The Institute of Transportation Studies (ITS) and Partners for Advanced Transportation Technology (PATH) at the University of California Berkeley and King Abdulaziz City for Science and Technology (KACST) in Riyadh, Saudi Arabia took great strides in their new collaboration to develop Truck Automation in Saudi Arabia when ITS and PATH hosted Hatim Bukhari Co-Director at Center of Excellence for Telecomm Applications (CETA) at KACST Aug. 7, 2018.
The new partnership will work to develop an automated truck, with both longitudinal and lateral control, which can be operated in most highway traffic scenarios in Saudi Arabia, including high heat, sandstorms, and congested roadways.
“We have a storied history in connected and automated truck automation research at PATH, and we look forward to moving into a new region and exploring the challenges that comes with it,” says Project PI Xiao-Yun Lu, research engineer at PATH.
The project will focus on automating heavy-duty trucks with longitudinal and lateral control to use in truck platooning, building on the UC Berkeley hardware already in use and ultimately implemented in Saudi Arabia with a new truck and a second truck with the technology installed.
“This is a very exciting opportunity to increase our collaboration in Saudi Arabia and implement our innovative technology we are developing in Berkeley in the rest of the world,” says PI and ITS Director Alexandre Bayen. “I look forward to working with the team at KACST and strengthening our ties and research.”
The recent visit to UC Berkeley by Bukhari, co-PI, introduced him to ITS and PATH and helped lay out the parameters, timeline, procurement and technology of the project, in addition to a demonstration of the technology with a ride on Highway 80 around Berkeley in the connected trucks. The visit will not be the last. The project includes more visits to train the KACST team at UC Berkeley with the Berkeley team, including Bayen, Lu and Graduate Student Researchers Saleh Albeaik, also a KACST Fellow at Bayen's Lab at Berkeley, and Research Associate at Center for Complex Engineering Systems (CCES) at KACST, Fang-Chieh Chou, and Abdul Rahman Kreidieh. Members of the UC Berkeley team will also travel to KACST to assist with the Saudi Arabia implementation and to learn about ecosystem challenges, needs and opportunities with respect to automation and transportation in Saudi Arabia.
The current project includes three phases over the next three years. Phase one and two will develop and demonstrate 2-truck CACC (Cooperative Adaptive Cruise Control) in Saudi Arabia with the first truck driven complete manually; and the second truck will be CACC mode with partial automation, meaning manual steering control and automatic acceleration/braking control. During phase three, the team will develop and demonstrate lateral (steering) control on one truck in Saudi Arabia.
Bayen says the UC Berkeley team is very excited about using deep reinforcement learning for truck automation. In particular, they are benchmarking approaches anchored in control theory and optimization with new algorithms leveraging recent advances in machine learning, the ability for deep learning to learn the behavior of trucks (from simulation and experiments), and the ability for reinforcement learning to create new control paradigms for truck platoons.
Beyond the research and technology, the program will serve as a first concrete step towards establishing an institutionalized partnership between ITS and KACST, and it will also help grow other dimensions of KACST – UC Berkeley interaction, in particular the involvement of KACST alums currently at UC Berkeley in the project.