icimagesEmerging medical applications are seeing a new breed of implantable and wearable devices for health monitoring and therapy. Portability and thus also battery lifetime poses stringent power requirements requiring ultra low power embedded electronics. Furthermore, the wireless streaming of raw data to the cloud is not viable in a number of these applications, either due to power constraints, the requirement for real-time closed-loop control or due to the life-critical nature of these systems. It is therefore essential to realise high performance embedded systems that provide ultra energy (and silicon area) efficient computation, beyond the current state-of-the-art. This can be achieved by exploiting hybrid processing (and signal representation) techniques, combining traditional computational methods with sub-threshold analogue and continuous-time (asynchronous) digital processing. Such hybrid application-specific systems can achieve orders of magnitude higher hardware-efficiency compared to traditional embedded systems.

My research utilises mixed-signal integrated circuit and microsystem technologies to address such challenges in implantable neural prosthetics, brain-machine interfaces, lab-on-chip/wireless capsule endoscope platforms and medical devices in general. My main focus is to develop high performance embedded systems that interface with neural pathways for restoring lost function in sensory, cognitive and motor impaired patients.

For more details on my research, visit: http://www.imperial.ac.uk/people/t.constandinou/research