The Geo-R2LLM project aims to build multimodal geographic Large Language Models (GeoLLMs) by rethinking the LLMs generation mode with retrieval and reasoning over diverse multimodal external knowledge sources to ground the prediction. These enhanced GeoLLMs will be integrated in a geospatio-temporal artificial intelligence (GeoAI) system prototype and evaluated on a pilot related to context-aware navigation system integrated into smart glasses which is tested in a complex urban environment in Helsinki, Finland. Navigation services are among the most critical and widely adopted location-based services in modern societies, giving the project potential impact beyond academia. The international consortium includes partners from Aalto University (Finland), University of Toulouse 3 (France), University of Leeds (UK), University of the Basque Country (Spain) and Ghent University (Belgium).

2025

2028