Reliability Assessment of MEMS IMUs based on Uncertainty Propagation in Positioning Algorithms during Accelerated Life Test

The widespread use of MEMS-based Inertial Measurement Units (IMUs) in mission-critical applications, such as aerospace, robotics, and autonomous systems, demands rigorous evaluation of their reliability. In particular, aging effects can significantly degrade the metrological performances of the sensor, impacting also the accuracy of positioning algorithms. This paper presents a methodology to assess the influence of IMU aging on orientation estimation, focusing on two widely used algorithms: the Complementary Filter and the Madgwick Attitude and Heading Reference System (AHRS). A six-month Accelerated Life Testing (ALT) campaign was conducted on 25 IMUs to emulate long-term degradation. Results are then expanded into a realistic life cycle using the Arrhenius degradation model. Drift and uncertainty propagation were analyzed in compliance with the ISO Guide to the Expression of Uncertainty in Measurement (GUM). For the Complementary Filter, analytical modeling using the Law of Propagation of Uncertainty (LPU) was performed, while a novel observation-window approach was proposed to evaluate uncertainty in the recursive Madgwick AHRS. Results showed non-linear performance degradation, with a critical mid-life phase characterized by increased drift and uncertainty. Monte Carlo simulations confirmed the validity of the proposed methods. The framework offers a practical tool for monitoring sensor aging and metrological reliability in long-duration IMU-based systems.