Fuzzy measures for sensor data fusion in industrial recycling

During recent years it has become more and more obvious that we need to take care of products when they are worn out if we are to be more environmentally responsible. To make this possible much of the related work has to be performed in an automized way. When a disassembly process is automized a number of sensors is needed. These sensors gather a large amount of information which must be fused in an effective way. To this end we need a method that is fast, flexible, robust and easy to adapt to all kind of sensors. The first attribute (fastness) can be obtained by using a distributed sensor data processing architecture with local processors which pre-process data close to the sensors. The other attributes are obtained by using a fusion algorithm based on a novel fuzzy operator. In this article, this fuzzy operator is described and its properties investigated. A simulation is performed, applying the fuzzy operator to a sensor-based classification by size of electrical motors aimed for disassembly. The simulation clearly demonstrates the superiority of our fuzzy operator when compared to common AND and OR operators.