Robust markers for visual navigation using Reed-Solomon codes

Abstract

Indoor navigation of unmanned vehicles in GPS denied environment is challenging but a necessity in many real-world applications. Although fully autonomous indoor navigation has been shown to work using simultaneous localization and mapping (SLAM), its accuracy and robustness are inadequate for commercial applications. A semi-autonomous approach is an option for indoor navigation can be achieved using visual markers such as ArUco. The errors caused by motion of robots, visual artifacts due to change in environmental conditions and other occlusion will impact the reliability of visual markers. In this paper, a new robust visual marker based on ArUco with error detection and correction capability is proposed using Reed-Solomon codes. A dictionary of 50 symbols is generated and tested under different conditions with good results in detection and identification.