Validating Extended Neglect Tolerance Model for Search & Rescue Missions Involving Multi-Robot Teams

Abstract

Although, neglect tolerance model have been widely adopted for estimation of robot performances, and robot attention demand in multi-robot teams wherein a single operator handle multiple robots towards accomplishing a given tasks. No systematic studies or theoretical analyses have been carried out to estimate the impact of zero erroneous interaction assumptions in this traditional model while estimating key metrics. In this paper, we put forward an extension of the traditional approach that accounts for any additional demand due to erroneous interactions whilst estimating performance and attention demand metrics in multi-robot scenarios. Experiments were performed with Spido-Bot, a  spider like search and rescue robot across semi-autonomous mode of autonomy where a team of robots searches for disaster victims in a debris filled environment. Measured robot performance, and robot attention demand were largely consistent with the proposed extended neglect tolerance model predictions. We also validate the efficacy of the proposed extended neglect tolerance model by comparing its results with the traditional approach for the same experimental design. The results showed that the proposed approach offers more realistic estimates of robot performance and robot attention demands, due to the inclusion of additional demands incurred due to erroneous interactions into the model.