Objective
UGV Reconnaissance / Surveillance in Zero-Visibility Conditions
This project is focused UGV Reconnaissance / Surveillance in Zero-Visibility Conditions through the integration of various advanced technologies. A key element of the project is object detection, which uses AI on the edge, using both RGB and thermal cameras. The project also incorporates mm-Wave radar technology, featuring the TI-AWR1642 board, equipped with a peak detection algorithm. Another critical aspect is voice gesture-controlled navigation, powered by the VOSK model for voice recognition in combination with Microsoft HoloLens. To enhance coordination, inter-robot communication is enabled through the use of visual cues, GPS, MongoDB, and MQTT. Furthermore, the project employs conversational visual question answering (VQA) using CARDS developed and fine-tuned with LLaVA1.5, VOSK, and GTTS, facilitating interaction between a robot and a remote operator.
Graces Quarters 2024 in collaboration with ARL,UMCP,BSU
