Reduced Deformation Transport of Flexible Objects using Decentralized Robot Networks

Abstract

This dissertation presents and evaluates new control approaches for flexible object transport using robot networks. Recent works have investigated bio-inspired strategies to transport objects using decentralized robot networks that only use local measurements without the need for communication between robots. However, current decentralized theories focus on ensuring state consensus at the end of the transition and not during transition. Deviation of states during transition causes large deformation, which can lead to damage of the object transported. With current methods, deformation can only be reduced by increasing the transport time. In contrast, this dissertation develops a delayed self-reinforcement (DSR) approach for transport tasks to reduce deformation during transport of flexible objects, without increasing transport time. An advantage of the DSR method is that it only uses a delayed self reinforcement of each robot’s actions utilizing prior available data and does not require additional information from the network. This dissertation presents the novel fully decentralized approach using DSR for transporting flexible objects with robot networks, which substantially reduces deformation.