This thesis addresses autonomous exploration and area mapping with a mobile robot
(iRobot Roomba). We propose a hybrid controller that combines frontier-based
global goal selection at the boundary between known and unknown space with a local
Bug2 algorithm that follows walls and returns to the direct line toward the current
goal. A Random Walk algorithm serves as the baseline. The Python implementation
logs telemetry, builds an occupancy map in real time, and automatically produces
coverage-over-time reports. We evaluate final coverage, path length, number of
collisions, and the times to reach 50% and 90% coverage. In a 2000mm×3000mm
environment, the proposed hybrid consistently achieves 100% coverage and converges
substantially faster than our baseline, with comparable or lower variability across
runs.
|
