Indoor Positioning Based on RSS Fingerprinting in a LTE Network: Method Based on Genetic Algorithms

Abstract

We study the performance of a composite positioning method based on Genetic Algorithms (GA) for determining the position of an indoor user using Radio Signal Strength (RSS) fingerprinting. RSS samples are obtained from field measurements in a test-bed with a dedicated LTE network. Radio maps are described by functions obtained by symbolic regression. The position of a user is then obtained by solving an optimization problem with the functions and instantaneous measured RSS values. The proposed method is benchmarked against a solution implemented with a Neural Network. Position accuracy is evaluated in scenarios with different characteristics. Conclusions are drawn about performance and complexity of the assessed methods.