Vision-Based Reaching Using Modular Deep Networks: from Simulation to the Real World. (arXiv:1610.06781v1 [cs.RO])

In this paper we describe a deep network architecture that maps visual input to control actions for a robotic planar reaching task with 100% reliability in real-world trials. Our network is trained in simulation and fine-tuned with a limited number of real-world images. The policy search is guided by a kinematics-based controller (K-GPS), which works more effectively and efficiently than $\varepsilon$-Greedy. A critical insight in our system is the need to introduce a bottleneck in the network between the perception and control networks, and to initially train these networks independently.

from cs.AI updates on arXiv.org http://ift.tt/2em31Mk
via IFTTT