A NEW STRATEGY FOR INDOOR PROPAGATION FAST COMPUTATION WITH MR-FDPF ALGORITHM

摘要

The multi-resolution frequency domain ParFlow (MR-FDPF) algorithm has been developed since 2001 for the purpose of Indoor wave propagation simulation. This algorithm derives from a TLM like formalism, and implements a multi-resolution (MR) approach to compute efficiently full coverage of several access points (AP). This approach is well-designed for a wireless LAN planning because the computation load required to process an AP candidate is fast. The heart of the proposed approach holds in the MR-nodes that are obtained by gathering recursively pixels into bricks. This requires a MR binary tree structure which should be adapted to the environment. In this paper, different strategies are evaluated to compute this binary tree. Computation load and memory consumption are evaluated on a real example. It is shown that an appropriate choice of the binary tree may decrease significantly the computation load associated with the method.