The rapid proliferation of tracking sensors, ranging from self-reporting systems to GPS, cameras, radars, and Earth observation, has led to unprecedented volumes of mobility data. However, these datasets are often fragmented, incomplete, and underutilized. Particularly, urban mobility and maritime surveillance generate massive, multi-modal datasets requiring advanced analytics.

Subsequent Knowledge Discovery and Extraction (KDE) hinges entirely on the quality and integrity of the underlying trajectories, demanding innovative preprocessing and fusion methods. Therefore, developing robust, quality-aware analytical pipelines is paramount for generating trustworthy insights from such data streams. These advancements are crucial for applications such as sustainable urban planning and maritime situational awareness, which are vital components for achieving climate-neutral cities and effective ocean protection.

This workshop, Multi-Sensor Trajectory Knowledge Discovery and Extraction (MuseKDE), aims to bring together researchers and practitioners from academia and industry to discuss novel methodologies for knowledge discovery and extraction from multi-sensor trajectory data. Contributions should employ recent advances in deep learning, graph neural networks, attention-based models and other data-driven techniques to open new possibilities for trajectory representation and inference with fragmented or underutilized data. Particular attention will be placed on addressing the challenges of fusing data from heterogeneous sources and on cross-domain representation models.

MuseKDE seeks submissions of novel, high-quality research papers (maximum 6 pages long) that represent original work not previously published or currently under review elsewhere. Additionally, the workshop welcomes short and vision papers (2 – 4 pages) to promote discussion of ongoing work and future directions. Specifically, these papers may report on preliminary results from ongoing research or propose conceptual frameworks for upcoming research efforts. Finally, the workshop will also put emphasis on FAIR principles and open datasets to create benchmark datasets and open-source tools.

The submission of tested solutions is highly encouraged. To evaluate their proposed research, participants can take advantage of some open mobility datasets provided by organizers, namely:

• MMDEC: Multimodal Maritime Dataset on the English Channel includes vessel tracks covering the western Celtic Sea, the English Channel, and a part of the North Sea over a three-month interval, complemented by several sensor data (including satellite imagery, weather data records and others). Available at https://zenodo.org/records/17491518.
• AegeaNET Syros AIS Dataset for Vessel Traffic Monitoring a collection of vessel activity of an extended area around the Cyclades in the Aegean Sea, Greece, spanning a continuous three-month period, encompassing peak maritime-traffic months in the region. Available at https://zenodo.org/records/18267789.