255 ACN121979255 ACN 121979255ACN-121979255-A

Abstract

The invention discloses a walking stability control method, a walking stability control system and a readable medium of a humanoid robot based on nonlinear model predictive control, wherein the walking stability control method of the humanoid robot comprises the following steps of S1, performing gait planning based on the current parameters of the humanoid robot, wherein the gait planning comprises a feasible region for defining a foot drop point and ZMP reference values, S2, performing NMPC optimization on the gait planning and adding flexible control to output new walking parameters to control the walking track of the humanoid robot so as to realize walking stability of the humanoid robot, and the walking parameters comprise the position of the final foot drop point, the optimized walking track, speed and/or acceleration. The technical scheme of the invention solves the problem of robust and efficient walking of the humanoid robot, optimizes the spring damping model and provides a theoretical basis for the flexibility and stability of the humanoid walking.

Inventors

  • WANG HELIN

Assignees

  • 上海应用技术大学

Dates

Publication Date
20260505
Application Date
20260115

Claims (10)

  1. 1. The walking stability control method of the humanoid robot based on nonlinear model predictive control is characterized by comprising the following steps of: step S1, gait planning is carried out based on current parameters of the humanoid robot, wherein the gait planning comprises the steps of defining a feasible region of a foot drop point and ZMP reference values; Step S2, NMPC optimization is carried out on gait planning, and flexible control is added to output new walking parameters to control the walking track of the humanoid robot so as to realize walking stability of the humanoid robot, wherein the walking parameters comprise the position of a final foot drop point, the optimized walking track, speed and/or acceleration, and the flexible control comprises the following steps: Establishing a virtual centroid model; the virtual force controller obtains virtual force through the virtual centroid model, and the virtual force resists the influence of external force so as to achieve compliant control.
  2. 2. The humanoid robot walking stability control method according to claim 1, characterized in that in the step S2, NMPC optimization of gait planning comprises the steps of: acquiring a landing point position parameter; Based on the landing point position parameters, future landing point parameters are obtained according to the following formula (1), and the final landing point position is obtained through NMPC optimization: Formula (1); Wherein, the Indicating the position of the support foot in each time step, The step size in the sampling interval is represented by M, the index value of the sequence summation is represented by M, the step index in the prediction time domain is represented by M, and the walking period is represented by T.
  3. 3. The humanoid robot walking stability control method according to claim 1, wherein the virtual centroid model is obtained by establishing the following formula (2): formula (2); Wherein m o is the mass of the robot, Is the position of the centroid of the robot, For the external force applied to the robot mass center, h represents the height of the gravity center, g represents the norm of the gravity vector, Representing the desired value of the zero moment point, Representing the true value of the zero moment point.
  4. 4. The humanoid robot walking stability control method according to claim 1, wherein the virtual force is obtained by calculation of the following formulas (3) and (4): Formula (3); Formula (4); Wherein, the A virtual force representing the actual output is provided, K 1 is a negative coefficient, Taking critical damping of the robot system.
  5. 5. The robot control device is characterized by comprising a walking mode generator, an anti-interference compliant controller and a nonlinear model predictive control optimizer; The anti-winding flexible controller is used for flexible control by adopting the humanoid robot walking stability control method according to any one of claims 1-4, and joint instructions are generated; The nonlinear model predictive control optimizer optimizes the position of a foothold by adopting the walking stability control method of the humanoid robot according to any one of claims 1-4, and generates a foothold position instruction; the walking pattern generator forms a walking track according to the joint instruction and the foot drop point position instruction.
  6. 6. A humanoid robot is characterized in that, the humanoid robot includes the robot control device according to claim 5.
  7. 7. A humanoid robot control system, characterized in that it comprises the following modules: The gait planning module is used for carrying out gait planning based on the current parameters of the humanoid robot, wherein the gait planning comprises a feasible region for defining a foot drop point and ZMP reference values; The optimizing module is used for NMPC optimization of gait planning and adding flexible control to output new walking parameters to control the walking track of the humanoid robot and realize walking stability of the humanoid robot, wherein the walking parameters comprise the position of a final foot drop point, the optimized walking track, speed and/or acceleration, and the flexible control comprises the following steps: Establishing a virtual centroid model; the virtual force controller obtains virtual force through the virtual centroid model, and the virtual force resists the influence of external force so as to achieve compliant control.
  8. 8. The humanoid robot control system of claim 7, wherein in the optimization module, NMPC optimizing gait planning includes the steps of: acquiring a landing point position parameter; Based on the landing point position parameters, future landing point parameters are obtained according to the following formula (1), and the final landing point position is obtained through NMPC optimization: Formula (1); Wherein, the Indicating the position of the support foot in each time step, The step size in the sampling interval is represented by M, the index value of the sequence summation is represented by M, the step index in the prediction time domain is represented by M, and the walking period is represented by T.
  9. 9. The humanoid robot control system according to claim 7, wherein the virtual centroid model is obtained by establishing the following formula (2): formula (2); Wherein m o is the mass of the robot, Is the position of the centroid of the robot, For the external force applied to the robot mass center, h represents the height of the gravity center, g represents the norm of the gravity vector, Representing the desired value of the zero moment point, Representing the true value of the zero moment point.
  10. 10. A computer-readable storage medium, wherein a computer program is stored on the storage medium, and when the computer program is executed by a processor, the humanoid robot walking stability control method according to any one of claims 1 to 4 is realized.

Description

Human-shaped robot walking stable control method and system based on nonlinear model predictive control and readable medium Technical Field The invention relates to the technical field of robot motion control, in particular to a humanoid robot walking stability control method and system based on nonlinear model predictive control and a readable medium. Background The humanoid robot has wide application prospect in the fields of service, rescue, industry and the like due to the natural compatibility of the humanoid robot with the human environment. However, the humanoid robot is susceptible to external disturbance (such as uneven ground and external force pushing) during walking, so that walking is unstable and even falls. The existing control method depends on zero moment point tracking, but the existing control method is poor in performance under complex terrain or continuous disturbance, and the problems of response lag, tracking error accumulation and the like are outstanding. In the prior art, a preview control method based on ZMP (zero moment point, hereinafter referred to as ZMP) is commonly used, and walking stability is realized by planning a ZMP track and combining linear model predictive control. The part of the method adopts a three-ring control structure with a current ring, a speed ring and a position ring, wherein the position ring and the speed ring are arranged on an upper computer, and the current ring is arranged on a joint driver and is mostly limited to a flat ground and a small disturbance scene. Although the above prior art realizes stable walking of humanoid robots to some extent, when facing complex, dynamic real environments, there are still the following technical fundamental drawbacks, between which there is a causal link: 1) The core goal of the prior art is to passively track a preset ZMP trajectory. The controller can only perform limited compensation when the robot is subjected to a continuous, unidirectional external force, such as walking on a slope, and the gravitational component generates a continuous side thrust. Due to the lack of an active "resistance" mechanism, ZMP can continue to deflect towards the supporting polygon edge, eventually leading to tilting or even falling of the robot. 2) Due to the fact that sensor signal processing delay, complex algorithm calculation time consumption and driver response inertia exist, the pre-observation measurement and control model is too simplified, and dynamic under disturbance cannot be accurately predicted. When sudden disturbance occurs, the response of the controller is slower than the dynamic change of the robot, so that ZMP generates tracking errors. 3) While the prior art includes error feedback in the control loop, the parameters of its control law are usually fixed, and the optimization objective is also mainly directed to ideal trajectory tracking. For unexpected, time-varying ZMP tracking errors due to the above-mentioned hysteresis and disturbances, there is a lack of a specific, strong, robust error convergence mechanism, which makes it possible that once the error is generated, it does not converge rapidly, but continues to exist or even enlarge over multiple gait cycles, eventually destroying the periodic stability of walking, manifested as robot shake or gait disturbances. Based on this, it is desirable to obtain a new humanoid robot walking stability control method that can be less susceptible to external disturbance in complex situations. Disclosure of Invention Aiming at the defects in the prior art, the invention aims to provide a walking stability control method, a walking stability control system and a readable medium for a humanoid robot based on nonlinear model predictive control. In order to achieve the above purpose, the present invention proposes the following technical scheme: In a first aspect, the present invention provides a walking stability control method for a humanoid robot based on nonlinear model predictive control, the walking stability control method for the humanoid robot comprising the following steps: step S1, gait planning is carried out based on current parameters of the humanoid robot, wherein the gait planning comprises the steps of defining a feasible region of a foot drop point and ZMP reference values; Step S2, NMPC optimization (NMPC refers to nonlinear model predictive control, and is called NMPC hereinafter) is performed on gait planning, and flexible control is added to output new walking parameters to control the walking track of the humanoid robot so as to realize walking stability of the humanoid robot, wherein the walking parameters comprise the position of a final foot drop point, the optimized walking track, speed and/or acceleration, and the flexible control comprises the following steps: Establishing a virtual centroid model; the virtual force controller obtains virtual force through the virtual centroid model, and the virtual force resists the influence of external fo