Kalman Filter Modeling of the Speed River Quality

The extended Kalman filter is applied to the results of two sampling surveys conducted on the Speed River, Ontario. Magnitudes of carbonaceous and nitrogenous BOD components for these surveys are determined. Parameter magnitudes are estimated using Rosenbrock's parameter estimation technique, and the measurement error convariance matrix for each measurement vector is established. Filtered estimates of carbonaceous oxygen demand, nitrogenous oxygen demand, and dissolved oxygen deficit concentrations are obtained. In addition, the carbonaceous and nitrogenous BOD reaction rates, the carbonaceous BOD sedimentation rate, and the maximum daily rate of photosynthetic oxygen production are estimated. The utility of the extended Kalman filter to use all available information coming from both the system model of the stream processes and measurements of the state variables to provide approximate minimum variance estimates of the state and confidence limits of these estimates is shown.