Explaining human interactions on the road by large-scale integration of computational psychological theory

Abstract

Interaction between road users is a societally important special case of human interaction, and a better understanding of such interactions is a key missing enabler for wide deployment of automated vehicles. Empirical studies implicate a variety of cognitive mechanisms, but these are studied and modeled in separate subfields of psychology. Here, we show how a range of these existing computational theories can be integrated into a single modeling framework, and demonstrate that to reproduce a set of well-established empirical phenomena in driver-pedestrian interaction, substantial theoretical unification of this nature is needed. Our results demonstrate the feasibility and value of large-scale integration of psychological theory, and underscore the formidable complexity of road user interaction, with strong implications for both mechanistic and machine-learned approaches to vehicle automation.

Subtle motion cues by automated vehicles can nudge human drivers’ decisions: Empirical evidence and computational cognitive model

“…Taken together, these findings strengthen the existing notion that noisy accumulation of evidence coming from relevant perceptual variables (distance, TTA, and acceleration) is a key mechanism underlying human tactical decisions in traffic (Pekkanen et al, 2022;Markkula et al, 2022). Besides added explanatory value, we believe our seemingly straightforward addition of the acceleration term to the previously proposed model makes an important step towards applications of the model for understanding and managing human-AV interactions in traffic.…”

“…This obscures the path to incorporating the obtained understanding of human behavior into computational frameworks for AV interaction planning. Our work exemplifies how insights from empirical research on human-AV interactions can be translated into the computational realm, contributing to the recent efforts in this direction (Markkula et al, 2022;Pekkanen et al, 2022;Rettenmaier and Bengler, 2020).…”

“…Recently, drift-diffusion models have been applied to human decisions in traffic, e.g. gap acceptance in pedestrian crossing (Markkula et al, 2022;Pekkanen et al, 2022) and unprotected left turns .…”

Beyond RMSE: Do Machine-Learned Models of Road User Interaction Produce Human-Like Behavior?

“…However, taking a more behavioral perspective forces one to engage with the high complexity of human road user behavior which otherwise gets hidden behind the high-level metrics. A very large number of human road user interaction phenomena have been identified in empirical research [15], [25], [27]. In this work, we have constrained our analyses to highway driving, to keep a feasible scope.…”

Intend-Wait-Perceive-Cross: Exploring the Effects of Perceptual Limitations on Pedestrian Decision-Making

“…The latter may be captured implicitly, however, such unobserved variables cannot be effectively learned as recent findings indicate [6]. Likewise, training on synthetic data with similar characteristics would increase sim-real gap, therefore simulations would greatly benefit from integration of more psychologically-plausible elements [7].…”