Rosario Arancio Febbo
Model-Based Design of Sensor Fusion Algorithms: Implementation and Experimental Evaluation on an Embedded Platform for Wearables.
Rel. Alessandro Rizzo. Politecnico di Torino, Corso di laurea magistrale in Mechatronic Engineering (Ingegneria Meccatronica), 2026
|
Preview |
PDF (Tesi_di_laurea)
- Tesi
Licenza: Creative Commons Attribution Non-commercial No Derivatives. Download (9MB) | Preview |
Abstract
The rapid advancement of MEMS technology has enabled widespread adoption of wearable devices for human motion tracking, yet robust 3D attitude estimation remains a significant mathematical and physical challenge. Raw data from Inertial Measurement Units (IMUs) suffer from gyroscope drift bias, high-frequency accelerometer noise, and magnetic distortions. The Attitude and Heading Reference Systems (AHRS) developed to fuse these signals present critical trade-offs between dynamic accuracy and embedded system computational limits. This thesis presents a comparative analysis of four prominent AHRS architectures: the Complementary Filter, Mahony, Madgwick, and the Extended Kalman Filter (EKF). The algorithms are implemented and experimentally evaluated on a resource-constrained Raspberry Pi Pico W microcontroller interfaced with an ICM-20948 MARG sensor.
Through a hybrid software approach combining automatic C-code generation with manual optimization, each filter is evaluated on two primary metrics: architectural efficiency (execution time and memory footprint) and dynamic robustness (disturbance rejection and transient response recovery)
Relatori
Anno Accademico
Tipo di pubblicazione
Numero di pagine
Corso di laurea
Classe di laurea
Aziende collaboratrici
URI
![]() |
Modifica (riservato agli operatori) |
