Recently a novel signal processing theory related with noise has grown and proven. Certain complex systems can improve performance with added optimal noise that classical theory cannot explain. Their behavior may be represented by a simplified scheme that combines both a deterministic and stochastic source. To that end, we are using noise in remote temperature sensing system to enhance their function without altering the system.
A new investigation of noise added scheme has been realized by an embedded heater for CMOS compatible thermoelectric infrared sensor. The design and fabrication of thermopile sensors are realized by using 1.2μm CMOS IC technology combined with a subsequent anisotropic front-side etching. We firstly develop an active thermopile with a heater embedded which is easily and naturally driven by a noise generation circuit. The stochastic resonance theory can be realized as a reduction in threshold of temperature detection. We have shown the possibility of improving the performance of remote temperature sensing system in the presence of noise. The strategy depends on the application. Stochastic resonance can reduce threshold detection resolution and greatly improve the temperature detection limit with a low cost scheme without using higher resolution ADC.