Real-Time High-Precision Indoor 3-D Human Localization by Integrating Inertial Sensors and a Barometer

In embedded systems, real-time 3D human localization primarily involves highly complex gait phase detection, high-precision step length estimation, and the particularly challenging task of vertical position estimation. However, due to hardware resource constraints, many advanced human 3D positioning algorithms are difficult to deploy on embedded platforms, while currently deployable lightweight algorithms generally suffer from low accuracy. To address this issue, this paper proposes an embedded real-time single-node gait recognition system that integrates an MCU, inertial measurement unit, barometric pressure sensor, and wireless communication module. To achieve high-precision human 3D positioning, three lightweight designs are introduced: First, a multi-threshold discrimination method is employed to detect the stance and swing phases of pedestrian gait, resolving the inaccuracy in step counting caused by gait segmentation bias. Second, an adaptive stride length estimation model under multimodal conditions is developed based on prior stride estimation and integration methods, thereby enabling high-precision stride estimation. Third, by combining barometric data, a complementary filtering and target error compensation fusion algorithm is proposed to correct altitude estimates in the vertical direction. Experimental results show that the system achieves a cumulative error of less than 3% during 60 meters of horizontal walking and 12 meters of altitude change. This study holds promise for broader applications in 3D human tracking, as well as pedestrian indoor positioning and navigation.