A common set of HSI metrics does not exist but are needed to assess the humans’ ability to track changes in and accomplish tasks with swarms. HSI metrics cannot necessarily be adapted from other domains (e.g., human-robot interaction) due to the unique characteristics swarms present. Swarms produce macro- and micro-level behavioral differences, have greater uncertainty, potentially visualization overload, and contain >50 entities.  The development of such a set of metrics enables concrete comparisons between alternate algorithms and his methods. Additionally, the relationships between metrics can be explored (e.g., the interactions between the operator performance, system performance and assessment of the task).

Biological swarm research uses a variety of metrics to capture the unique swarm attributes. The biological swarm metrics are useful for artificial swarms because artificial swarm member movement is inspired by biological swarm member movement. The metrics will assess human-swarm interaction. A subset of human-swarm interaction metrics are adapted from human-robot interaction. Human behavior and interaction capabilities are not investigated in biological swarm literature, but have been measured in human-robot interaction with multiple-robot teams. All of the challenges that swarms bring to HSI cannot be addressed with human-robot interaction metrics alone, but they cannot be addressed solely with biological swarms metrics either. A combination of metrics from both domains will create a useful set of HSI metrics.

Related Publications

2014

Inproceedings

Harriott, Caroline E; Seiffert, Adriane E; Hayes, Sean T; Adams, Julie A

Biologically-Inspired Human-Swarm Interaction Metrics Inproceedings

In: Proceedings of the Human Factors and Ergonomics Society Annual Meeting, pp. 1471-1475, Chicago, Illinois, USA, 2014.

Links | BibTeX

Personnel

Caroline E. Harriott

Sean T. Hayes

Collaborators

 Howie Choset

Support

This research is supported by ONR Award Number  N00014-12-1-0987.