Enhanced 3D indoor position estimation from smartphone's inbuilt IMU and pressure sensor via hybrid approach
Sharma, Shiva,Kaur, Manjit,Kumar, Naresh
DOI: https://doi.org/10.1007/s10586-024-04473-9
2024-05-14
Cluster Computing
Abstract:This research unveils an improved indoor positioning method by utilizing inbuilt Inertial Measuring Unit (IMU) and pressure sensor of smartphone and enhancing accuracy in Global Positioning System (GPS)-challenged environments. Initially, Ground Truth (GT) trajectories in 3 cases were obtained by measuring different points with measuring-tape. Each GT trajectory represents actual Positional Data (PD) which is compared with the calculated PD trajectories generated from PD 1, 2, 3, 4, and 5 obtained with 5 different methods. Applying Rotational Transformation (RT), it amalgamates accelerometer and gyroscope data to derive earth linear accelerations (ELA) for PD 1. Applying a Fast Fourier Transform (FFT)-based High Pass Filter (HPF) enhances ELA, refining it into PD 2. Incorporating the magnetic field through the Kalman Filter produces PD 3. Linear Regression (LR) models on IMU and pressure sensor data generate PD 4. Through LR model training with past Positional Data 3, IMU, pressure sensor data, and Positional Data 1, the algorithm achieves superior accuracy in PD 5. Demonstrating a substantial 19.49–78.57% improvement over existing literature data, PD 5 exhibits the least average Root Mean Square Error of 0.51 m.
computer science, information systems, theory & methods