In this paper, we propose a novel general technique, based on renewal
theory, for analyzing mobility models in ad hoc networks. Our
technique enables an accurate derivation of the steady state
distribution functions for node movement parameters such as distance
and speed. We first apply our technique to the random waypoint model
and provide alternative proofs for previous claims about the
discrepancy between the steady state average speed and the average
speed associated with the simulated distribution [1]. Our main
contribution is a new methodology for simulating mobility which
guarantees steady state for node movement distributions from the start
of the simulation. Our methodology enables the correct and efficient
simulation of a desired steady state distribution, and can be
implemented in a manner transparent to the user. We support our claims
through both formal proofs as well as extensive simulations.