Stochastic Spiking Wireless Multimodal Sensory Systems

Introduction

SWIMS vision is to propose a bio-inspired paradigm change for the design and hardware of future smart wireless multimodal sensory systems that will operate with stochastic spikes. This approach offers a breakthrough with unrivalled energy efficiency at the system level for event detections and communication.

Scientific Breakthroughs

The scientific breakthroughs proposed here are sine qua non advancements for a sustainable deployment of billions of future Internet of Things nodes. These advancements will support smart economy and society development, with large energy savings and limited impact on the environment.

Synergistic Interaction

The synergistic interaction of four complementary skilled PIs enables the realization of a radically-novel end-to-end stochastic analog spiking neuromorphic concept for SWIMS nodes. This concept offers solutions to challenges of:

1. Sensor spiking signal generation
2. Processing
3. Communication

These solutions include disruptive innovations at all levels, inspired by the biological model of a small insect.

Technological Advancements

SWIMS involves many beyond state-of-the-art scientific advancements from technology to system level to enable a neuromorphic architecture with:

i) An input neuron layer featuring new heterostructure spiking sensor arrays based on transition metal oxides/2D semiconductor, covering infrared, ultraviolet, acoustic, and electromagnetic detections.

ii) Hidden layers in tiny spiking neural networks based on novel CMOS Fe-FET concepts capable of efficiently dealing with inherent stochastic noise when processing spiking signals on-chip.

iii) A spiking emitter as an output layer for event-driven wireless transmission using optimized spike modulation and encoding.

iv) A modeling framework embedding stochastic effects in task-based electronic system design and biologically inspired recurrent neural networks.

Experimental Validation

The synergistic interaction will enable unique design and experimental validations of first-of-their-kind event-driven demonstrators. These demonstrators will feature optimized all spiking multi-modal sensor nodes with energy consumption more than 100x lower than existing state of the art.