Dead Reckoning-Based Update Scheduling against Message Loss for Improving Consistency in DVEs

Maintaining a consistent presentation of the virtual world among participants is a fundamental problem in the Distributed Virtual Environment (DVE). The problem is exacerbated due to the limited network bandwidth and errorprone transmission. This paper investigates Dead Reckoning (DR) update scheduling to improve consistency in the DVE against message loss. Using the metric of Time-Space Inconsistency (TSI), which combines the spatial magnitude and temporal duration of inconsistency, we analytically derive the impact of message loss on TSI when using a DR-based update mechanism. To improve consistency against message loss, a naive update scheduling algorithm is first proposed, in which the expected spatial difference is calculated by taking message loss into account. In order to reduce the TSI to the case without transmission failures, a compensation update scheduling algorithm is further proposed by reducing the DR threshold according to the message loss rate. Using these algorithms, a budget-based mechanism is developed to meet the network bandwidth constraint. We show through experiments using a racing car game that the budget-based mechanism using the compensation update scheduling algorithm makes the best use of available network bandwidth to reduce the inconsistency and its impact on the participants. In addition, it ensures fairness among participants in spite of widely varying message loss rates.