Proposition of sensor agent for estimation of air-pollution direction and its experimental simulation

Air pollutants appear at various areas on the earth due to economic growth, and harm the health of the residents of those regions. Various countries around those areas are also polluted. Many kinds of air-polluting factors must be grasped, for example, the place of origin, kind of pollutant, its direction and speed. These factors are also grasped quickly over a wide area. Various kinds of chemical sensors have been investigated and developed to detect air pollutants. In this study, metal oxide odor sensors are installed at 12 points in an experimental residence and experimental simulations are carried out. Each sensor with a microprocessor has the ability of an agent, which can autonomously make many advanced decisions and estimations. This sensor system is named sensor agent. The sensor agents can derive pollution direction by cooperation and collaboration with other sensor agents. Derived directions are compared with daily resident behaviors and reliability is examined. Results show that polluting direction is estimated with good reliability and a clue to practical use is achieved. (C) 2000 Published by Elsevier Science S.A.