Temperature control of semiconductor metal-oxide gas sensors by means of fuzzy logic

The electrical conductivity of metal-oxide gas sensors depends on the gas atmosphere and on the temperature of the sensor. Temperature control systems for gas sensors were developed as a conventional PID-controller. Problems appear in cases of hard disturbances (i.e. rough gas flows) to fix a temperature point that has its origin in the high non-linearity of the controlled system. A new concept based on fuzzy logic was developed to gain better temperature stabilization in atmospheres with hard disturbances. This control system was adapted in the laboratory and tested in a real sensor application. The fuzzy system can optimize the system for different disturbances.