Improving the Mobility Performance of Autonomous Unmanned Ground Vehicles by Adding the Ability to 'Sense/Feel' Their Local Environment

摘要

This paper explores how a 'learning' algorithm can be added to UGV's by giving it the ability to test the terrain through 'feeling' using incorporated sensors, which would in turn increase its situational awareness. Once the conditions are measured the system will log the results and a database can be built up of terrain types and their properties (terrain classification), therefore when it comes to operating autonomously in an unknown, unpredictable environment, the vehicle will be able to cope by identifying the terrain and situation and then decide on the best and most efficient way to travel over it by making adjustments, which would greatly improve the vehicles ability to operate autonomously.