| 研究概要 |
The project is about investigating and developing realistic models to study distributed fault-tolerant algorithms for groups of mobile robots. During the research period, we have built several small robots in order to understand what is realistic and what is not. We have come to the conclusion that discrete models better match reality than continuous ones.
In FY 2013, we have obtained results on several problems: formal study of stigmergy and path planning with fundamental bounds in graphs (submitted); fault propagation in graphs (e.g., sensor networks) with simulations (submitted) and theoretical results (in preparation); reliable information propagation in a graph where malicious nodes may tamper with the information (submitted); deterministic and probabilistic gathering under crash and Byzantine faults in a continuous environment (under revision for journal); feasibility bounds for reaching distributed consensus under moving Byzantine faults (submitted).
In addition, motivated by the need to improve robot software reliability, we have also worked on software reliability modeling, with results published in two international conference and two journal papers.
Throughout the research period, we have mostly focused on the question of robot model definition. Our conclusion is that discrete models can provide a very good basis to bridge formal and algorithmic studies and application to practical problems.
|
