Deliberative control for satellite-guided water quality monitoring

This paper addresses the issue of efficient monitoring of Lake Winnipeg water quality by employing the power of computational intelligence methods in the processing of multi-spectral remote sensing data. The large size of Lake Winnipeg (the 10th largest lake in the world) and its susceptibility to algal blooms makes satellite technologies indispensable in monitoring the quality of the lake's water. The remote sensing data have to be complemented by in-situ measurements due to the requirements for the calibration of satellite imagery, for precise local measurements, as well as because of the variations in water conditions. A method for the path planning of a ship equipped with water sample acquisition and processing facilities is presented. Given the complexity of the planning task (acquisition of different types of samples of different informational values, use of ancillary environment data, dependence on the results of satellite data processing, etc.), an inclusion of the deliberative level in the ship trajectory planning and control scheme is postulated. A deliberative control architecture is proposed which features a multi-model classification/regression system for the determination and forecasting of spatial distribution of water pollutants, in particular chlorophyll-a, and a cost optimizing path planner. A fuzzy system which handles different control strategies depending on the surrounding environment supervises the reactive level operational control.