Ali Hamza, Widener University – Understanding the Future of Automated Self-Driving

On Widener University Week:  Lighting and weather conditions can make self-driving cars struggle to see.

Ali Hamza, assistant professor of electrical engineering, outlines a technological fix.

Ali Hamza is an assistant professor of electrical engineering at Widener University. His research interests include statistical signal and array processing, radar signal processing, communication systems, sparse arrays, convex optimization, and RF sensing for assisted living and remote patient monitoring.

Understanding the Future of Automated Self-Driving

The reality of fully automated self-driving is within view. To achieve this, vehicles require high levels of perception of the surrounding environment. When coupled together, camera and Lidar sensing technologies can potentially yield accurate and precise perception of the surrounding environment and enable superior object recognition and ranging information for self-driving purposes. Their performance, however, is impeded by lighting and weather conditions such as rain and fog which drastically limits the operating range. Radar technology, on the other hand, shows the greatest promise due to its comparatively low cost, low power and resiliency in a wide variety of environmental conditions. In contrast to the other sensor technologies, radar is the only sensing modality offering 4D sensing capability including range, 2D angular localization and velocity.

In my research, I collaborate with industry leaders to develop high-resolution imaging radar technology at a relatively long range to improve the safety and reliability of autonomous vehicles. The goal is to provide significant improvements to the overall radar resolution by developing new high-resolution 2D direction of arrival algorithms that will more accurately discern approaching objects such as pedestrians and other vehicles. This makes autonomous vehicles much safer and more widely accepted. My work is pursuing novel machine learning and AI algorithms to process high resolution radar point cloud data for reliable object recognition which had always been a challenge due to lack of available radar data.

Exploring the capabilities of radar technology will bring us closer to the arrival of safe and affordable self-driving innovation.