Advances in Visual Simultaneous Localisation and Mapping Techniques for Autonomous Vehicles: A Review

Area of Science:

• Robotics and Computer Vision

Background:

• Autonomous vehicle (AV) development relies heavily on advanced Information and Communication Technology (ICT) for enhanced vehicular safety.
• Perception and localization are critical functions for AVs, necessitating efficient environmental understanding and 3D mapping capabilities.

Purpose of the Study:

• To present a comprehensive review of Visual Simultaneous Localisation and Mapping (V-SLAM) techniques applied to AV perception and localization.
• To highlight the strengths, limitations, and challenges of V-SLAM deployment in autonomous systems.

Main Methods:

• Overview of Simultaneous Localisation and Mapping (SLAM) principles.
• In-depth review of various V-SLAM schemes and their performance metrics.
• Analysis of current challenges and future research trajectories in V-SLAM for AVs.

Main Results:

• V-SLAM enables AVs to perceive surroundings, map environments, and determine their precise location.
• Identified various V-SLAM techniques with distinct advantages and disadvantages for AV applications.
• Outlined key challenges hindering widespread V-SLAM adoption and potential solutions.

Conclusions:

• V-SLAM is a pivotal technology for advancing AV perception and localization.
• Further research is needed to overcome V-SLAM deployment challenges and enhance its robustness.
• Continued development in V-SLAM will significantly contribute to the future of autonomous driving.