The tracking filter plays a key role in the accurate estimation and prediction of maneuvering a vessel’s position and velocity when attempting to enhance safety by avoiding collision. Therefore, in order to achieve accurate estimation and prediction, many oceangoing vessels are equipped with the Automatic Radar Plotting Aid (ARPA) system. However, the accuracy of prediction depends on the tracking filter’s ability to reduce noise and maintain a stable transient response. The purpose of this paper is to derive the optimal values of the gain parameters used in tracking a High Dynamic Warship. The algorithm employs a α-β-γ filter to provide accurate estimates and updates of the state variables, that is, positions, velocity and acceleration of the high dynamic warship based on previously observed values. In this study, the filtering coefficients α, β and γ are determined from set values of the damping parameter, ξ. Optimization of the damping parameter, ξ, is achieved experimentally by plotting the residual error against different values of the damping parameter to determine the least value of the damping parameter that results in the optimum smoothing coefficients leading to a reduction in the noise corruption effect. Further investigation of the performance of the filter indicates that optimal smoothing coefficients depend on the initial and average velocity of the target.

The maritime industry is expanding at an alarming rate hence there is a perpetual need to improve situation awareness in the maritime environment using new and emerging technology. Tracking is one of the numerous ways of enhancing situation awareness by providing information that may be useful to the operator. The tracking module designed herein comprises determining existing states of high dynamic target warship, state prediction and state compensation due to random noise. This is achieved by first analyzing the process of tracking followed by design of a tracking algorithm that uses α-β-γ tracking filter under a random noise. The algorithm involves initializing the state parameters which include position, velocity, acceleration and the course. This is then followed by state prediction at each time interval. A weighted difference of the observed and predicted state values at the nth observation is added to the predicted state to obtain the smoothed (filtered) state. This estimation is subsequently employed to determine the predicted state in the next radar scan. The filtering coefficients ,  and  are determined from a pre- determined value of the damping parameter, . The smoothed, predicted and the observed positions are used to compute the twice distance root mean square (2drms) error as a measure of the ability of the tracking module to manage the noise to acceptable levels.