Path finding and motion planning in climbing

Abstract

Path planning is a notoriously NP-Hard problem that has intrigued researchers in Computer Science. The automation of such a problem, within a specific given environment, would eventually be capable of identifying possible paths. In most cases such paths would be the shortest or the approximate shortest paths. Some systems, like bi-pedal robots walking on uneven terrain, are able to interact and react to the terrain of the environment which they are exposed to. The system must find a shortest path between a source and the corresponding destination whilst producing walking steps which will keep the robot upright and prevent it from stumbling due to the uneven terrain. Climbing can be considered as different from the casual team oriented sports. This is because success depends solely on the choice done by the climber per move and body position. A wrong move can lead the climber to failure in completion of the route. This can result from reaching a dead end through a sequence implying that there would be no continuation to the ending of the climb. Another reason would be that the chosen holds with which the climber positions his body, would not permit the climber to continue because they put him/her off balance causes him/her to fall. Therefore climbing can be related to a specific set of problems and path plam1ing is considered as the solution to the problem. The aim of this thesis is to build a system which when given a climbing wall, the system would be able to find a sequence of stances that will lead the climber to the top of the wall in such a way as to minimize effort and to maximize comfort. As it is a proof of a concept the system is constrained to 2D environments. The simpler version of the problem, which is, the tackling of solely vertical walls. Yet the same concepts would need to be kept in mind and applied to be able to produce a system able to interact with 3D climbing environments.