An insect-based approach to autonomous mobile robot navigation

Autonomy is an important capacity to a robot which works in the circumstances of real world. Navigating a robot in an unfamiliar terrain is a major topic of autonomy. This paper presents an insect-based approach to autonomous mobile robot navigation in an unknown world. In the "N-antennae robot method", the robot is considered as an simple insect with N antennae with which it detects the surrounding environment. Each antenna is activated if it hits an obstacle. The robot architecture is presented and different levels of behavior are defined. The N-antennae robot method is then described. It can be divided into 4 parts: 1. Simulation of an antenna, 2. Relations between the environment and the antennae status, 3. Principles of "N-antennae robot" method, 4. Rule format and program structure. The method was programmed and tested in various environments. Several are only simulated, others are reals. An example of the robot's path illustrates the method. We conclude on the efficiency of the method and we present future developments of this approach using a multi-agent system.<>