By coupling State Space Models (SSMs) with Spiking Neural Networks (SNNs), the potential of neuromorphic solutions for positioning and tracking can be successfully explored. A spiking SSM based on the Legendre Memory Unit (LMU) is shown to effectively predict the state of an Extended Kalman Filter (EKF) using Inertial Measurement Unit (IMU) and Global Positioning System (GPS) inputs, with memory reduction up to 96% compared to Deep Learning (DL) solutions. Additionally, such SNN-based architecture enables GPS error correction through data fusion of IMUs, wheel encoders, and gyroscopes, highlighting the efficiency of spike-based solutions over traditional models.
Positioning and tracking enhancement through a spiking Legendre memory unit / Tilocca, Salvatore; Fra, Vittorio; Pignata, Andrea; Macii, Enrico; Urgese, Gianvito. - In: IFAC PAPERSONLINE. - ISSN 2405-8971. - ELETTRONICO. - IFAC-PapersOnline:(In corso di stampa). (Intervento presentato al convegno 7th IFAC Conference on Intelligent Control and Automation Sciences (ICONS 2025) tenutosi a Padova nel 15-18 September 2025).
Positioning and tracking enhancement through a spiking Legendre memory unit
Salvatore Tilocca;Vittorio Fra;Andrea Pignata;Enrico Macii;Gianvito Urgese
In corso di stampa
Abstract
By coupling State Space Models (SSMs) with Spiking Neural Networks (SNNs), the potential of neuromorphic solutions for positioning and tracking can be successfully explored. A spiking SSM based on the Legendre Memory Unit (LMU) is shown to effectively predict the state of an Extended Kalman Filter (EKF) using Inertial Measurement Unit (IMU) and Global Positioning System (GPS) inputs, with memory reduction up to 96% compared to Deep Learning (DL) solutions. Additionally, such SNN-based architecture enables GPS error correction through data fusion of IMUs, wheel encoders, and gyroscopes, highlighting the efficiency of spike-based solutions over traditional models.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/11583/3003699
Attenzione
Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo