Adaptive Causal Models for Fault Diagnosis and Recovery in Multi-Robot Teams

Parker, Lynne E.; Kannan, Balajee

doi:10.1109/iros.2006.281993

Cited by 22 publications

(22 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our earlier experience on developing multi-robot fault-diagnostic architectures [15], [12], convinces us that it is unlikely that the human designer can anticipate every possible fault that the multi-robot system may encounter a priori. Thus, a multi-robot team can achieve a much higher level of faulttolerance if it is able to autonomously adapt over time.…”

Section: Related Workmentioning

confidence: 94%

“…The second architecture that we implemented, called LeaF [12], is an adaptive method that uses its experience to update and extend its causal model to enable the team, over time, to better recover from faults when they occur. LeaF combines the existing CMM strategy with case-based learning algorithm to adapt and categorize a new fault and add it to the causal model for future use.…”

Section: Evaluation Of Metricsmentioning

confidence: 99%

“…At a higher level, the entire process of fault representation and diagnosis can be viewed as a fully connected graph, with nodes representing faults and edges highlighting the relation between the faults. Much more detail on LeaF can be found in [12] and [8].…”

Section: Evaluation Of Metricsmentioning

confidence: 99%

“…Despite LeaF's ability to recover from all the encountered faults over the course of the experiments, it is important to note that the value for robustness for LeaF is misleading. From our earlier work [12], we know that the larger the number of interacting components in an environment, higher is the probability of failure. We believe eventually LeaF will also encounter fault(s) that it does not handle well, similar to the sonar case in the box pushing task.…”

Section: B Experiments II -Deployment Taskmentioning

confidence: 99%

“…However, while redundancy by itself is a useful measure [6], it is incomplete as an evaluation metric, since a system can also be effectively fault-tolerant through other reasoning methods. Additionally, in our earlier work ( [12]), we detailed empirical results that show the inability of a system to identify all faults that may occur without incorporating an experience based online-learner. As a consequence, we can say that for a system to be truly robust to change, it needs to exhibit adaptability to counter the dynamic nature of its operating environment.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Metrics for quantifying system performance in intelligent, fault-tolerant multi-robot teams

Kannan

Parker

2007

2007 IEEE/RSJ International Conference on Intelligent Robots and Systems

Self Cite

View full text Add to dashboard Cite

Any system that has the capability to diagnose and recover from faults is considered to be a fault-tolerant system. Additionally, the quality of the incorporated fault-tolerance has a direct impact on the overall performance of the system. Hence, being able to measure the extent and usefulness of faulttolerance exhibited by the system would provide the designer with a useful analysis tool for better understanding the system as a whole. Unfortunately, it is difficult to quantify system fault-tolerance on its own for intelligent systems. A more useful metric for evaluation is the "effectiveness" [6] measure of faulttolerance. The influence of fault-tolerance towards improving overall performance determines the overall effectiveness or quality of the system. In this paper, we outline applicationindependent metrics to measure fault-tolerance within the context of system performance. In addition, we also outline potential methods to better interpret the obtained measures towards understanding the capabilities of the implemented system. Furthermore, a main focus of our approach is to capture the effect of intelligence, reasoning, or learning on the effective fault-tolerance of the system, rather than relying purely on traditional redundancy based measures. We show the utility of the designed metrics by applying them to different fault-tolerance architectures implemented for multiple complex heterogeneous multi-robot team applications and comparing system performance. Finally, we contrast the developed metrics with the only other existing method (HWB method [6]) for evaluating (that we are aware of) effective fault-tolerance for multi-robot teams and rate them in terms of their capability to best interpret the workings of the implemented systems. To the best of our knowledge, this is the first metric that attempts to evaluate the quality of learning towards understanding system level fault-tolerance.

show abstract

Section: Related Workmentioning

confidence: 94%

Section: Evaluation Of Metricsmentioning

confidence: 99%

Section: Evaluation Of Metricsmentioning

confidence: 99%

Section: B Experiments II -Deployment Taskmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations