The FU Berlin parallel lateration-algorithm simulation and visualization engine

Abstract

We introduce a simulation engine to visually evaluate and compare distance based lateration algorithms and deployments called the FU Berlin Parallel Lateration-Algorithm Simulation and Visualization Engine (LS2). Our engine simulates a scenario which consists of given anchor positions and evaluates all positions of a playing field in parallel, instead of only randomly selected positions. At the end of a simulation run, the user is able to judge the strengths and weaknesses of the algorithm in a picture that displays the spatial position error distribution, a representation of positions in a plane with their errors. Understanding spacial distribution of error is especially valuable, because we observed that it depends on the placement of anchor nodes. This enables the developer to optimize his algorithm or aid him in selecting an algorithm for his application. The simulator's design separates the simulation engine, the lateration algorithms, and the error models. The simulator can be easily extended with additional lateration algorithms and error models. The engine is written in GNU C99 and uses the SSE or AVX vector extensions of modern microprocessors. Thus, it is able to scale fully to all cores. Beside extendability, the main focus is set on execution speed.