A comparison of different approaches to the automatic online, data-driven calibration of assisted gearshift settings for a motorcycle is presented. An objective function associated with the component stress and clutch resynchronization time is exploited and optimized during operation using different strategies: from naïve space-filling approaches to learning-based black-box optimization algorithms. The performance of various methods is compared in real-world experiments using metrics related to the experimental convergence rate and the quality of the best found result.
Automatic Learning-Based Calibration of Assisted Motorcycle Gearshift: A Comparative Study
Catenaro Edoardo;Sabug Lorenzo;Panzani Giulio;Ruiz Fredy;Fagiano Lorenzo;Savaresi Sergio Matteo
2025-01-01
Abstract
A comparison of different approaches to the automatic online, data-driven calibration of assisted gearshift settings for a motorcycle is presented. An objective function associated with the component stress and clutch resynchronization time is exploited and optimized during operation using different strategies: from naïve space-filling approaches to learning-based black-box optimization algorithms. The performance of various methods is compared in real-world experiments using metrics related to the experimental convergence rate and the quality of the best found result.File in questo prodotto:
| File | Dimensione | Formato | |
|---|---|---|---|
|
CBW_SMGO_accepted.pdf
Accesso riservato
:
Post-Print (DRAFT o Author’s Accepted Manuscript-AAM)
Dimensione
1.61 MB
Formato
Adobe PDF
|
1.61 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
