TTS Lab Seminar | Sep 17, 2020

Abstract: Understanding decisions of human drivers is essential for the development of safe and efficient transportation systems. However, current models of decision making in drivers provide little insight into the underlying cognitive mechanisms. Laboratory studies of abstract, highly controlled tasks point towards noisy evidence accumulation as a key mechanism governing decision making in humans. Yet it is unclear whether the cognitive processes implicated in these tasks are as paramount to decisions ingrained in more complex behaviors, such as driving. I will present the results of a recent project aimed at addressing the gap between studies of naturalistic decision making in drivers and state-of-the-art cognitive models of decision making. We investigated drivers' decision making during unprotected left turns in a simulated driving task, and modelled the cognitive process underlying these decisions. The model explained and predicted drivers’ choices and response times, suggesting that accumulation of dynamic perceptual evidence under time pressure is the major cognitive mechanisms underlying gap acceptance decisions in human drivers. The results of this work exemplify how modern cognitive process models can help us to understand and predict human road user behavior. 

Speaker bio: 

Dr. Arkady Zgonnikov is currently a postdoctoral researcher at AiTech and Department of Cognitive Robotics (3mE). His interdisciplinary background spans cognitive science, computer science and engineering, and applied mathematics. Arkady has previously contributed to the fields of human motor control and decision making by combining mathematical modeling, behavioral experimentation, and machine learning. His current research at TU Delft focuses on enhancing meaningful human control of autonomous vehicles through cognitive modeling of human road user behavior.

Interested in attending this seminar? Drop us an email at ttslab-citg@tudelft.nl, then we can send you the link.