Efficient Computing for Low Energy Robotics

Abstract: This talk discusses efficient algorithm and hardware co-design for energy-efficient autonomous navigation processors, which are widely used in low-power and miniature mobile robotics. The proposed solutions deliver orders of magnitude reduction in energy consumption, while also increasing processing speed, for tasks that include localization, exploration, motion planning, and depth estimation. More at https://lean.mit.edu/ Work highlighted in this talk was done in collaboration with Sertac Karaman and Joel Emer Invited talk at 2020 International Conference on Computer-Aided Design