Case Study Abstracts

Case Study 1

Who's the Boss? - A Haptic Interface for Negotiating Control Authority between Human Drivers and Automation Systems

Contributors:  
Faculty Tulga Ersal, Brent Gillespie, Jeffrey Stein (University of Michigan)
Post-doc Amirhossein Ghasemi (University of Michigan)
Students Akshay Bhardwaj, Huckleberry Febbo, Yingshi Zheng (University of Michigan)
Government Paramsothy Jayakumar (US Army TARDEC)
Industry Mitch Rohde and Steve Rohde (Quantum Signal)
John Walsh (Ford)

        Recognizing the potential for autonomous systems to dramatically increase mobility in both military and civilian applications and their critical role in the Third Offset Strategy of the Department of Defense, this case study brings together three projects to explore how control actions generated by an autonomous system and a human driver can be combined to achieve greater mobility than achievable when either control action acts alone. We first present an autonomous navigation algorithm designed to fully exploit the dynamic limits of a vehicle to maximize its mobility when navigating through an obstacle field. Supposing that intermittent sensor faults might preclude fully autonomous operation, we then use the steering wheel as a haptic interface to facilitate smooth negotiation of control between a remote human driver and the automation system. We explore a control sharing paradigm in which the autonomous system applies its control effort through a finite mechanical impedance such that the human driver can feel and in effect “edit” the autonomous system’s control actions. Likewise, the automation can edit the human driver’s control actions. We compare performance at lane keeping and obstacle avoidance under this haptic control sharing paradigm to performance under more traditional alternatives using a fixed base driving simulator. In addition, we consider communication delays between the remote human driver and the vehicle and explore the benefits of a novel predictor framework to attenuate the negative impact of delays on mobility performance in this haptic shared control paradigm.

Case Study 2

Finding the MARVEL in the Hay Stack:
A Case Study on Modular Adaptive Resilient VehiclE FLeets

Contributors:  
Faculty Bogdan Epureanu, Panos Papalambros (University of Michigan)
Zissimos Mourelatos (Oakland University)
Post-docs A. Emrah Bayrak, Mert Egilmez (University of Michigan)
Students Xingyu Li, Arianne Collopy (University of Michigan)
Themistoklis Koutsellis (Oakland University)
Government Matthew Castanier, Richard Gerth, Michael Kerr, Chad Wilson (US Army TARDEC), Jeff Bradel (Code 30), Ra'ed Seifeldin (Vencore)
Industry Edward Umpfenbach (General Motors), Terrance Wagner (Ford), Randy Jaeger (Applied Minds), Clint Hope (Applied Minds), William Ross (Nevada Automotive Test Center)

        Revolutionary adaptability and resilience are two of the key needs of the 3rd offset. Paradigm changes are required in fleet operation and design to accomplish these needs. Modularity promises to provide significant benefits in terms of adaptability and resilience especially when combined with autonomy through self-assembly and self-reconfiguration. Modular autonomous systems are complex systems of systems which require advanced design and management approaches. This case study presents a unique approach to the design, dynamic operation and reliability of modular vehicle systems. The design study refers to a systematic process for decisions to generate modular design concepts to create the vehicle fleets of the 3rd offset; dynamic operation addresses the fleet management to schedule and conduct on-base and in-theatre missions such as vehicle assembly, resupply and convoy operations considering the entire fleet dynamics; and reliability analysis models the failures in the fleet as a repairable system based on limited observations, and provides quantitative predictions for future failures based on previous ones to enhance the overall system reliability. The analysis in this case study is applied to the Vehicle Agnostic Modularity (VAM) program of the Office of Naval Research to support decisions from design to deployment and operation of innovative modular vehicle fleet concepts; VAM seeks to assess the efficacy of modularity to the USMC ground vehicle fleet.