Research
Laser tracker for dashboard display interaction inside car
I proposed two new interaction devices using a laser pointer with hardware switch as well as laser pointer with eye gaze switch which do not require drivers to physically touch a display.
Gaze controlled HUD inside car
I developed a see-through Heads-Up display (HUD) and explored hands free interaction techniques in the form of eye gaze-controlled interface. The gaze-controlled interface allows the user to look at the intended location on the screen to select the intended target.
Realtime cognitive load estimation of drivers in cars
I developed a method to estimate cognitive load of drivers while driving. In this video, the green circles represent the driver's gaze position and red circles indicate cognitive spikes.
User Acceptance of Interaction Technologies in cars
(Qualitative Research)
A qualitative study on wearable laser-controller based HDD and eye gaze-controlled HUD contributed to explore and understand concerns of professional drivers in India and how these concerns could be addressed
Real-time Multimodal Alert System for Drivers in cars
I developed a method which can detect the increase in cognitive load of the driver while operating secondary task. I also integrated drowsiness detection and finally framed an alert system to give auditory as well as visual alerts.
Comparison of HMI using cognitive load from Pupil Dilation
I developed a procedure to evaluate different HMIs based on the cognitive load of participants using the HMIs. The cognitive load metric is calculated using Pupil Dilation.
Analysis of EEG, Pupil Dilation and Head movement for distraction detection of drivers
I analyzed EEG, pupil dilation and head movement (Kinect) to estimate cognitive load for detecting distraction of drivers while operating secondary tasks.
On-road Obstacle Detection using Deep learning in ADAS
Detection of on-road objects are implemented in GPU using Faster-RCNN for boosted performance in ADAS. This project was done in collaboration with TATA Elxsi.