Research On Trajectory Correction And Positioning Calibration Of Substation Inspection Robot

Substation is an important hub link for voltage level transformation and power resource distribution.With the proposal of the concept of "Big Data,Cloud Computing,Internet of Things,Mobile Internet and Intelligent Cities" in the power grid,the automation process of the power system is also advancing,and the automation and intelligence of inspection methods have attracted more and more attention.The substation inspection robot is based on the concept of automatic inspection,and is gradually replacing the important role of manual inspection in daily inspection to a certain extent.Because the actual circumstances of the substation are complex and there are many disturbances in it,the stability of autonomous walking of the inspection robot could be affected when it is carrying out tasks,ending up with deviating from the planned path or a large deviation to the location of the mission point.On the premise that the basic path planning has been completed,this paper focuses on the problems of trajectory correction and positioning calibration of the substation inspection robot.This paper firstly analyzes the functional characteristics of the substation inspection robot in the actual workflow,and conducts a modular design to build its overall system framework.Secondly,compared with the adventage and disadvantage of various navigation methods used in the actual circumstances of the substation,we choose the path-guided machine vision as our navigation scheme aiming at the interference of the open complex circumstances in substation.In the visual navigation mode,the collected road images are preprocessed by image segmentation,grayscale,and binarization,the center line of the guiding trajectory is extracted,and the virtual calibration line is introduced to calculate the trajectory offset parameters.In view of the problems of overshoot and long time to steady state in the traditional deviation correction algorithm,this paper adds fuzzy control on the basis of the traditional PID control algorithm and Single-neuron PID control algorithm,and gives the gain K dynamic adjustment of the Single-neuron PID control function.Through the MATLAB software simulation,the control ability of the algorithm is compared and analyzed.It can be seen from the results that the Fuzzy Single-neuron PID algorithm has no overshoot and tends to the steady state faster,and improves the trajectory correction control algorithm for the inspection robot.Aiming at the situation that the inspection robot has a large deviation to the location of the mission point,the LANDMARC algorithm,which collects multiple reference tag signals with different weights for position calibration in RFID technology,is introduced to assist the positioning.The positioning accuracy experiment of the algorithm is carried out by MATLAB software simulation,and the results show that the positioning accuracy has been greatly improved when the number of reference tags is 4.In this paper,the simulation test platform based on STM32 is selected to test the overall trajectory correction performance of the inspection robot.After comparing and analyzing the experimental data fed back by the host computer,it is found that in the straight walking area,the maximum value of the trajectory correction time of the inspection robot is 2.2s,and the experimental average is 1.03s;in the turning area of the curve,the maximum value of the trajectory correction time of the inspection robot is 2.5s,and the experimental average is 1.33 s.On the selected simulation test platform,the deflection control effect is better with reference to the industry test standard,which can provide an effective reference for the trajectory deflection control of the substation inspection robot.