Navigating robots with precision and efficiency is a fundamental challenge in the field of robotics. Central to this challenge is the critical aspect of odometry, the ability to estimate a robot's motion relative to its environment. In this context, this paper presents an evaluation of the generalizability and effectiveness of a monocular visual odometry method in the context of navigation on a wheeled robot. The study aims to assess TartanVO's performance in real-time motion estimation and its ability to handle various challenges encountered in indoor and outdoor environments. For this purpose, we designed our methodology framework to evaluate the real-time effectiveness of the TartanVO method by utilizing data streams from a robot's on-board sensors. To validate the performance of TartanVO, we compared its pose estimations against ZED pose estimations, analyzing the mean absolute error of the trajectories produced by each method. We collected time-synchronized data from both TartanVO and ZED positional estimate methods, enabling simultaneous position estimation from both methods. Experimental results reveal that the TartanVO method demonstrates impressive real-time efficiency and generalizability, positioning it as a promising solution for odometry in robots operating in various environments. However, challenges were identified, including scale drift and suboptimal pose estimation in lowlight conditions and open outdoor areas when tested with the Jackal robot. These findings underscore the need for further refinement in addressing specific environmental nuances, while acknowledging the overall potential of the method in real-time motion estimation.

The development of autonomous vehicles has prompted an interest in exploring various techniques in navigation. One such technique is simultaneous localization and mapping (SLAM), which enables a vehicle to comprehend its surroundings, build a map of the environment in real time, and locate itself within that map. Although traditional techniques have been used to perform SLAM for a long time, recent advancements have seen the incorporation of neural network techniques into various stages of the SLAM pipeline. This review article provides a focused analysis of the recent developments in neural network techniques for SLAM-based localization of autonomous ground vehicles. In contrast to the previous review studies that covered general navigation and SLAM techniques, this paper specifically addresses the unique challenges and opportunities presented by the integration of neural networks in this context. Existing review studies have highlighted the limitations of conventional visual SLAM, and this article aims to explore the potential of deep learning methods. This article discusses the functions required for localization, and several neural network-based techniques proposed by researchers to carry out such functions. First, it presents a general background of the issue, the relevant review studies that have already been done, and the adopted methodology in this review. Then, it provides a thorough review of the findings regarding localization and odometry. Finally, it presents our analysis of the findings, open research questions in the field, and a conclusion. A semisystematic approach is used to carry out the review.