Qadir - Self-Sufficiency of an Autonomous Reconfigurable Modular Robotic Organism

dSPACE GmbH, Paderborn, Germany

Raja Humza Qadir

Email:

Different principles in artificial autonomous systems have been devised to mimic the behavior of biological systems, observed in nature, for instance, the division of labor, trophallaxis, task sharing, leaving pheromone traces, etc. But, it is still challenging to establish both physical and virtual collaboration between multiple autonomous robotic module of a modular robotic organism. In the past, to keep the energetic autonomy of an autonomous robotic system in a swarm the problem of self-sufficiency has been explored and presented in a variety of scenarios, for instance, harvesting energy using biological means, e.g., sugar (Wilkinson 2000), digesting slugs (Ieropoulos et al. 2005b), and through artificial means, like fuel cells (Kelly et al. 2000), solar panels (Boletis et al. 2006; Landis and Jenkins 1997), recharging from stationary power stations (Munoz et al. 2002), swapping of battery packs (Ngo 2008), etc. But, so far the said principle has not been explored from the perspective of a modular robotic organism.

The self-sufficiency principle in a similar way for a modular robotic organism can be interpreted as the collective ability of autonomous robotic modules in an organism to find fuel and refuel. To accomplish this objective collectively, the robotic modules in a modular robotic organism are therefore must be remained operational till it successfully docks to an available recharge station in the arena. In this regard, to keep the physical collaboration between the robotic modules in a modular robotic organism despite their varying individual energetic status, the hardware design of the robotic modules must support power sharing and fault tolerance to withstand abrupt endogenous and exogenous faults and failures. Studying the platform designs of existing state-of-the-art reconfigurable robotic systems like SuperBot (Salemi et al. 2006) and ATRON (Lund et al. 2005) revealed that their solution although support power sharing but lack certain design features that are considered by the author as vital for dynamic power sharing between the robotic modules of an organism. For example, from the perspective of power sharing, a robotic module must be able to measure and control the current flowing from its battery pack to its on-board system components and to the docked robotic modules. Depending on a systems implementation, the current flow measurements at different components may allow a robotic module to dynamically adapt its behavior during different behavioral states in a modular robotic organism. In addition, the existing state-of-the-art reconfigurable robotic systems platform designs lack fault tolerance at the hardware layer that is essential to keep the physical collaboration of the robotic modules despite faults and failures in a modular robotic organism.

Considering the morphology of a modular robotic organism, in addition, the system design of an autonomous robotic module must bear features that allow it to control its power sharing behavior in conjunction with the docked robotic modules. In other words, the application software components of a robotic module must have an access to the innate system features, embedded in the hardware design, in order to cope different situations in a modular robotic organism, in particular, non-uniform or varying energetic status of the robotic modules and collective behavior of robotic modules in the presence of faults or component failures. So far, no such research work has been seen or conducted that explore the behavior of a modular robotic organism from the perspective of self-sufficiency.