This thesis aims to develop predictive models capable of anticipating customer driving and charging behaviour based on large-scale fleet telematics data. By leveraging machine learning techniques and statistical analysis, the research will extract meaningful patterns from historical usage data, including trip frequency, parking duration, charging cycles, environmental conditions, and driving styles. The objective is to enable predictive battery charge management strategies that optimize energy efficiency, minimize battery degradation, and enhance user convenience while ensuring vehicle readiness. Particular attention will be paid to model robustness and integration into real-time energy management systems.
The thesis will be carried out in cooperation with a major automotive manufacturer, providing access to anonymized fleet datasets.
