Interest scheme

Abstract

Most multiplayer network games use dead-reckoning vectors (DR) to predict movement of an entity controlled by participating game players. Generally, DR contains the current position and velocity of the entity. To get more accurate prediction, time-stamp and acceleration are also included. In this paper, we propose a solution named Interest Scheme (IS) to achieve much more prediction accuracy when network latency is unsteady and package loss is frequent. IS assumes that path prediction of a given player has relations to its nearby objects and players in game space. In IS, the surroundings of a given entity is taken into account in path prediction. Moreover, we consider that different prediction method should be used for different network latency. So, a hybrid method, which is a combination of IS and traditional method, is introduced. We use a 2D tank game for experiment and compare the results of our solution with those of traditional methods. Simulation shows that our method achieves significant improvement in path prediction.