RBE Colloquium Series Presents
Dr. Brian Mekdeci
Engineering Systems of Autonomous Systems for Survivability
Abstract: Stakeholders of large-scale systems, such as public transportation, package delivery and emergency services, are considering replacing or supplementing manned assets with highly autonomous machines. The larger these systems are, the more critical it is that they survive internal and external disturbances. Since building prototypes of large systems is often infeasible, accurate models are necessary for engineers to assess the performance of design choices before and after disturbances. Despite the label, “autonomous” machines require human attention for critical tasks that are too complicated or too risky. Unfortunately, the behavior of humans-robot teams is notoriously difficult to model, particularly when the humans (or sometimes the robots) behave irrationally. To understand and predict the performance of these intelligent entities, “soft” issues must be addressed, in areas such as psychology and organizational behavior. Although engineers are typically responsible for designing and building systems of autonomous systems, their success may depend upon how comfortable the engineers are at handling non-STEM problems.
This lecture introduces some of the research done by Dr. Brian Mekdeci to model human-robot teams and improve the survivability of systems of autonomous systems. Topics such as organizational behavior, the Yerkes–Dodson law of human performance, queuing theory, design of experiments, serious games and agent-based modeling are all tools that Dr. Mekdeci has used in his survivability of autonomous systems research and will be discussed.
At the end of the lecture, Dr. Mekdeci will also give some brief opinions on how robotics engineering should be taught based on his academic and industry experience around the world. Dr. Mekdeci will compare schools that focus on theory with those that take a more practical “hands-on” approach, weighing the pros and cons of each.
Bio: Dr. Mekdeci has studied and worked with autonomous systems for over a decade. He has a B.A.Sc. and M.A.Sc. from the University of Waterloo in Canada, and a Ph.D. from M.I.T. He spent a semester at the Singaporean University of Technology and Design researching how autonomous vehicles can be used to improve maritime security systems and most recently worked at the University of South Australia modeling the impact of autonomous technology on the Australian Defense Force. Dr. Mekdeci has worked in industry for several years including positions as an embedded software developer, a systems engineer and a consultant. Prior to studying at M.I.T., he was the video and human factors engineer at CDL Systems in Calgary (later acquired by Lockheed Martin) where he helped design and test software that controlled unmanned aerial vehicles (UAVs) for the U.S. Army. While at CDL Systems, Dr. Mekdeci co-authored the VCS-4586 Style Guide which was adopted as the official U.S. Army style guide for UAV ground control stations.
Friday, Jan. 25, 2019 - 2:00 p.m. - 3:00 p.m. - 60 Gateway Park, GP 1002