Vehicle level health and usage monitoring of electric unicycles using end user telemetry

Abstract Electric unicycles (EUCs) have evolved into high-performance personal vehicles capable of daily commuting at substantial speeds and energies, yet they remain largely outside the scope of established vehicle health and usage monitoring practices. We interpret vehicle-level health as the ability of the combined battery and power stage to sustain performance with acceptable voltage and thermal margins under representative operating conditions — a practical definition tailored to the limited telemetry available on commercially deployed EUCs. Sudden loss of propulsion, often suspected to be linked to electrical or thermal limitations, can lead to severe rider injuries, while existing mobile applications focus mainly on real-time dashboards and trip logging without providing long-term, vehicle-level health assessment. This paper introduces EUC_SOH, a vehicle health and usage monitoring system for electric unicycles that operates on ride logs collected over several thousands of kilometers. EUC_SOH derives empirical, physically interpretable indicators from pack-level telemetry and usage data, including a telemetry-derived equivalent resistance proxy, battery-dominated resistance normalized to 25 $$^{\circ }$$ C, power MOSFET resistance at operating temperature, voltage sag metrics, and current and temperature-based stress indicators. An optional stylized ageing simulator and dedicated anomaly and degradation detection modules, including CUSUM-like change detection, are used to characterize both slow degradation trends and rare out-of-family events at the vehicle level under realistic data constraints. We evaluate EUC_SOH on real-world datasets from multiple EUCs, complemented by simulated ageing and fault scenarios. Results from the real-world logs show stable or changing intra-wheel indicator trajectories over several thousand kilometres, while the controlled synthetic scenarios demonstrate detection of major injected deviations over tens of thousands of simulated kilometres. We discuss how such vehicle-level health indicators, built from end-user-accessible logs, could support safer operation, early maintenance decisions and more systematic consideration of electric unicycles as serious vehicles within the broader framework of health and usage monitoring systems.