Combination of Localization Techniques for Mobile Sensor Network for Precise Localization

Abstract

A precise localization is required in order to maximize the usage of Mobile Sensor Network.
As well, mobile robots also need a precise localization mechanism for the same reason.
In this paper, we showed a combination of various localization mechanisms for precise localization in three different levels.
Localization can be classified in three big categories: wide area and long distance localization with low accuracy, medium area and distance localization with medium accuracy, and small area short distance localization with high accuracy.
In order to present localization methods, traditional map building technologies such as grid maps or topological maps can be used.
We implemented mobile sensor vehicles and composed mobile sensor network with three levels of localization techniques.
Each mobile sensor vehicles act as a mobile sensor node with the facilities such as autonomous driving, obstacle detection and avoidance, map building, communication via wireless network, image processing, extensibility of multiple heterogeneous sensors, and so on.
For localization, each mobile sensor vehicle has abilities of the location awareness by mobility trajectory based localization, RSSI based localization and computer vision based localization.
With this set of mobile sensor network, we have the possibility to demonstrate various localization mechanisms and their effectiveness.
In this paper, the result of computer vision based localization, sensor mobility trail based localization and RSSI based localization will be presented.