Obstacle Avoidance Path Planning For Deicing Robot On Transmission Lines

The continuous development and stable operation of power system not only provides ahuge momentum to the development of the national economy, but also brought greatconvenience to people’s life. The safety of transmission of electricity is the primary safeguardto stable operation of power system and ice and snow constantly threatening the safeoperation of power system. Hence it has important sense to research the deicing robot, and theautonomous obstacle surmounting of planning is key technology in robot research. Based onthe analysis of the deicing robot working environment, deicing robot obstacle-navigation pathplanning method is studied in this paper.In this paper, the research background of deicing robot are reviewed, and the difficulitiesand key technologies of obstance path planning are analyzed. Then the mechanical structureof deicing robot is introduced, the kinematics analysis is carried out on the deicing robot,kinematics model of mechanical arm(robotic arms) is established, and the constraintconditions of mechanical arm(robotic arms) is used to solve inverse kinematics equation.Based on the analysis of the structure of power transmission line, several kinds of cableaccessories are described.According to the type of obstacle is known, puts forward obstacle recognition methodbased on local feature matching. First, image SURF features are extracted, and BBF prioritysearch method is used to match the feature points of template images and images. Then, thePROSAC robust estimation algorithm is presented to eliminate the error of matched pointsand identify obstacles exactly. Distance measuring system is designed combining laser andinfrared, and the sensor date is fused using Bayesian estimation. This method can effectivelyeliminate the measured value of the error, improve the measurement accuracy. The distancemeasuring system realizes accurate ranging of obstacles, to ensure the deicing robot canperform actions at a safe distance. Experimental results show that at the premise of real-timeperformance, the ranging accuracy is improved after adopting multiple-sensor fusionmeasuring system.Path planning with obstacles avoidance is addressed based on expert system, by adoptingthe combination of offline and online planning approach. Obstacle-climbing motions areplanned for different obstacles in the laboratory environment and the planned actionsequences are stored in repository. Depending on the type of obstacles, deicing robot choosethe appropriate sequence of actions to perform obstacle actions. Through the sensorinformation regulate deicing robot, and update the repository. This way of the combination ofoffline and online planning reduces the amount of computation, and improves the accuracy and real-time performance. The experimental results show that the control strategy cancomplete the obstacle path planning.