Comparison of Mamdani and Sugeno fuzzy based data aggregation models for developing smart aquaculture system

In this paper a data aggregation model using Machine Learning (ML) techniques like sugeno and mamdani fuzzy model has been developed and compared based on performance characteristics for smart aquaculture system which improves productivity. Trapezoidal and triangular membership functions are used for defining the input variables such as pH, Dissolved oxygen, Salinity, Organic carbon and Pollution and output variable (Aggregation) respectively. AND operator, implication by truncation method, defuzzificationby weighted average and centroid of gravity,and1875 fuzzy rules are employed for aggregating the input values in sugeno and mamdani fuzzy model. This model divides the aggregated area of each aqua site in the datasets into four categories: poor, good, very good, and excellent, with ratings ranging from 1 to 4 for sugeno and 0 to 3 for mamdani. The result shows sugeno model responds to the full range of aggregation value i.e. [1]-[4] and excecution time for sugeno(6.784 minutes) is less than mamdani(7.457 minutes). Thus data aggregation model using sugeno fuzzy based ML technique improves the reliability and accuracy of the sensed information and also helps the planners and the decision makers for the development of smart aquaculture.