A Method of Using Personal Habits for Path Prediction in Network Games

Abstract

In almost all of multiplayer network games, dead- reckoning (DR) is used to predict movement of game players. According to received DR vectors, game clients can predict `future' movement of other players. However, DR does not work well under bad network conditions. In this paper, we propose a solution to achieve much more prediction accuracy. Personal habits influence path prediction of players. However, our method introduces extra computation burden and searching delay. So, a hybrid method, which is a combination of DR and personal habits, is introduced. We use a 2D tank game for experiment and compare the results of our solution with those of traditional methods. To obtain habitual movement, we carry out 30 minutes playing observation on each participator. Simulation shows that our method achieves significant improvement in path prediction.