Neuromorphic Engineering

Neuromorphic engineering is a new interdisciplinary discipline that takes inspiration from biology, physics, mathematics, computer science and engineering to design artificial neural systems, such as vision systems, head-eye systems, auditory processors, and autonomous robots, whose physical architecture and design principles are based on those of biological nervous systems [from Wikipedia].
The neuromorphic engineering approach to solve real-world problems typically stems from three important constatations: [From: my Ph.D. Thesis]

Silicon Neurons

Neuromorphic processors (figure at right) have been proposed as an alternative solution to power hungry digital processors, because by emulating the structure and function of biological cells (neurons) and systems (locomotory), they can be used in compact, low-power, and implantable devices.
The silicon Central Pattern Generators that we have developed are composed of 10 and 24 integrate-and-fire neurons using digitally programmable synapses and analog synapses, respectively. Conceptually, each neuron can be divided into three biologically-inspired compartments, one each for the dendrites, soma, and axon.
All the synapses on every neuron integrate charge on a large “membrane” capacitance C_m, with associated membrane potential V_m (figure below).



We use these neurons in appropriately connected networks to produce the same patterns required to produce locomotory gaits. These have been tested on biped robots, and could help in the development of the next generation of neuroprostheses. Also see the videos page for demonstrations of these networks.

Biosignal processing

Due to the significantly higher number of degrees of freedom that characterize the upper limb as opposed to the lower limb, it is crucial that we first be able to decode the intended movements prior to trying to replicate them in a prosthesis. The end goal is to then restore as much of the functionality of the hand as possible. By extracting time-domain features from surface EMG signals from the forearm and upper arm, successful decoding of individual finger movements from a transradial amputee and 4 able bodied subjects was achieved. Furthermore, Kruskal-Wallis tests showed that there is no significant difference in decoding accuracy between able bodied subjects and the transradial amputee. These results pave the way towards intuitive control of individual fingers on the next generation of hand/arm prostheses. This is part of the Defense Advanced Research Project Agency's Revolutionizing Prosthetics 2009 program. The link below provides more details about the project as a whole.
Revolutionizing Prosthetics 2009 >