Neuromorphic Sensor-Perception Systems for Immersive Displays

An immersive display implementing enhanced virtual and augmented reality demands a higher level of human-machine interaction and requires input from a variety of sensors that monitor human activities. However, the processing of a large volume of sensory data challenges the classical von Neumann computing architecture in terms of latency and energy efficiency. Made possible by the sharing of a common metal-oxide (MO) thin-film transistor (TFT) technology, the monolithic integration of a sensor array and a neuromorphic signal processor has been reported as an approach to meeting such requirement. Reviewed presently is the realization of an analog front-end human-machine interfacing system and its application to the acquisition of bio-potential signals. Active-matrix chemical and tactile sensor arrays integrated with biomimetic artificial neural networks based on dual-gate MO TFTs for neuromorphic signal processing are described. Finally, the challenges and prospects of enhanced neuromorphic sensor-perception systems for immersive displays are discussed.