Analysis of Robotic Fish Using Swarming Rules with Limited Sensory Input

Abstract

Many species of fish school, or swarm, to increase their chances of survival in the wild. For this, they rely on their ability to sense the environment around using a multitude of sensory inputs. In robotics, researches have often tried to mimic both the fish's ability to swarm but also their ability to perceive their environment. In this paper, the effects of different sensor characteristics are coupled with simple swarming rules to determine what a robotic fish would require to be able to interact with other agents in a swarm. The purpose is the further development of a bio-inspired robotic fish designed to interact, and one-day monitor, a school of fish. Due to time, cost and the repeatability to test the design and swarming rules with real fish, the design and development of the robotic fish is being undertaken in a virtual environment. Simulating the robotic fish also allows the parameters such as the type of sensor used to be easily and quickly changed. It also has the additional benefit of being able to use Virtual Reality, which makes modifying and interacting with the simulation easier. This paper demonstrates what effect the different sensor types will have on a robotic fish's ability to swarm with real fish.