Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Robot multi-joint self-adaptation compensation method based on perceptron model and stabilizing device

An adaptive compensation and robot technology, applied in the control field, can solve the problems of fixed parameters, limited compensation range of a single-joint stabilizer, and inability to deal with interference, so as to eliminate the center of mass tracking error and body posture error, and avoid the parameter calculation and tuning process. , Improve the effect of network convergence speed

Active Publication Date: 2020-01-21
TONGJI UNIV
View PDF6 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] 1) Due to the characteristics of nonlinearity, strong coupling, and large lag of the humanoid robot system, the calculation and tuning of the parameters of the traditional stable controller are difficult and cumbersome;
[0006] 2) Traditional stability controllers are generally only suitable for a specific interference situation, poor adaptability and robustness, fixed parameters, unable to self-adjust, and unable to cope with complex and changeable environments;
[0007] 3) When the disturbance is large, a single ankle joint or hip joint stabilizer will have a large load on the robot joints, which will reduce the service life of the robot;
[0008] 4) The traditional single-joint stabilizer has a limited compensation range and cannot cope with large disturbances

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Robot multi-joint self-adaptation compensation method based on perceptron model and stabilizing device
  • Robot multi-joint self-adaptation compensation method based on perceptron model and stabilizing device
  • Robot multi-joint self-adaptation compensation method based on perceptron model and stabilizing device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0050] Such as figure 1 As shown, this embodiment is a multi-joint adaptive compensation method for a robot based on a perceptron model, and the method includes the following steps:

[0051] S1: Obtain error data based on the robot inertial sensor and gait generator;

[0052] The error data includes attitude angle error, attitude angular velocity error, robot center of mass tracking error and first order difference of center of mass error. Attitude angle e θ and angular velocity error It can be directly measured by the robot's inertial sensors (gyroscope and accelerometer); the gait generator plans the trajectory of the center of mass, and makes a difference from the actual measured center of mass to obtain the center of mass error e CoM , and then make a first-order difference to get the centroid error differential term

[0053] S2: Input the error data into the pre-established perceptron model, the perceptron model updates the network weight parameters based on the as...

Embodiment 2

[0076] This embodiment is a robot multi-joint adaptive compensation stabilizer based on the perceptron model. The control objects of this embodiment are the trajectory of the robot's center of mass and the posture of the upper body; in order to improve robustness, resist large disturbances, and reduce the load on a single joint of the robot , the actuator is no longer the hip joint or the ankle joint, but the compensation is distributed to all joints of the leg (ankle joint, knee joint, hip joint). Considering that the traditional stabilizer often adopts PD control, and in order to eliminate the center of mass error and attitude error during the walking process of the robot at the same time, the input of the stabilizer in this embodiment is determined as body attitude angle error, angular velocity error, center of mass tracking error and first order difference of center of mass error ; The output is the compensation amount of each joint of the robot leg. In this embodiment, a ...

specific Embodiment approach

[0107] The experimental platform of this example is the Nao humanoid robot of SoftBank Company. The MPC gait planner is used to generate the basic walking gait of the robot. On this basis, the multi-joint self-adaptive compensation stabilizer of the present invention is used to compensate the leg joints of the robot. Eliminate the tracking error of the center of mass trajectory during walking and adjust the posture of the upper body of the robot to improve the stability of walking and adaptability to the environment.

[0108] 1. Parameter initialization

[0109] Such as Figure 4 As shown, the network hyperparameters in this stabilizer include output gain (K) and weight learning rate η. Figure 4 are the structural parameters of the Nao humanoid robot used in this example, wherein Nao's torso is 85mm away from the hip joint (hip), 185mm away from the knee joint (knee), and 287mm away from the ankle joint (ankle). The sensitivity of the center of mass is roughly inversely pro...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention relates to a robot multi-joint self-adaptation compensation method based on a perceptron model and a stabilizing device. The method comprises the steps that firstly, based on a robot inertial sensor and a gait generator, error data are obtained; secondly, the error data are input into the pre-built perceptron model, the perceptron model updates a network weighting parameter based onan associative learning strategy, and compensation value is calculated and output; and thirdly, based on the compensation value, a robot is subjected to motion compensation, and the compensation valueoutput by the perceptron model comprises hip joint compensation value, knee joint compensation value and ankle joint compensation joint. Compared with the prior art, the associative learning strategyis adopted, the network weighting parameter is updated through a supervisory Hebb learning rule, robustness is improved, self-adaptation control is achieved, compensation amount is dispersed to all joints, such as the ankle joint, the knee joint and the hip joint, of a leg, the joint loads are reduced, and the service life of a robot is prolonged.

Description

technical field [0001] The invention relates to the field of control technology, in particular to a perceptron model-based multi-joint self-adaptive compensation method and a stabilizer for a robot. Background technique [0002] The realization of stable walking of humanoid robot generally includes two parts, gait planner and compensation stabilizer. When the robot executes the center-of-mass trajectory and gait generated by the planner, due to the model error of the robot itself and the interference or disturbance of the external environment, there will often be a large difference between the actual state and the target gait when walking, and the walking state of the robot is no longer Stable or even fall. Therefore, in the actual walking process of the robot, it is necessary to design a stabilizer, and make certain compensations to the planning parameters according to the actual state of the robot, so as to eliminate tracking errors, resist interference, and realize the c...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): B25J9/16
CPCB25J9/163B25J9/1653B25J9/1612
Inventor 刘成菊周浩然陈启军
Owner TONGJI UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products