Deletion of data breaks using nonlinear smoothing algorithms

The Range Accuracy and Monitoring Program (RAMP) was instituted to quantify instrumentation accuracy and to determine any necessary improvements to the existing tracking systems used at the Atlantic Fleet Weapons Training Facility (AFWTF). The RAMP is conducted by the AFWTF Software Contractor. RAMP procedures collect range, azimuth, and elevation data from various tracking systems. Such data are used to generate performance evaluation reports on tracking systems. These reports may lead to system improvements whose goal is to provide the AFWTF customer with higher quality data. Data obtained through the RAMP procedures require smoothing to delete data breaks, such as peaks and outliers. Currently, data breaks are removed via visual inspection and manual deletion. Such processes require an average of 30 minutes per data set. This research project studied the use of nonlinear filtering techniques to automate the elimination of data breaks. MATLAB engineering application was used as the development environment for the seven smoothing algorithms evaluated. These algorithms were implemented and tested using two sets of range data from different RAMP procedures. The evaluation process showed that the Wilcoxon and the Impulse detection filters were the best in terms of algorithm complexity and output response of the filter. This conclusion is based on comparisons of the filtered data versus the data corrected via the current manual method.