Inference of Wireless Link Performance in MANET

The link performance (loss rate and delay probability) inference is important for evaluating mobile ad hoc network (MANET). The crucial step is MANET topology identification. Since the majority of existing mobility models (MM) may not correctly reflect the true MANET performance, a novel MM based on circle movement is presented. With help of the model, we identify the MANET topology and its live time. MANET internal links performance inference process, which consists of unicast back-to- back packets probes sent from a sender to pairs of receivers, and the performance inference algorithms are presented. NS2 simulations validate the presented model and algorithms.