Experiments in learning distributed control for a hexapod robot

Abstract

This paper reports on experiments involving a hexapod robot. Motivated by neurobiological evidence that control in real hexapod insects is distributed leg-wise, we investigated two approaches to learning distributed controllers: genetic algorithms and reinforcement learning. In the case of reinforcement learning, a new learning algorithm was developed to encourage cooperation between legs. Results from both approaches are presented and compared.

A biologically inspired approach to feasible gait learning for a hexapod robot

“…In the walking machines domain, GA-based methods have been used to evolve gaits for a number of legged robots, including hexapods (Barfoot et al, 2006;Gallagher et al, 1996;Lewis et al, 1994) and octopods (Jakobi, 1998;Luk et al, 2001). An evolutionary algorithm has also been successfully applied in the process of developing dynamic gaits for four-legged Sony entertainment robots (Hornby et al, 2005).…”

“…Belter and P. Skrzypczyński et al (2006) to evolve walking gaits. Hence, robots are not confronted with the real world as it is, because the additional equipment may cause errors or bias in the reinforcement signal (e.g., due to wheel slippage) or even introduce factors that are not present in the natural environment (Barfoot et al, 2006). Therefore, real robot-based gait learning systems are not completely free from the reality gap problem.…”

Evolutionary-Based Control Approaches for Multirobot Systems

“…Next we report on a number of approaches we have investigated to control and coordinate groups of robots. Our discussion is framed in the context of four tasks motivated by space exploration: heap formation , tiling pattern formation (Thangavelautham & D'Eleuterio, 2004), a walking robot (Barfoot et al, 2006) (wherein each leg can be thought of a single robot), and resource gathering (Thangavelautham et al, 2007). This is followed by a discussion of the common findings across these experiments and finally we make some concluding remarks.…”

Decision Making Process of Hexapods in a Model of Complex Terrains

“…Hexapod represents a six-legged arthropod with a set of three pairs of legs controlled by a nervous system, [1] while relying on the concept of six-legged undercarriage previously introduced in robotics. The model of the cognitive system used for modelling of a six-legged insect`s gait [2] employed a set of sensors (eyes) and actuators (leg muscles), [1,2].…”

“…The model of the cognitive system used for modelling of a six-legged insect`s gait [2] employed a set of sensors (eyes) and actuators (leg muscles), [1,2]. Generally, the cognitive system sends information to the motoric system and receives information from the sensory system, which is a concept known as the sensor-actuator loop [2].…”