Our Work

Develop a deep learning based full-scene recognition of vehicle environment from a vision sensor. Examples are vehicles, pedestrians, bicyclists, traffic signs, buildings, curbs, etc.

By analyzing over 25,000 high resolution images available on Google Street View across the country, we were able to determine that grass, metal guardrail, concrete divider and curbs are the most common roadside boundary objects.

Estimate the Residual Safety Problem after Integrated Safety Systems (ISS) are deployed in the future. ISS consists of all active (auto braking for vehicle, pedestrian, bicyclist, lane keeping, etc.) and passive safety systems (advanced airbag, curtain shield airbag, roof strength, pedestrian protection active hood, etc.).

Provide theoretical and mathematical framework of how drivers communicate at an intersection.

Identify what kind of communication we have with other road users (e.g., pedestrians, other vehicles) with cutting-edge technology of computer vision.

The objective of this project was to increase the scientific understanding of typical front-seat passenger postures through a naturalistic study observing in-vehicle behaviors.

The objective of this study was to quantify key adult and child occupant kinematic, kinetic and muscle responses from sudden evasive vehicle maneuvers. This data is useful to develop future tools to better assess injury risk in crashes preceded by crash avoidance.

Naturalistic driving data was collected from 110 drivers over the course of a year capturing vehicle telematics data and camera footage, to better understand how bicyclists and cars interact and identify dangerous events.

This study observed volunteer passengers experiencing unexpected abrupt evasive maneuvers, including hard braking and swerving. Our goal was to understand passenger responses to abrupt vehicle maneuvers, to inform the development of onboard safety systems.

We conducted a naturalistic driver study to understand how drivers interact with automated systems in everyday commuting, examining periods where a driver is likely to be highly vigilant along with incidents where driving task may be in low demand.