Our work focuses on developing and evaluating enhanced inverse reinforcement learning and imitation learning methods regarding 4D trajectories, although not only for those trajectories. Thus, we follow a supervised leaning approach, and treat historical data provided by “experts” as demonstrations that a machine learning algorithm should exploit to learn the corresponding models of costs/rewards and action policies. Then, learned models can be exploited to plan or predict trajectories.
GEOINFORMATICA, 23 , 2019.
Big Data Analytics for Time Critical Mobility Forecasting: Recent Progress and Research Challenges Proceeding
Increasing Maritime Situation Awareness via Trajectory Detection, Enrichment and Recognition of Events Proceeding
Springer, 10731 , 2018, ISBN: 978-3-319-73520-7.
A Stream Reasoning System for Maritime Monitoring Proceeding