Using a combination of recent developments, ranging from computer vision to decentralized estimation and control, this project will develop a deep-learning Bayesian-optimization framework hinging on sparse features for mobile cooperative scene perception. The methods developed in this project will be tested using real video data from Cornell’s campus as well as virtual data generated using a realistic game engine.
Cornell Chronicle: Researchers link robots to surveillance teams
recode: These surveillance robots will work together to chase down suspects
Cornell Research: Collaborative Robotic Surveillance
Our recent work using Ms. Pac Man as a benchmark problem for optimal control strategies has been featured on BGR, Tom’s Guide, and in the Cornell Chronicle! You can read more about how our control strategy beat the previous Ms. Pac Man AI record here!
BGR: Scientists built an AI that is really, really good at ‘Ms. Pac-Man’
Cornell Chronicle: Engineers eat away at Ms. Pac-Man score with artificial player
Tom’s Guide (FR): Cette IA est imbattable à Miss Pac-Man
The New Yorker: Could Ms. Pac-Man train the next generation of military drones?
Our paper by Hongchuan Wei, et. al. “Information value in nonparametric Dirichlet-process Gaussian-process (DPGP) mixture models,” has been published in Automatica, Volume 74, December 2016, p. 360-368. The paper is available on Science Direct here or a PDF is available here.
We had two papers accepted to CDC 2016!
- P. Zhu, J. Morelli, S. Ferrari, “Value Function Approximation for the Control of Multiscale Dynamical Systems,” Proc. of the IEEE Conference on Decision and Control, Las Vegas, NV, December 2016, in press. [PDF]
- T. S. Clawson, S. Ferrari, S. B. Fuller, R. J. Wood, “Spiking Neural Network (SNN) Control of a Flapping Insect-scale Robot,” Proc. of the IEEE Conference on Decision and Control, Las Vegas, NV, December 2016, in press. [PDF]