Abstract
Area coverage is a crucial feature of robots developed for painting, lawn mowing, inspection, and maintenance applications. Reconfigurable robots with shape-changing abilities have been introduced to coverage applications to overcome the limitations of fixed-shape robots. Although the state-of-the-art reconfigurable robots can change their shapes, only a small set of predefined shapes, such as three or seven, are considered resulting in limited coverage performance. Therefore, this thesis proposes context-aware reconfiguration strategies that consider beyond a small set of predefined shapes for reconfiguration to improve coverage performance. The inability to adequately cover near obstacles has been identified as one of the main limitations of state-of-the-art methods. This limitation has been resolved by proposing an obstacle-specific morphology synthesizing strategy that considers beyond a limited set of predefined shapes. The proposed method can significantly improve the area coverage compared to the stateof- the-art methods. The method has also been extended for a robot with any number of blocks. The low-level model and controllers required to facilitate the reconfiguration and navigation beyond a small set of predefined shapes have been developed. The ability of internal rehearsals has been incorporated into the coverage strategy to determine a set of coverage and reconfiguration parameters. Incorporating the internal rehearsal ability has resulted in a significantly improved coverage performance. Continuous navigation and shape-changing ability has been developed to facilitate moving through narrow spaces for coverage which is not possible with the single-step obstaclespecific reconfiguration methods. A Global Coverage Path Planning (GCPP) method to be used in conjunction with the method explored for continuous navigation and reconfiguration has been introduced. Online and offline variants of the GCPP have been formulated. Area coverage and coverage time are often conflicting entities. New designs of reconfigurable robots that can change their size and shape have also been proposed as a part of the thesis. The size adaptability can significantly reduce the coverage time while maintaining the same level of area coverage.