Distributed Self-assembly Method Based on Motion-chain for Swarm Robotics
YANG Hongan,KONG Jie,CAO Shuai,ZAN Wenpei,SHEN Gaopan
DOI: https://doi.org/10.3901/jme.2020.07.016
2020-01-01
Journal of Mechanical Engineering
Abstract:： For the versatility and flexibility of robots in complex non-structural environments, a self-assembly approach for a two-dimensional user-specified shape is proposed, which is based on the local interaction and cooperation of distributed and large-scale lattice robots. With the stratified mechanism including layer-by-layer separation, layer-by-layer filling and iteration loop, the macro-level behavior in swarm robotics is transformed to local action of individuals within the edge layers of the current aggregate, which makes the autonomous shaping of complex shape possible. Under the guidance of the stratified mechanism, a self-assembly approach integrating the planning and execution of motion-chain is proposed. With the mechanism of layer-by-layer separation, a motion-chain, a collection of individuals that currently have right to move preferentially, are planned in the edge layer of the current aggregate through local interaction and information propagation; then with the mechanism of layer-by-layer filling, the individuals within a motion-chain move along the outer edge of the remaining aggregate to fill into the edge-filling layer. With the mechanism described above, the problem of “who goes first, how to go and where to go” in the process of self-assembly is solved. Finally, the feasibility and scalability of this novel approach are verified by simulation-based experiments, and the self-assembly approach is implemented on the Rubik robot, a hardware system developed in the lab. With the novel form that a set of particles becomes the multi-function tools by machine-integration, a swarm robot can be controlled to form the user-specified shape autonomous and orderly in a programmable manner.