AUTHOR=Zafar Muhammad Hamza , Moosavi Syed Kumayl Raza , Sanfilippo Filippo 

TITLE=Enhancing unmanned ground vehicle performance in SAR operations: integrated gesture-control and deep learning framework for optimised victim detection

JOURNAL=Frontiers in Robotics and AI

VOLUME=Volume 11 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/robotics-and-ai/articles/10.3389/frobt.2024.1356345

DOI=10.3389/frobt.2024.1356345

ISSN=2296-9144

ABSTRACT=This paper introduces a novel approach to enhance search and rescue (SAR) operations in disaster environments by integrating cutting-edge technologies into unmanned ground vehicles (UGVs). A prime example is a quadruped robot tailored for SAR tasks. The methodology consists of two key components: gesture-controlled UGV operation and camera-based human detection.The gesture-controlled UGV operation addresses the challenge of precise control in confined spaces, utilising a deep learning (DL) model to accurately interpret and translate hand gestures into real-time control commands. This intuitive interface empowers human operators to guide UGVs through intricate spaces, improving situational awareness and control precision. The second component focuses on human detection using cameras mounted on the UGV, employing an innovative deep learning architecture to identify individuals amidst disaster-induced debris and chaotic surroundings. Notably, the YOLOv8 network is employed, trained, and tested using a Human Dataset specifically curated for disaster scenarios, ensuring the model's adaptability and effectiveness in real-world SAR operations. The integration of these components forms a holistic framework that significantly advances SAR capabilities. Experimental results in simulated disaster scenarios demonstrate the efficacy and real-world viability of the proposed methodology, showcasing the potential of advanced technologies, including DL and camera-based approaches, in evolving UGV technology for disaster response. This research contributes to the ongoing efforts to improve operational outcomes and ultimately save lives in disaster-stricken areas.