3056 FXB BuildingFrançois-Xavier Bagnoud Aerospace Building 1320 Beal Avenue Ann Arbor, MI 48109-2140
University of Michigan
PhD Computer Science and Engineering ’99
MS Computer Science and Engineering ’95
Massachusetts Institute of Technology
MS Aeronautics and Astronautics ’90
BS Aeronautics and Astronautics ’88
AERO 552 Aerospace Information Systems
AERO 450 Flight Software Systems
AERO 201 Introduction to Aerospace Engineering
ROB 550 Robotic Systems Laboratory
AERO 740 (now 552) Aerospace Information Systems
ENGR 151 Accelerated Intro. to Computers & Programming (new)
ENGR 101 Intro. to Computers & Programming
Atkins is motivated by the needs of the Aerospace application, and so are many students. As engineers we are looking to make the world a better place. For aviation, she pursues autonomy research to improve safety of flight and enable new missions. For space, she is interested in augmenting onboard decision systems and supporting closer astronaut-robot collaboration. Increasingly autonomous systems for any vehicle must be resilient to failures and must continue to make rational decisions in the presence of unexpected or anomalous events. These challenge problems require advances in sensing and decision-making. Cloud-based data sources can be fused with real-time sensor feedback to better inform decision systems.
Atkins overall research goal is to identify, adapt, and advance an appropriate set of models and algorithms from the control systems and computer science communities to best solve key Aerospace research challenges. Her PhD research revealed challenges and opportunities in defining the best representation or abstraction for decision-making, particularly when the set of features and values might be incomplete or incorrect. Many presume “state” is fully-defined by a vector of real numbers, yet human cognition is based on symbols that translate to objects, actions, and measures or attributes of each object or action. Dynamics and control system researchers have developed capable and mathematically-correct motion planning, guidance, navigation, and feedback control techniques. A central challenge in application-driven autonomy research is when and how to apply existing techniques versus defining a new abstraction or new algorithm that might be a more effective strategy.
Section “c” of Atkins’ CV describes her past, current, and proposed research projects. These projects span a variety of fundamental and application topics, but they all involve “systems” problems best solved with multidisciplinary models and methods. Atkins’ first project in emergency landing planning relied on an established geometric path construction method, the Dubins path, to connect an initial aircraft state with a landing runway given a no-thrust (gliding) failure case. Perhaps the most important contribution of this work was not in path construction but instead in landing site selection. The simple multi-objective cost function combining common-sense utility terms didn’t receive too much attention, yet every pilot informed of this approach has agreed that inclusion of the more “practical” utility terms beyond time and energy use is important (and novel). This initial research has led to a long series of “emergency flight planning” studies, many in collaboration with researchers who provide essential adaptive control and system identification capabilities underlying the landing site selection and emergency landing planning layer on which my work has focused. The more recent extension to flight safety assessment and management (FSAM) addresses a long-standing challenge of how automation and crew can monitor and serve as safety backups to each other. While neither the deterministic (timed automaton) nor stochastic (Markov Decision Process) modeling constructs is fundamentally novel, perhaps the most important research contributions of this work are in abstracting the state space to forms that efficiently capture the decision space and that can be explained and understood by human operators and air traffic controllers.
Atkins’ research in cyberphysical systems (CPS) has resulted in two important research contributions, discussed in the context of collaborations and supported students in her CV. Both the “co-regulation” and “co-optimization” concepts show promise for offering a new dimension in multidisciplinary optimization. As emerging small UAS, CubeSats, and a variety of other small robotic systems become prevalent in research and commercial applications, we will see an increasing number of cases where computing, communication, and physical sensing and actuation systems must negotiate resource sharing in real-time rather than assuming one subsystem (physical or cyber) dominates.
As increasingly autonomous systems are proposed, Atkins envisions a wealth of new opportunities to inform and exploit cloud-based data, real-time perceptions, and appropriate model abstractions to make optimal decisions for long-duration autonomy and for collaborative human-machine systems. Autonomy is great to study, but an autonomous system is useless unless it ultimately accomplishes a mission that we, the humans, want to accomplish. We must harness the power of increasingly autonomous systems to educate and improve quality of life for people worldwide, not fall into a trap where the next generation grows dependent on autonomous systems without gaining a new evolutionary advantage.
Dr. Ella Atkins is a Professor in the Department of Aerospace Engineering at the University of Michigan, where she is director of the Autonomous Aerospace Systems (A2SYS) Lab. Dr. Atkins holds B.S. and M.S. degrees in Aeronautics and Astronautics from MIT and M.S. and Ph.D. degrees in Computer Science and Engineering from the University of Michigan. She previously served on the Aerospace Engineering faculty at the University of Maryland, College Park. Dr. Atkins is past-chair of the AIAA Intelligent Systems Technical Committee, AIAA Associate Fellow, IEEE senior member, small public airport owner/operator (Shamrock Field, Brooklyn, MI) and private pilot. She served on the National Academy’s Aeronautics and Space Engineering Board (ASEB) (2011-2015 term), was a member of the Institute for Defense Analysis Defense Science Studies (DSSG) Group (2012-2013), and recently served on an NRC committee to develop an autonomy research agenda for civil aviation (2013-2014).
Honors and Awards
Click here for a full list of publications