Includes supplementary material: sn.pub/extras

Introduction.- Overview of Probability and Statistics.- Primer on Matrices and Determinants.- Robot Kinematics.- Computer Vision.- Machine Learning.- Bullet Kinematics.- Experimental Results.- Future Direction.

This SpringerBrief reveals the latest techniques in computer vision and machine learning on robots that are designed as accurate and efficient military snipers. Militaries around the world are investigating this technology to simplify the time, cost and safety measures necessary for training human snipers. These robots are developed by combining crucial aspects of computer science research areas including image processing, robotic kinematics and learning algorithms. The authors explain how a new humanoid robot, the iCub, uses high-speed cameras and computer vision algorithms to track the object that has been classified as a target. The robot adjusts its arm and the gun muzzle for maximum accuracy, due to a neural model that includes the parameters of its joint angles, the velocity of the bullet and the approximate distance of the target. A thorough literature review provides helpful context for the experiments. Of practical interest to military forces around the world, this brief is designed for professionals and researchers working in military robotics. It will also be useful for advanced level computer science students focused on computer vision, AI and machine learning issues.