Towards minimally complex motion planning in a robot-inclusive space

Abstract

Much research is being conducted on domestic robots and on their introduction to home environments. Still, the complex and dynamic nature of the home is restricting the utilization of the full potential of existing robot technology. Algorithms used in navigating the robot may not ensure smooth manoeuvring unless high-end processors are used, as the complex and cluttered environment of the home normally demands such processors. In our previous research, we presented a case study that compared two commercially available highly complex robots to a proposed minimally complex autonomous three-wheeled domestic robot, and we noted that under the favourable conditions of a robot-inclusive space (RIS), a simpler robot can perform at a level that is comparable to that of a more complex robot. Later, we implemented suitable SLAM algorithms on a three-wheeled domestic robot and selected a low complexity mapping and localization method suitable for RIS environments by tuning the mapping parameters. In the present paper we consider a four-wheeled version of the robot that offers higher mechanical strength and stability. We also investigate methods for motion planning, leading to the selection of a suitable and minimally complex strategy that combines a sampling-based algorithm with appropriate RIS modifications to the home. This strategy is found to enable a robot of relatively simple architecture to navigate efficiently with minimal computational overhead. Additionally, the self-exploring autonomous capability of the algorithm potentially contributes to the primary objective of assisting the elderly and impaired to have independent living.