<p>➀ MIT engineers have developed a training method for multiagent systems that ensures safe operation in crowded environments. This method can automatically scale to larger groups by training a few agents.</p><p>➁ In real-world tests, small drones successfully switched positions midflight and landed on moving vehicles. Simulations confirmed that the same training, applied to a few drones, could be extended to thousands.</p><p>➂ The method introduces GCBF+, a new approach for ensuring safe navigation in multiagent systems. It simplifies the computation of safety zones by considering an agent’s sensing radius and using simulations to guide agents while maintaining safety.</p>
Related Articles
- Secretary of Energy Chris Wright ’85 visits MIT14 days ago
- Undersea cable attacks drive sea drone development — Stark's Vanta unmanned vessels could be an affordable solution to protecting vital infrastructure14 days ago
- Concrete “battery” developed at MIT now packs 10 times the power15 days ago
- Palladium filters could enable cheaper, more efficient generation of hydrogen fuel15 days ago
- MIT engineers develop a magnetic transistor for more energy-efficient electronics24 days ago
- Ukraine reveals jammer-resistant Kamikaze strike drones — 31-mile range ordinance promises ’a new level of enemy destruction far behind the front lines’27 days ago
- “Bottlebrush” particles deliver big chemotherapy payloads directly to cancer cellsabout 1 month ago
- New method could monitor corrosion and cracking in a nuclear reactorabout 2 months ago
- Shape Shifting Antenna For Wireless Communicationabout 2 months ago
- Eco Driving Cuts Vehicle Emissions2 months ago