Hyper-reductant snake robot for military search and rescue operations

A search and rescue effort is critical during an emergency, such as hostages trapped within a building. It is critical to assess the prisoners’ status in advance so that proper safeguards may be taken. Most traditional search and rescue robots fail to enter buildings through small areas due to their incapacity to adapt to the intricate geometry of entry infrastructure such as pipes. The redundant structure of snake robots gives for more flexibility in such scenarios. As a result, these snake robots can assist the military with search and rescue tasks in small structures. This study focuses on developing snake robots that can be delivered into buildings via tiny pipes and recognizing humans using machine learning frameworks such as deep convolutional neural networks. To put the notion to the test, we built a snake robot with 12 redundant degrees of freedom. After that, the simulation was done in ROS Gazebo and rviz. Human models and structures were also used to imitate the real-world environment. As a result, the snake robot’s forward kinematics were refined in order to better align the surveillance camera in order to speed up the detection process. Furthermore, using deep convolutional neural networks, the snake robot detected people with about 98% accuracy.