Department of Informatics, University of Minho, Portugal

Title of the Invited Talk: Collaborative Multi-Agent System for Multimodal Object Detection

Abstract: I will present a collaborative multi-agent system for multimodal object detection on edge devices, integrating video, sound and radar data to enhance detection accuracy and reliability. The proposed system leverages Docker containers for efficient model deployment and homographic transformations for precise alignment of multimodal data. By fusing video, sound and radar inputs, the architecture overcomes the limitations of single-modality approaches, particularly in challenging scenarios like occlusions and non-line-of-sight conditions. The system incorporates continuous learning to dynamically adapt to changing environments, making it scalable and robust. This solution is well-suited for applications in smart cities and autonomous systems, demonstrating significant potential for real-time object detection and situational awareness.