Research On Intelligent Obstacle Avoidance Methods For Power Lines Inspection With UAVs

With the development of aviation technology, UAVs (Unmanned Aerial Vehicles) are playing asignificant role in military and civilian fields. But in industrial areas, owing to the complex flyingenvironment and many constraints, establish planning environment and effective planning algorithmhas become the key solution of path planning. This paper, aimed at the characteristics of the powertransmission lines, addresses the following obstacle avoidance algorithms of UAVs.1）Firstly, design the obstacle avoidance control system, including the overall structure and thevarious components. At the same time elaborate sensors of UAVs for obstacle avoidance in detail.2）Secondly, according to the distribution of obstacles in the UAVs flight area, the columnarspace which is the basis of the overall planning of UAVs is established. Also, according to thecharacteristics of the power transmission lines which generate an electromagnetic field, theelectromagnetic field distribution model of the power transmission lines is established and thesimulation software is written, which lay the foundation of real-time obstacle avoidance of UAVs.3）Then, an algorithm based on improving A*algorithm and columnar space and the another onebased on Support Vector Machine and columnar space are proposed. In the first algorithm, affects onUAVs from barriers are introduced to the evaluation function to re-design the heuristic function. Thealgorithm is applied to avoid collision and smooth the path planning. At the same time, the conceptionof Support Vector Machine is introduced to the obstacle avoidance UAVs, and UAVs avoid collisionfrom another perspective.4）Finally, combining obstacle avoidance system of UAVs with distribution model of electricfield around the transmission lines, the research based on fuzzy neural network for UAVs obstacleavoidance methods is studied. By selecting input and output, simplifying the processing, andspecifying the fuzzy rules to establish fuzzy neural network controller, the real-time obstacleavoidance of UAVs is implemented.A series of simulation tests are done for the method proposed above, the simulation test resultsshow that the proposed method is feasible avoidance and effective, and they can satisfy obstacleavoidance for Power Lines well.