Autonomous navigation system

Abstract

Even after three decades, there are still few real-world strategies that can be used because most techniques in use are based on exaggeratedly optimistic presumptions. When using multiple robots, the agents besides steer clear of obstacles but also prevent collisions with one another, the effects of having unknowledgeable or only vaguely understood maps and states estimate drift increases dramatically [11]. Robot Operating System (ROS) provides a system architecture that enables numerous agencies to independently navigate between user-selected local interest points, the implementation of frequently utilized functionality, message-passing between processes, and administration among the solutions of what an operating system would typically provide. The purpose of this project is to develop low-level device control and hardware abstraction [10]. We suggest a multi-robot global planning pipeline that is hierarchical that prioritizes the agents to prevent collisions between them and compute their trajectories toward the specified target. The generated worldwide pathways are distributed to the agents and used as recommendation data by each agent's internal local planner. The objective that will be done in this research is that we will have to feed the multi-robot simulation to ORB-SLAM3 standalone. After that will introduce covins into the simulation. The project is incomplete due to the issue faced while trying to run ORB-SLAM3 standalone and is unsolved, after uncountable times of trying to solve it.