Learning motor skills for sports or performance driving is often done with professional instruction from expert human teachers, whose availability is limited. Our goal is to enable automated teaching via a learned model that interacts with the student similar to a human teacher. However, training such automated teaching systems is limited by the availability of highquality annotated datasets of expert teacher and student interactions as they are difficult to collect at scale. To address this data scarcity problem, we propose an approach for training a coaching system for complex motor tasks such as high performance driving via a Multi-Task Imitation Learning (MTIL) paradigm. MTIL allows our model to learn robust representations by utilizing self-supervised training signals from more readily available non-interactive datasets of humans performing the task of interest. We validate our approach with (1) a semi-synthetic dataset created from real human driving trajectories, (2) a professional track driving instruction dataset, (3) a track-racing driving simulator human-subject study, and (4) a system demonstration on an instrumented car at a race track. Our experiments show that the right set of auxiliary machine learning tasks improves prediction of teaching instructions. Moreover, in the human subjects study, students exposed to the instructions from our teaching system improve their ability to stay within track limits, and show favorable perception of the model's interaction with them, in terms of usefulness and satisfaction.

Field sobriety tests and breathalyzers are commonly used to prevent alcohol-impaired driving, but are expensive and time-consuming to administer. We propose a set of sobriety tests which, in contrast, can feasibly be automated and deployed to modern vehicles equipped with a driver monitoring camera. Our tests are inspired by research on the physiological effects of alcohol, with particular focus on eye movements and gaze behavior. We run an exploratory in-lab study with N=50 subjects (20 alcohol-impaired, 30 control), and train a variety of models to detect alcohol impairment. We find that, using only 10 seconds of observations of the driver, one of the four proposed tests performs comparably to existing non-breathalyzer field sobriety tests. We make our code and data available to support further research efforts to combat alcohol-impaired driving: https://toyotaresearchinstitute.github.io/IV25-beyond-breathalysers/.

Near the limits of adhesion, the forces generated by a tire are nonlinear and intricately coupled. Efficient and accurate modelling in this region could improve safety, especially in emergency situations where high forces are required. To this end, we propose a novel family of tire force models based on neural ordinary differential equations and a neural-ExpTanh parameterization. These models are designed to satisfy physically insightful assumptions while also having sufficient fidelity to capture higher-order effects directly from vehicle state measurements. They are used as drop-in replacements for an analytical brush tire model in an existing nonlinear model predictive control framework. Experiments with a customized Toyota Supra show that scarce amounts of driving data – less than three minutes – is sufficient to achieve high-performance autonomous drifting on various trajectories with speeds up to 45mph. Comparisons with the benchmark model show a 4x improvement in tracking performance, smoother control inputs, and faster and more consistent computation time.

New sensing and steering technologies enable safety systems that work with the driver to ensure a safe and collision-free vehicle trajectory using a shared control approach. These shared control systems must constantly balance the sometimes competing objectives of following the driver's command and maintaining a feasible trajectory for the vehicle. This paper presents a novel technique for creating haptic steering feedback based on a prediction of the system's need to intervene in the future. This feedback mirrors the tension between the two controller objectives of following the driver and maintaining a feasible path. The paper uses simulation and experiment to investigate the impact of varying the prediction horizon on system performance. A novel in-vehicle driver study based on decoupling visual and haptic cues demonstrates that this feedback provides a statistically significant improvement in response time and reduced time to collision (TTC) in an obstacle avoidance task.

New sensing technologies allow modern vehicles to perceive the environment around them even when human visual perception is limited due to poor lighting or fog. Steer-by-wire technology enables active steering capability in which the driver's command to the roadwheels is augmented for maintaining safety. Predictive controllers can leverage both of these technologies to create shared control safety systems that work with the driver to ensure a safe and collision-free vehicle trajectory. The earlier the system intervenes, the smoother the intervention but the more it interferes with the driver's control authority. Ideally, predictive controllers should still intervene late but also indicate upcoming environmental threats to the driver as early as possible. Haptic feedback provides a good means of communicating information to the driver early. Together with a controller still providing a late intervention fallback, this regime provides an ideal framework for predictive shared control systems. This paper presents a novel technique for creating haptic steering feedback, based on future differences between the predictive controller and the driver. This feedback mirrors the tension between the sometimes competing controller objectives of following the driver and maintaining a feasible path. The paper uses simulation and experiment to investigate the inherent trade-offs of predictive haptic feedback and qualitatively discuss its impact.