A Predictive Low Latency Handover Scheme for Mobile IP

Many wireless protocol simulation studies are performed under the assumptions that (i) nodes travel in a random fashion in unobstructed environments, and that (ii) radio wave propagation modeling is based on the freespace principle. We propose an urban mobility model, where mobile nodes traverse city streets using a variety of movement modes and where the urban environment is used as an input in the radio wave propagation modeling process. We present a low latency Mobile IP handover scheme, deployed in these urban settings. The network learns the mobility patterns of the mobile nodes and can then make predictions on their next handover. By pre-emptively setting up tunnels and estimating the tunnel activation times, very good handover conditions can be obtained. We present results drawn from our OPNET simulation model.