Abstract: A continuous-time recurrent neural network (CTRNN) is described that exploits the nonlinear dynamics of micro-electro-mechanical system (MEMS) devices to model a neuron in accordance with a neuron rate model that is the basis for dynamic field theory. Each MEMS device in the CTRNN is configured to simulate a neuron population by exploiting the characteristics of bi-stability and hysteresis inherent in certain MEMS device structures. In an embodiment, the MEMS device is a microbeam or cantilevered microbeam device that is excited with an alternating current (AC) voltage at or near an electrical resonance frequency associated with the MEMS device. In another embodiment, the MEMS device is an arched microbeam device that is excited with a direct current voltage and exhibits snap-through behavior due to the physical design of the structure. A CTRNN can be implemented using a number of MEMS devices that are interconnected, the connections associated with varying connection coefficients.

Abstract: A continuous-time recurrent neural network (CTRNN) is described that exploits the nonlinear dynamics of micro-electro-mechanical system (MEMS) devices to model a neuron in accordance with a neuron rate model that is the basis for dynamic field theory. Each MEMS device in the CTRNN is configured to simulate a neuron population by exploiting the characteristics of bi-stability and hysteresis inherent in certain MEMS device structures. In an embodiment, the MEMS device is a microbeam or cantilevered microbeam device that is excited with an alternating current (AC) voltage at or near an electrical resonance frequency associated with the MEMS device. In another embodiment, the MEMS device is an arched microbeam device that is excited with a direct current voltage and exhibits snap-through behavior due to the physical design of the structure. A CTRNN can be implemented using a number of MEMS devices that are interconnected, the connections associated with varying connection coefficients.

Abstract: A continuous-time recurrent neural network (CTRNN) is described that exploits the nonlinear dynamics of micro-electro-mechanical system (MEMS) devices to model a neuron in accordance with a neuron rate model that is the basis for dynamic field theory. Each MEMS device in the CTRNN is configured to simulate a neuron population by exploiting the characteristics of bi-stability and hysteresis inherent in certain MEMS device structures. In an embodiment, the MEMS device is a microbeam or cantilevered microbeam device that is excited with an alternating current (AC) voltage at or near an electrical resonance frequency associated with the MEMS device. In another embodiment, the MEMS device is an arched microbeam device that is excited with a direct current voltage and exhibits snap-through behavior due to the physical design of the structure. A CTRNN can be implemented using a number of MEMS devices that are interconnected, the connections associated with varying connection coefficients.

Abstract: A micro-electro-mechanical-systems (MEMS) device is disclosed that is configured to operate as a reservoir computer including performing sensing and computing co-locally. The MEMS device includes circuitry for: generating a modulated input signal based on an input signal; generating a MEMS deflection signal based on the modulated input signal and a time-delayed MEMS deflection signal; sampling the MEMS deflection signal N times during a time internal T to generate a MEMS deflection matrix, wherein MEMS deflection matrix has a size M×N, wherein N corresponds to a number of virtual nodes of the reservoir computer and M is a number of time steps of time interval T; receiving a trained weight matrix, wherein the trained weight matrix is trained by linear regression; and multiplying the MEMS deflection matrix by the trained weight matrix to generate an output signal that classifies the input signal.

Abstract: Systems and methods to amplify the response of a MEMS micro-oscillator by driving the MEMS device at its electrical and mechanical resonance frequencies, simultaneously. This enhances the MEMS mechanical sensitivity to electrical excitation and increases the voltage across the MEMS capacitor. Moreover, using a combination of two input signals at different frequencies (beat signal) may be used to achieve double resonance in any MEMS device, even if its natural frequency is far from its electrical resonance.

Abstract: A continuous-time recurrent neural network (CTRNN) is described that exploits the nonlinear dynamics of micro-electro-mechanical system (MEMS) devices to model a neuron in accordance with a neuron rate model that is the basis for dynamic field theory. Each MEMS device in the CTRNN is configured to simulate a neuron population by exploiting the characteristics of bi-stability and hysteresis inherent in certain MEMS device structures. In an embodiment, the MEMS device is a microbeam or cantilevered microbeam device that is excited with an alternating current (AC) voltage at or near an electrical resonance frequency associated with the MEMS device. In another embodiment, the MEMS device is an arched microbeam device that is excited with a direct current voltage and exhibits snap-through behavior due to the physical design of the structure. A CTRNN can be implemented using a number of MEMS devices that are interconnected, the connections associated with varying connection coefficients.

Abstract: Systems and methods to amplify the response of a MEMS micro-oscillator by driving the MEMS device at its electrical and mechanical resonance frequencies, simultaneously. This enhances the MEMS mechanical sensitivity to electrical excitation and increases the voltage across the MEMS capacitor. Moreover, using a combination of two input signals at different frequencies (beat signal) may be used to achieve double resonance in any MEMS device, even if its natural frequency is far from its electrical resonance.

Abstract: A controller device and a method for controlling a system that utilizes an adaptive mechanism to self-learn the system characteristics and incorporates this adaptive self-learning ability to predict a control parameter correctly to provide precise control of a system component.