GOALS: To design an advanced driver assistance system (ADAS) for automobiles to enhance driver safety, e.g., by tracking lane markers to keep the driver in the correct lane and alerting the driver to other cars and pedestrians. While image processing algorithms exist to track lanes, cars, and pedestrians, the challenge in this project is to implement these algorithms on an embedded board (with CPUs and GPUs) and a very limited power budget.
KEY ELEMENTS: Image processing, real-time object recognition, embedded systems, low-power design, parallel programming, graphics, user interface design, sensors, assistive technologies.
RESEARCH ISSUES: Low-overhead and accurate real-time object recognition; parallel software implementations of image processing algorithms on GPUs; real-time sensing and tracking; low-power optimizations; driver user-interface design.
MEETING TIME: Wed, 2:30 – 3:30
- Sudeep Pasricha (ECE), firstname.lastname@example.org
PARTNERS & SPONSORS: U.S. Department of Energy, General Motors, MathWorks, dSPACE, Freescale.
MAJORS, PREPARATION, INTERESTS:
- ECE – Background/interest in embedded systems, parallel programming, low-power design, sensors, CPU/GPU computing.
- CS – Background/interest in image processing, real-time object recognition, HCI/User-interface design, GPU programming .
- ME – Background/interest in automotive electronics, control systems, driver assistance systems.