Thesis: Multimodal Spatiotemporal Intelligence for Fleet-Level Intelligence
Core
Develop a shared spatiotemporal representation of sensor observations and vehicle signals to maintain a dynamic model of site conditions for fleet-level decision support.
Role type
Master's thesis student (multimodal AI for autonomous fleets)
Builds
Dynamic site-condition models and evidence-grounded recommendations for fleet operations
Domain
Autonomous transport, machine learning, multimodal sensor fusion
Deliverable
research
Required skills
Python, deep-learning frameworks (PyTorch/TensorFlow), multimodal data handling, spatiotemporal modelling, machine-learning experiments
Preferred skills
Autonomous systems, machine perception, real-world sensor data analysis, fleet-level intelligence
Technologies
PyTorch, TensorFlow, GPU compute, CAN signals, LiDAR, RADAR, camera data
Responsibilities
Develop and compare timestamp-based single-modality and multimodal methods for modelling changing site conditions; Combine information from vehicle sensors, CAN signals, maps, and operational events; Measure detection quality, warning lead time, and robustness to missing sensors; Analyze results and document findings in a scientific report.
Seniority
Master's level research project
