IEEE Communications Letters · 2009 · 27 citations · 4 references
ComputingElectrical EngineeringMobile ComputingTime Series AnalysisWireless CommunicationsAd Hoc Mobile NetworksIntelligent Transportation SystemsKalman FilteringMobile Ad Hoc NetworksWireless NetworksWireless Sensor NetworksHuman Mobility PatternsKalman Filter PredictionMobility ModelsMobility PredictionNonlinear System IdentificationNonlinear System ModelingUnscented Kalman FilterUnscented Kalman Filtering
We propose an algorithm to predict the link lifetime in MANETs by the unscented Kalman filter (UKF). The algorithm recursively computes the UKF states, modeled as a nonlinear system, using periodically measured distances as inputs. The UKF states are then utilized to compute the estimates of the remaining link lifetime. Evaluation of the proposed algorithm demonstrates robust performance in computing the MANET link lifetime for various mobility models and in the presence of velocity changes.
4