US automaker Ford today announced that it is beginning to develop in Israel a decision-making system for driving autonomous cars. Such a system is now a major challenge facing autonomous car developers, because difficult decisions must be made within a fraction of a second in a car surrounded by moving traffic. Sources close to Ford say that the company will invest $12 million in the new activity.
The new team will operate as part of SAIPS, the company acquired by Ford two years ago, and which has since been responsible for developing important parts of Ford’s autonomous vehicle program. Among other things, the company develops a visual cognitive system that makes it possible to detect and identify pedestrians and vehicles in the vicinity of the autonomous vehicle even in difficult weather and lighting conditions and another core system that makes it possible to draw a completely automatic high-resolution 3D map of cities in which the car will drive itself.
Prof. Shie Mannor, a senior faculty member at Technion, Israel Institute of Technology and one of the world’s leading researchers in reinforcement learning, will lead Ford’s new activity. In this field of research, now one of the hot topics in machine learning, the algorithm interacts with the surroundings, thereby learning how to improve its decision-making according to the positive or negative reinforcement it receives. Four graduates of Mannor’s research laboratory are joining him at SAIPS: Dr. Eli Marom, Dr. Gal Dalal, Dr. Assaf Hallak, and Dr. Elad Gilboa. This team will be led by machine learning expert Dr. Boris Sherman. An undisclosed number of additional employees will also be recruited.
Ford CTO and VP research and advanced engineering Dr. Ken Washington said, “Development of safe and reliable autonomous vehicles requires going beyond the limits of current knowledge in computer science and artificial intelligence (AI). In the past two years, SAIPS has been doing this with great success, thereby making a substantial contribution to Ford’s autonomous vehicle program. I am confident that the new addition to the team will expedite innovation and enable us to meet the aggressive targets we set for ourselves on the journey to building smart vehicles.”
SAIPS, which currently has 30 employees, was founded in 2013 by CEO Udy Danino, principal algorithms engineer Noga Zieber (manager of the company’s US branch), and CTO Rotem Littman. “The combination of SAIPS’s proven capabilities and the expertise of the research team under Prof. Mannor’s guidance will enable us to generate innovation all along the road to developing the autonomous vehicle,” Danino said today.
Sources also told “Globes” that Audi, part of the Volkswagen group, is considering the establishment of an active research and development center in Israel to develop advanced driver assistance systems (ADAS). The activity will be conducted through Automated Intelligent Driving (AID) company founded last year by Volkswagen for the purpose of constituting the group’s “expertise center” and developing software and hardware that will make fully autonomous driving possible by 2021.