Parker Mitchell

Computer Engineer

NeoN: Neuromorphic Navigation

A Platform for Neuromorphic Navigation for Autonomous Robots

Report forthcoming — expected late spring/summer 2017

What is Neuromorphic?

Neuromorphic Computing, in this context, is a computer architecture which is derived from neuroscience either extracting concepts or attempting to emulate a biological brain. For the development of NeoN, an FPGA implementation of DANNA (Dynamic Adaptive Neural Network Arrays) will be utilized.

What will NeoN do?

Admittedly, not much. The goal of NeoN is to establish the feasibility of using and training Neuromorphic systems for autonomous robotic navigation. NeoN's initial implementation will use LIDAR and will be design to roam an area and use LIDAR to detect potential obstacles. The system will be completely self contained on an FPGA (Xilinx Kintex 7 325T) with all the inputs and outputs tied to FPGA pins, and the FPGA will implement a DANNA grid and the necessary I/O functionality. NeoN will also feature live power monitoring, but because it will use an FPGA, it will not truely showcase the power benefits of Neuromorphic computing.