Issue link: https://iconnect007.uberflip.com/i/1471044
72 PCB007 MAGAZINE I JUNE 2022 e irony of the SAGE computer is that in its last few years of operations, replacement vacuum tubes had to be purchased from the Soviet Union, as they were no longer manufac- tured in the West. PCB007 References 1. The Chinese Sunway TaihuLight is the world's fastest supercomputer. It became active in 2016 at a cost of $273 million. Wikipedia.com. 2. History of SAGE: Semi-Automatic Ground Envi- ronment, Lincoln Laboratory, MIT. 3. "Old Missile Site at McGuire is Still Tainted, Kean Says," New York Times, July 10, 1985. 4. "The United States developed the first hyperson- ic missile in 1949," We Are the Mighty, April 13, 2022. 5. Locklin on science: The largest computer ever built, March 28, 2013. 6. "The Computer Museum Member's First Field Trip to Northbay AN/FSQ7 SAGE Site and to the Canadian National Museum of Science and Tech- nology," by Gordon Bell, Oct. 10, 1982. 7. The SAGE System, Computer History Museum. 8. IBM SAGE specs, bitsavers.trailing-edge.com. 9. "The SAGE Air Defense System, A Personal History," by John F. Jacobs, The MITRE Corporation. Happy Holden has worked in printed circuit technol- ogy since 1970 with Hewlett- Packard, NanYa Westwood, Merix, Foxconn, and Gentex. He is currently a contrib- uting technical editor with I-Connect007, and the author of Automation and Advanced Procedures in PCB Fabrication, and 24 Essential Skills for Engineers. To read past columns or contact Holden, click here. In a new study, MIT researchers demonstrate a machine-learning approach that can learn to control a fleet of autonomous vehicles as they approach and travel through a signalized intersection in a way that keeps traffic flowing smoothly. Using simulations, they found that their approach reduces fuel consumption and emissions while improving average vehicle speed. The technique gets the best results if all cars on the road are auton- omous, but even if only 25 percent use their control algorithm, it still leads to substantial fuel and emis- sions benefits. Typical approaches for tackling intersection con- trol problems use mathematical models to solve one simple, ideal intersection. That looks good on paper, but likely won't hold up in the real world, where traffic patterns are often about as messy as they come. While humans may drive past a green light with- out giving it much thought, intersections can pres- ent billions of different scenarios depend- ing on the number of lanes, how the sig- nals operate, the number of vehicles and their speeds, the presence of pedestri- ans and cyclists, etc. Researchers approached the prob- lem using a model-free technique known as deep reinforcement learning, a trial- and-error method where the control algorithm learns to make a sequence of decisions. It is rewarded when it finds a good sequence. With deep reinforce- ment learning, the algorithm leverages assumptions learned by a neural network to find shortcuts to good sequences, even if there are billions of possibilities. (Source: MIT News) AI Helps Autonomous Vehicles Avoid Idling at Red Lights