The Identification of Terrains for Mobile Robots Using Eigenspace and Neural Network Methods

摘要: autonomous vehicles operate within an increasingly larger set of environments compared to earlier research in which environ- ments were more controlled. In particular, unmanned ground vehi- cles (UGV's) must be able travel on whatever terrain the mission offers, including sand, mud, or even snow. These terrains can af- fect performance and controllability vehicle. Like a hu- man driver who feels his vehicle's response takes appropriate steps compensate, UGV that autonomously perceive its also make necessary changes control strategy. This article focuses development de- tection algorithm based features extracted from induced vehicle vibration. Research is conducted reduce correlation traversing at different speeds. Procedures are presented remove dependencies speed through eigendecompositon methods applying probabilistic neural network for classi- fication between nonlinear boundaries. Experimental results iRobot's ATRV Jr demonstrate iden- tify multi-differentiated broadly defined as grass, asphalt, gravel.