Neuro-inspired Architectures for Inference and Control in Massively Scalable Multi-Agent Systems

Vertical Tabs

Neural and motor systems in biological organisms have evolved over millennia to provide complex functionalities with simple building blocks. It is therefore natural to turn to these as a source of inspiration for the design of large-scale manmade multi-agent systems: scale requires simplifying the design of individual agents (which are analogous to individual neurons or muscle fibers), but these simple agents must interact as in the neural and motor systems to produce complex behavior.

Vertical Tabs

The broad goal of this project is to take multi-agent systems and sensing to new levels of scalability and reliability, drawing inspiration from both the high-level organization and low-level “front end” processing of the neural and motor control systems. The proposed neuro-inspired designs draw upon the close collaborative ties that we have established with neuroscientists within ICB over the past couple of years, and require a close coupling of, and innovations in, both control and signal processing. Our work continuous to be focused on the following key enabling scientific objects: understanding sensory front ends, understanding neuro-inspired information processing, and understanding neuro-inspired control architectures.

University: 

UCSB

Task Order: