Intelligent Robotic Rescue Assistant

In this comprehensive project, we undertook the development of several advanced robotics systems with diverse capabilities. One of these robots was specifically engineered for aquatic environments, equipped with specialized sensors and guided by sophisticated pathfinding algorithms. Its primary function was to autonomously navigate through water bodies, such as the sea, and detect individuals in distress. Utilizing a heat sensor, this robot could identify signs of drowning and promptly reach the person in need to provide assistance.

Additionally, another significant aspect of our project involved integrating a high-resolution 4K camera with a Jetson device. Leveraging the computational power of CUDA cores, we deployed a custom-trained deep learning model capable of real-time human detection. This system was designed not only to recognize complete human figures but also to identify individuals based on partial body features, making it invaluable in scenarios where traditional visibility is limited, such as under earthquake rubble. The captured images were efficiently compressed using ffmpeg’s H264 technique and transmitted to designated receivers via MicroTik infrastructure, ensuring swift and effective communication for immediate response and support operations. This multifaceted approach showcased our team’s proficiency in robotics, deep learning, and real-time data transmission technologies, emphasizing innovation and practical application in critical situations.