Conception of the Channel Robot Navigation

Mobile robot navigation has gained increasing interest in academical and industrial fields. Navigation task mainly consists of three main parts including object detection and mapping, path planning and motion execution [1]. Various sensors exist which separately or in fusion can acquire different kind of information from surrounding environment recognizing distances, shapes, colors etc. Nevertheless, navigation tasks are becoming more difficult requiring to adjust to varying static and dynamic obstacles and unknown environments. To efficiently deal with adaptive navigation tasks it is important not only to acquire correct data from environment but also to process this data using algorithms for path planning. Adaptivity challenges classical path planning methods like potential field histogram (PFH), visual graph (VG) and Dijkstra [2] and require to look into heuristic methods and in some cases the use of artificial intelligence. This paper contains a review of classical and heuristic methods of path planning and possible combinations analyzing main advantages and disadvantages. Moreover, taking into account presented analysis specific obstacle detection and path planning method is presented for autonomous mobile robot used to navigate in under soil tunnels. Finally, conclusions and research situations are presented.