The Key Technology Of Automatic Attitude Control And Follow-up Detection Of Ground Penetrating Radar

Tunnels,pipe corridors,bridges and other engineering facilities are easily affected by internal and external environmental factors during their service life,which inevitably leads to the occurrence of diseases such as cracks and voids inside the engineering structure,affecting their service life,and may cause serious economic losses and major safety hazards.It is very important to realize regular inspection of internal defects of engineering structures to ensure the safe operation of various engineering facilities.Ground penetrating radar(GPR)has become the preferred equipment for detecting internal defects in engineering structures.However,conventional engineering structure detection mainly adopts the detection method of manually holding the geological radar or using extension tools to lift the geological radar,which is timeconsuming and laborious in the detection process.Therefore,research on automatic inspection technology for internal defects in engineering structures based on GPR has important research significance and engineering practical value.This paper focuses on the above requirements and studies the automatic attitude control and follow-up detection technology of GPR,and the main research results are as follows:(1)The selection of the GPR survey line position is crucial for automatic inspection of internal defects in engineering structures.Improper selection of the survey line position will directly affect the adjustment range of the mechanical arm that holds the GPR,causing the GPR posture to not be adjusted in a timely manner and collisions to occur during the movement detection process.To address these issues,an optimization method for mechanical arm posture based on the optimal survey line position was proposed;an engineering structure internal defects detection robot with a mechanical arm as its core was developed,which can hold GPR and be mounted on various mobile platforms;based on this,a method for constructing a partial contour model of engineering facilities was established,and a local optimal survey line position search algorithm for the mechanical arm was studied.The mechanical arm posture was optimized with the maximum adjustment range of the mechanical arm as the target to achieve automatic selection of the optimal survey line position for GPR and automatic alignment with the initial posture.A control software platform for engineering structure internal defects detection robots was constructed to achieve visualization of partial contour models of engineering facilities and monitoring of robot status.(2)During the patrol inspection of engineering structure defects,the GPR is required to approach the surface of the engineering structure "while moving and detecting".It’s the core goal to achieve GPR to closely follow the fluctuations in the surface of engineering facilities for follow-up detection.Therefore,a GPR follow-up detection control method based on "PD control+ Kalman filtering" was studied.The relative pose between the GPR and the surface of the engineering structure was obtained through multiple laser ranging sensors,and then a GPR dynamic detection control method was designed by focusing on the fusion of PD speed control algorithm and Kalman filtering state estimation,which solved the problem of GPR jitter and dynamic fitting distance fluctuation caused by mechanical arm adjustment lag in the process of mobile detection,and realized adaptive adjustment of distance between "GPR-engineering facility surface" and automatic detection of internal defects of engineering structures during mobile inspection.(3)Aiming at the requirement of automatic obstacle avoidance in the follow-up detection process of GPR,a GPR automatic obstacle avoidance technology based on solid-state laser radar was studied.According to the characteristics of solid-state laser radar point clouds,an obstacle detection algorithm was constructed,and a height-based segmentation method for measured surface point clouds was designed.This method used multiple laser ranging sensor data as a priori information to accurately segment and remove the point cloud on the measured surface,which reduced the loss of obstacle point clouds,solved the problem of accurate detection and size estimation of small obstacles,achieved automatic obstacle avoidance during GPR follow-up detection,and avoided potential safety hazards brought by obstacles such as line pipes,distribution boxes,and rivets on the surface of engineering facilities.