A Fuzzy Logic Method for Autonomous Robot Navigation in Dynamic and Uncertain Environment Composed with Complex Traps

Abstract

Nowadays, mobile robotics is becoming one of the most open research areas due to the fact that it includes many quite complex advanced mechanisms that are difficult to get a thorough mastery of. In fact, the control task becomes much more complex. The problem is how to guarantee safety for the robot while moving toward a known goal. For this reason, simulation plays a very important role in robotics. It is believed to be the most suitable procedure that may offer a fast prototyping of mobile robots, by considering all options even before fixing the technological devices. Multi-agent simulation models are widely used to solve problems in mobile robotic. In fact, there are similarities between the agent and the robot. However, the most popular methods are based on reactive local navigation schemes that tightly couple the robot actions to the sensor information. But the robot needs not only to avoid collision at real time, but also to anticipate environment evolution in order to take suitable decisions. In this paper, we present a multi-agent simulation model of an autonomous robot that takes into account the reactivity and the anticipation of blocking situations in dynamic and uncertain environments by using only ultrasonic sensors.