ENHANCED COMPUTATIONAL METHOD TO PREDICT FIRE PARAMETERS IN AN ARTILLERY FIRE

This paper proposes an enhanced computational technique to compute fire parameters such as firing angle and azimuth, during fire prediction in an artillery fire. The current fire prediction methods, are in-general based on two approaches wherein either a firing table containing several look-up tables is used in conjunction with simple mathematical equations to manually compute the fire parameters for non-standard conditions of fire or alternatively computing the fire parameters iteratively using a trajectory-based fire prediction software. Trajectory-based fire prediction algorithms are much accurate and usually begins with an initial guess firing angle, obtained by interpolating the range-elevation dataset, computed for standard conditions. A step further, to reduce the number of iterations, the final firing angles computed using a digital firing table is more often used as guess firing angle, wherein the corrections to range and line for every non-standard parameter is computed from a set of polynomial equations obtained by curve fitting the dataset. This paper proposes the use of machine learning based algorithms which are much robust and accurate as compared to the polynomial functions obtained by traditional curve fit techniques.