The project is the result of a collaboration among ICoNIoT researchers from four Brazilian universities and is among the nominees for the Best Paper Award at SBRC 2026
VAMOS! is the result of a collaboration between researchers Carnot Braun (State University of Campinas – UNICAMP), Daniel L. Guidoni (Federal University of Ouro Preto – UFOP), Eduardo Cerqueira (Federal University of Pará – UFPA), Joahannes B. D. da Costa (Federal University of São Paulo – UNIFESP), Leandro Villas (UNICAMP), and Allan M. Souza (UNICAMP).
The agent developed by the team, titled VAMOS (Vehicular Agent for Multi-Objective Optimization and Semantics), functions as an intelligent system capable of formulating personalized routes by interpreting the environmental context and the individual priorities of each user. It outperforms traditional navigation systems, which prioritize metric efficiency, such as time and distance, but fail to interpret more complex and context-dependent human intentions.
Unlike conventional navigation systems, this agent uses a Large Language Model (LLM) to suggest strategic stops, such as gas stations or grocery stores, based on continuous learning about the traveler’s profile. The technology’s unique feature lies in its ability to process complex information to optimize routes without the need for overly specific geographic commands.
The project faces the technical challenge of balancing robust processing on external servers with the feasibility of running smaller models directly on mobile devices.
Read the researchers’ paper published in the proceedings of SBRC 2026