Biorobotics Laboratory BioRob

Self-reconfiguration for Adaptive Furniture

Roombots are modular robots developed at the BIRG. These robots have a much wider possibility of movement due to their three degrees of freedom than previous YaMoR modules, a former modular robot previously used at the BIRG, and can be used for different tasks. One of these task is to be able to reconfigure themselves into different shapes of furniture.

This project presents an algorithm for self-reconfiguration for roombots. The reconfiguration problem consists of determining the different actions that should be undertaken by a group of attached robots so as to pass from a certain configuration in space to another.

This problem represents a challenge as the configuration space grows exponentially when new roombots are added. So, heuristic methods need to be used to obtain some solutions. The goal is to be able to find a viable solution, as optimal as possible, in a reasonable computation time.

This work is based on the work of Masoud Asadpour that builds a centralized algorithm capable to define a list of actions that will reconfigure a group of YaMoR modules. His algorithm represents spacial configuration as graphs and uses graph edit distance and graph signature.

An adaptation of Asadpour's algorithm for the roombot modules is presented here. It also extends Asadpour's algorithm by changing the definition of a configuration.

Example

Here is a videos that show the resulting reconfiguration of a group of roombot modules. Four modules have to reconfigure themselves from a line to a loop.