In this project, we employ event cameras to allow high-speed sensing of a robotic manipulator's environment. The cameras can be stationary mounted to observe the work space; and/or attached on the moving manipulator.
We train spiking neural networks (SNN) to process the incoming camera data in real time; thereby allowing low-latency responses.
We train spiking neural networks (SNN) to process the incoming camera data in real time; thereby allowing low-latency responses.
[Picture]
Sensing to Motor Actions