Traditional imitation learning provides a set of methods and algorithms to learn a reward function or policy from expert demonstrations. Learning from demonstration has been shown to be advantageous for navigation tasks as it allows for machine learning non-experts to quickly provide information needed to learn complex traversal behaviors. However, a minimal set of demonstrations is unlikely to capture all relevant information needed to achieve the desired behavior in every possible future operational environment. Due to distributional shift among environments, a robot may encounter features that were rarely or never observed during training for which the appropriate reward value is uncertain, leading to undesired outcomes. This research is developing techniques to overcome uncertainty despite a minimal set of human demonstrations in order to operate safely in dynamic environments. It also seeks to leverage quantitative methods from reliability and resilience engineering in support of rigorous assessment of autonomous systems.
3. C. Ellis, M. Wigness, and L. Fiondella, “A Mapping of Assurance Techniques for Learning Enabled Autonomous Systems to the Systems Engineering Lifecycle” In Proc. IEEE International Conference on Assured Autonomy (ICAA), Fajardo, Puerto Rico, Mar 2022.
2. C. Ellis, M. Wigness, J. G. Rogers, C. Lennon, and L. Fiondella, “Risk Averse Bayesian Reward Learning for Autonomous Navigation from Human Demonstration” in Proc. International Conference on Intelligent Robotics and Systems (IROS), Oct 2021
1. A. Gula, C. Ellis, S. Bhattacharya, L. Fiondella, “Software and System Reliability Engineering for Autonomous Systems incorporating Machine Learning” In Proc. 66th Annual Reliability and Maintainability Symposium (RAMS), Palm Springs, CA, Jan 2020. Society of Reliability Engineers Stan Ofsthun Best Student Paper Award.
Presentations and Tutorials
4. C. Ellis, M. Wigness, and L. Fiondella, Assurance Techniques for Learning Enabled Autonomous Systems, Presented to WG 35 AI and Autonomous Systems, 90th Military Operations Research Symposium (MORS 2022), Quantico, VA, June 2022.
3. C. Ellis, M. Wigness, J. Rogers, C. Lennon and L. Fiondella, Risk Averse Autonomous Navigation from Human Demonstrations, Presented to WG31 Operational Environments and WG35 AI and Autonomous Systems, at the 89th Military Operations Research Symposium (MORS 2021), Quantico, VA, June 2021.
2. A. Gula, C. Ellis, S. Bhattacharya, and L. Fiondella, Relationships between Machine Learning and Reliability Engineering, Presented to WG17 Logistics, Reliability and Maintainability and WG35 AI and Autonomous Systems, at the 88th Military Operations Research Symposium (MORS 2020), New London, CT, June 2020.
1. A. Gula, C. Ellis, S. Bhattacharya, and L. Fiondella, Reliability Engineering of Autonomous Systems incorporating Machine Learning, presented at the 7th Annual Systems Engineering Research Center (SERC) Doctoral Students Forum, a University-Affiliated Research Center of the US Department of Defense, Stevens Institute of Technology, Washington DC, November 2019.