A Review of Learning-Based SLAM Approaches of Autonomous Unmanned Vehicles (AUV)

Artificial Intelligence (AI) is becoming a hot topic in the field of robotic research last few decades. Autonomous Unmanned Vehicles (AUV) are being used for different tasks like rescue, search, monitoring, aerial operations as well underwater even AUV can aid where human reachability impossible. Localization, tracking, and mapping fundamentals an autonomous system. The main problem AUV, which attracts researchers, simultaneous localization mapping, no external positioning source available GPS. There many proposed techniques algorithms be to solve this GPS (Global Positioning System), Motion Capture System (MCS), Visual System, etc. with limitations. Some probabilistic solutions Graph SLAM, EKF based SALM, Fast SLAM also problem. non-linear model but has issues (like inconsistency) when map becomes large complex. It does not work non-Gaussian distribution. algorithm provide high speed computation good accuracy during resampling processes particle depletion degeneracy. On other hand, deal complex maps process number landmarks accurately it perform much better than SLAM.