Autonomous Robot Navigation in Diverse Terrain Using a Fuzzy Evolutionary Technique

Abstract

Autonomous navigation of mobile robots is essential for many applications. An increasing number of applications require the robots to operate in diverse terrain conditions, such as earth and space exploration, surveillance, and search and rescue. While optimal path planning has been addressed by many researchers, the traversal of diverse terrain in outdoor or some indoor environments adds another level of complexity. Navigation algorithms utilizing evolutionary computation or fuzzy logic have been proposed. However, they fail to consider the impact of varying terrain conditions. Additionally, they normally focus on either path planning or motion control, but not both. Most lack the ability to adapt to a dynamic environment. This paper proposes a navigation approach that addresses the impact of diverse terrain conditions on determining the optional path using fuzzy sets and evolutionary computation. Fuzzy linguistic variables are used to accommodate the uncertainty and imprecision inherent in determining and representing terrain conditions. The proposed algorithm is capable of rapidly adapting to changes in the environment.