Research Article
An Improved Hidden Markov Model for Indoor Positioning
@INPROCEEDINGS{10.1007/978-3-031-34790-0_31, author={Xingyu Ren and Di He and Xuyu Gao and Zhicheng Zhou and Chih-Chun Ho}, title={An Improved Hidden Markov Model for Indoor Positioning}, proceedings={Communications and Networking. 17th EAI International Conference, Chinacom 2022, Virtual Event, November 19-20, 2022, Proceedings}, proceedings_a={CHINACOM}, year={2023}, month={6}, keywords={Indoor Positioning Hidden Markov Model IMU Machine Learning}, doi={10.1007/978-3-031-34790-0_31} }- Xingyu Ren
Di He
Xuyu Gao
Zhicheng Zhou
Chih-Chun Ho
Year: 2023
An Improved Hidden Markov Model for Indoor Positioning
CHINACOM
Springer
DOI: 10.1007/978-3-031-34790-0_31
Abstract
This paper proposes an indoor positioning method combines machine learning, IMU (Inertial Measurement Unit) and an improved HMM (Hidden Markov Model). HMM is the base framework, the latent states correspond to the location grid, which uses two methods to divide the area into grids with different granularity. The fine-grained grids are used to compute transition probability, and the coarse-grained ones for emission probability. IMU data can help to adjust transition probability between fine-grained grids, and some machine learning methods to estimate the emission probability in each coarse-grained grid. And for the particularity of the model, there’s an improved Viterbi algorithm proposed to calculate the most likely path, which is a more robust version compared with the original one.
