|
Abstract:
|
Wireless signal propagation in indoor environments is generally affected by various factors such as multi-path fading, temperature and humidity. It is thus challenging to use the simple metric of Received Signal Strength (RSS) for indoor localization. Prior work often requires manual profiling of each indoor environment to achieve good accuracy. We, instead, propose a simple generic localization algorithm that can be applied in any indoor environment without profiling. Our algorithm makes use of a simple signal propagation model and accepts a broad characterization of the environmental effect. The algorithm's simplification results in some accuracy loss compared to profiling based algorithms. We, however, believe that our algorithm's zero-profiling requirement would make it favorable in many scenarios, especially when extensive profiling is not affordable or it is favored to trade a little accuracy for saved human effort. Evaluations also show that deployment of additional landmarks can be combined with our algorithm to compensate the accuracy loss, thus achieving good accuracy without profiling.
|
