Local exploration strategies for a mobile robot in a highly dynamic environment

摘要

In this paper, three strategies for navigation in highly dynamic environments are developed and their exploration effectiveness compared. These strategies include a collision avoidance strategy, a random wandering strategy and a goal planning strategy. Existing exploration and surveillance techniques make extensive use of metric maps for localisation and navigation which can be of little use in an environment that has many moving obstacles. While the strategies compared here are not specifically designed for exploration, they allow navigation without the need to store and maintain a map. The evenness or positional distribution over time as well as the speed of exploration are taken as measures of exploration efficacy and applicability to surveillance. The comparison of the strategies shows that the goal planning strategy performs far better than the other two, warranting a significant increase in the complexity and thus computational resources required for successful implementation onto physical platforms.