2025/01/01 by Lei Cao, Wenchao Zhang, Dongyan Wei +1
Engineering · Computer Science · #Inertial Sensor and Navigation #Robotic Path Planning Algorithms #Robotics and Sensor-Based Localization
paper · doi:10.1109/tim.2025.3553223
Existing pedestrian dead reckoning (PDR) that relies on inertial sensors is primarily based on the strap-down inertial navigation-based model (INS-PDR) using the IMU mounted on the foot or the step-based model (S-PDR) using the IMU mounted on the foot. These approaches face limitations due to the specialized equipment requirements and lower localization accuracy, which impede their practical application. The methods based on INS-PDR using trunk-mounted IMUs largely adopt the nonholonomic constraint (NHC), which is often used in the inertial navigation of vehicles, but this scheme simplistically assumes zero lateral velocity for pedestrians, which does not accurately reflect their actual motion. To address this issue, this article proposes an inverted pendulum model (IPM) based on the analysis of pedestrian motion during walking. This model is used to calculate 3-D velocity and displacement during pedestrian movement, thereby correcting the cumulative localization errors from inertial navigation and achieving accurate pedestrian location information. Experimental results demonstrate that the proposed method improves localization accuracy by 30% compared to the most advanced existing methods.