Experiments with Diversified Models for Fault-Tolerant Planning
Benjamin Lussier, Matthieu Gallien, Jérémie Guiochet, Félix Ingrand, Marc‐Olivier Killijian, David Powell · 2007
Autonomous robots make extensive use of decisional mechanisms, such as planning. These mechanisms are able to take complex and adaptative decisions, but are notoriously hard to validate. This paper reports an investigation of how redundant, diversified models can be used to tolerate residual design faults in such mechanisms. A fault-tolerant temporal planner has been designed and implemented using diversity, and its effectiveness demonstrated experimentally through fault injection. The paper describes the implementation of the fault-tolerant planner and discusses the results obtained. The results indicate that diversification provides a noticeable improvement in planning reliability with a negligible performance overhead. However, further improvements in reliability will require implementation of a on-line checking mechanism for assessing plan validity before execution.