FPGA based sensor linearization application

Sensors are fundamental measurement elements used in wide variety of applications. A sensor is a converter that measures a physical quantity and converts it into a signal which can be read by an electrical instrument. Since many kinds of sensors have non-linear input-output behaviour, to be able to calculate the measured physical quantity by using the sensor data, special electrical circuits and algorithms are required. In the scope of this thesis, the advantages of the Programmable Gain Amplifier (PGA) and Artifical Neural Networks (ANN) are combined and realized experimental applications on an FPGA platform to linearize a non-linear sensor behaviour. The Negative Temperature Coefficient (NTC) type temperature sensor is choosen. The simulation of the suggested linearization method is realized in the software environment and afterwards applied on an FPGA experiment set. The resulting performance of the suggested method is compared with other methods in respect to mean squared error, process speed, memory usage and cost.KEYWORDS: sensor linearization, ANN, PGA, FPGA, NTC, MLP