Error-predictable adaptive dead reckoning algorithm

A dead reckoning algorithm was presented which is suitable for first-person shooting games. Avatars' moving directions change frequently and their traces consist of continuous broken lines. According to the features, first order derivative polynomial and linear convergence algorithms were selected for position extrapolation and convergence respectively. Avatar's moving state was used to predict extrapolation error and adjust convergence procedure dynamically. Extrapolation policy was changed with the distance between avatar and its observer in order to reduce state update packets. Experiment results show that the position extrapolation algorithm can reduce state update packets and deviation of reckoning trace in first-person shooting game.