Development of an interface for digital neuromorphic hardware based on an FPGA

TL;DR

This paper presents the development of an FPGA-based interface for a digital neuromorphic hardware platform, enabling integration with conventional sensors to facilitate experimental research and system monitoring in neuromorphic engineering.

Contribution

It introduces a novel FPGA-based interface for the SPINN-3 neuromorphic hardware, expanding its compatibility with standard sensors for research purposes.

Findings

Robust interface system enabling sensor integration

Enhanced monitoring capabilities for neuromorphic experiments

Facilitates new research approaches in neuromorphic hardware

Abstract

Exploring and understanding the functioning of the human brain is one of the greatest challenges for current research. Neuromorphic engineering tries to address this challenge by abstracting biological mechanisms and translating them into technology. Via the abstraction process and experiments with the resulting technical system, an attempt is made to obtain information about the biological counterpart. One subsection of Neuromorphic Engineering (NE) are Spiking Neural Networks (SNN), which describe the structures of the human brain more and more closely than Artificial Neural Networks (ANN). Together with their dedicated hardware, SNNs provide a good platform for developing new algorithms for information processing. In the context of these neuromorphic hardware platforms, this paper aims to develop an interface for a digital hardware platform (SPINN-3 Development Board) to enable the use of industrial or conventional sensors and thus create new approaches for experimental research. The basis for this endeavor is a Field Programmable Gate Array (FPGA), which is placed as a gateway between the sensors and the neuromorphic hardware. Overall, the developed system provides a robust solution for a wide variety of investigations related to neuromorphic hardware and SNNs. Furthermore, the solution also offers suitable possibilities to monitor all processes within the system in order to obtain suitable measurements, which can be examined in search of meaningful results.