Target Capturing in an Ellipsoidal Region with a Swarm of Quadcopter Agents with Different Social Value Orientation Parameters

Abstract

This paper is an extension of [1] and builds on the results in [2] and [1], and investigates the effect of a Social Value Orientation (SVO) parameter on the target capturing problem in an ellipsoidal region by a swarm composed of quadrotor agents. Social Value Orientation is a concept taken from social psychology which has not been studied from the perspective of control of multi-agent systems. Here SVO is encoded as a parameter in the control protocol, resulting in different agent behaviors or in a sense different agent personalities such as social (swarming), prosocial or asocial (selfish) agents. It is guaranteed that agents eventually enter a user-defined ellipsoidal ring around the target and the target will be captured/enclosed. The obtained results show that agents with different personalities act differently while moving towards the target and have different steady-state behaviours or configurations. In particular, selfish agents converge to steady state faster than their more social counterparts. In contrast, social agents interact more with the swarm, and therefore, they stay more cohesive but converge slower to steady state. We present analytical derivations as well as simulations with agents with quadcopter dynamics for both stationary and moving targets.