Algorithms for High-Performance Networking in the Presence of Obstacles

Abstract

This work develops novel algorithms for high-performance networking in the presence of obstacles based on a method for communicating via ultrasonic rays reflected at the obstacles. The rays are curves determined by the variable speed of sound and initial conditions and we develop ultrasonic ray models based on a system of differential equations. We present new parallel algorithms and software for shape and trajectory reconstruction of moving obstacles and show how the reconstructed reflection point of a ray at an obstacle is a natural router for messages between the ray's transmitter and receiver and discuss the advantages of the new architecture. We discuss how the new algorithms and software improve the performance and properties of the network architecture.