The human ear distinguishes sounds with astonishing sensitivity. It is able to resolve frequencies differing by 0.2% over 130 decibels of intensity. To achieve this feat the ear combines different strategies which we will implement in sensors of unprecedented sensitivity. The ear uses the power of Brownian noise in the cochlear fluid to enhance the detection of weak periodic signals. Each frequency is analyzed by specialized auditory hair cells which have different lengths and use a property called negative stiffness to enhance detection sensitivity. The incoming fellow has recently proposed a theoretical scheme for amplification by negative stiffness which he has demonstrated using a bench top experiment. Through a series of recent papers, the host group has conceived and studied a biomimetic neuron which sums and thresholds electrical pulses. In particular, the neuron demonstrates the amplification of useful signals by random noise using a property known as stochastic resonance. The present project will integrate artificial hair cells with semiconductor neurons to deliver highly sensitive, low power, scalable mechano-electrical transducers. We will study the amplification properties of the hair-neuron system. We will then integrate artificial hairs on neurons to make an artificial cochlea of microscopic size. We will fabricate hairs of different length/diameter aspect ratio to detect specific audio frequencies and code the sound intensity in the firing rate of the neuron. This fellowship will thus prepare the next generation of hearing aids, hydrophones and voice recognition systems by incorporating advances in non-linear physics and nanoscience in our cochlear sensor. The sensor will improve the range of sub-marine detection and lead to smarter, smaller hearing aids thanks to the monolithic integration of the receiver with the neural network. The fellowship will establish a EU-Korea pole of collaboration in this multidisciplinary emergent field.
|Effective start/end date||1/08/10 → 31/07/11|