Discussing the Reality Gap by Comparing Physics Engines in Kilobot Simulations

Abstract

The difference between real experiments and their simulated counterparts, the so-called reality gap, is dependent on various factors and a challenging issue for every simulation-based robot experiment. The reality gap in robot experiments is often caused by software, which is not always able to sufficiently capture particular details and process them properly. In order to minimize this difference, this paper strives to assess different simulation physics engines in V-REP, the simulation framework for the Kilobot swarm robots. The outcome of the simulation software adaption is based on a unified multi robot experiment applied to Kilobots. This paper proposes a simulation attitude, which reflects the outcomes of a real world experiment with Kilobots accurately.

Testing the reality gap with kilobots performing two ant-inspired foraging behaviours

“…These differences, however, followed opposite trends and these observations were confirmed when performing combination tests to control for initial search time of leaders in tandem pairs and for the high variation observed across individual random searches. Our results highlight how difficult it is to use simulations to model the behaviour of kilobots in real life - something that has been reported before in the literature [15], [16]. We discuss these results in the context of ant foraging behaviour as well as in relation to the reality gap in robotics and other applications of kilobots and robotics to biological scenarios.…”

“…Simulations are fundamental to inform planning experiments with real kilobots but, at the same time, cannot fully capture the complexity of the environment a robot is acting in - which is nevertheless significantly simpler than the real world where organisms live and behave. Even in the super-simplified foraging-arena of a robotic lab there are many factors such as individual mass or collisions and traction among robots [14] that might have an influence on the outcome of the robot-robot interactions, but are only partially considered by the simulation software [15]. Some help could come from more advanced physics engines, the programs that simulate physics phenomena or carefully selecting among the available physics engines available in a simulator.…”

“…Some help could come from more advanced physics engines, the programs that simulate physics phenomena or carefully selecting among the available physics engines available in a simulator. A recent study, in fact, reported that Open Dynamics Engine or ODE provides results that are the closest to the output of a real environment among all tested physics engines [15].…”

“…While it has been suggested that experiments with real robots can be avoided if using a different model for the simulated behaviour [12] [13], this approach might miss important aspects of a real life scenario. While this reality gap has been thoroughly documented for swarm robotics [14], [15], [16] there are no works, to our knowledge, that explored the reality gap using kilobots to compare two different strategies of foraging behaviour.…”

“…In our study, we moved one step further and used these fascinating behavioural strategies displayed by ant colonies to implement a series of experiments with physical swarm robots, that were then compared with computer simulations. Past studies, in fact, have revealed how the direct comparison between real life experiments with kilobots and simulation present numerous issues, contributing to the reality gap challenge [15] [14]. These studies combined the use of real robots and simulation to reproduce both stochastic and deterministic foraging behaviours [35].…”