Multimodal Solutions for the Detection of Safety-Related Railroad Objects under Adverse Weather: Architectures, Proof-of-Concepts, and Performance Results
The emergence of high-resolution automotive grade lidar sensor technologies and the development of advanced deep- learning frameworks open the door for accurate and real-time perception of the surroundings of mobile trains for improved safety and operational efficiency. Under this project, we developed key components and proof-of-concepts of lidar-based multimodal solutions that rely on advanced deep- learning methods and utilize object detection approaches to demonstrate the feasibility of real-time detection of objects and events that could represent major safety issues for moving trains. The detected objects could be vehicles or people trespassing a railroad track, or they could be hazardous objects such as falling rocks obstructing a moving train. The project developed multiple deep -learning– based software algorithms that demonstrate the feasibility of a robust real-time railroad safety system. The project also integrated Next-Generation automotive-grade lidars and cameras, which are readily deployed by Autonomous Vehicles (AVs) and Advanced Driver Assistance Systems (ADAS), with deep learning-based object detection solutions to demonstrate the feasibility of achieving high levels of object detection accuracy and robustness under challenging weather conditions. Using automotive-grade sensor technologies can lead to significant savings in cost for the railroad industry while enabling the deployment of accurate real-time detection solutions that have been improving over the past decade, and which continue to improve as new sensors, deep- learning architectures, and other AI solutions emerge.
Hayder Radha
Hayder Radha, professor of electrical and computer engineering, is an international leader in the broad fields of multi-dimensional signal processing and visual analysis. He is especially known for his pioneering work in scalable video coding, with broad applications for Internet and wireless streaming. He is the founder and director of the CANVAS program and a leader in MSU's mobility and autonomous vehicles research. His current research areas include deep learning and statistical signal processing for autonomous systems; multi-modal sensor data fusion; distributed object detection, tracking and forecasting for connected and autonomous vehicles.
He has a B.S. with honors from Michigan State University; an M.S. from Purdue University; and a Ph.M. and a Ph.D. from Columbia University, all in electrical engineering. He joined MSU in 2000, following an outstanding career at AT&T Bell Laboratories, where he received the rank of Distinguished Member of Technical Staff; and Philips Research, where he held the rank of Consulting Scientist and Fellow. As a Philips Research Fellow, he led a team that enabled the rollout of digital HDTV services in the U.S. and another team that developed coding methods, which became part of the MPEG-4 video standard.
He has more than 300 peer-reviewed papers and holds 38 patents. He is a Fellow of the IEEE and the recipient of two Google Faculty Research Awards; the Amazon Research Award; the Microsoft Research Award; the AT&T Bell Labs Ambassador and Circle of Excellence Awards; the MSU William J. Beal Outstanding Faculty and College-of-Engineering Withrow Distinguished Scholar Awards.
-
Contact Us
San José State University One Washington Square, San Jose, CA 95192 Phone: 408-924-7560 Email: mineta-institute@sjsu.edu
