Artificial hormone network for adaptive robot in a dynamic environment

Abstract

Adaptability is undoubtedly one of the most important characteristics of autonomous robots working in real-world outdoor applications. Due to the fact that real-world outdoor environments are dynamic and uncertain, an ability to adapt to cues provided from environments is a potentially important aspect for adaptable robots. This paper proposes an artificial hormone network as a mechanism to alter robot behaviours in order to lessen the effects of environmental changes. Hormone dynamics are subjected to environmental cues and used to help an autonomous robot to deal with environmental variations. The proposed system is tested on a physically-realistic simulated robot and environment. The results are promising in helping the robot to deal with both internal and external environmental changes.