AUTONAV: A Toolfor Autonomous Navigation of Robots
Mir Md Sajid Sarwar, Sudip Samanta, Rajarshi Ray
TL;DR
AUTONAV addresses autonomous navigation by unifying mapping, localization, and path planning within a modular framework that supports rapid comparative evaluation. It leverages Cartographer for 2D SLAM and Amcl for localization, and formulates motion planning as a constraint-satisfaction problem solved by the z3 SMT solver, producing a sequence of waypoints $(x_t, y_t)$ with an upper bound $M$ on the number of steps. The main contributions are the integrated architecture, a detailed SMT-based planning formulation including obstacle inflation and obstacle-free-path guarantees, and a comparative performance study against BFS and A* in indoor simulations. The framework enables flexible experimentation with mapping, localization, and planning components and has practical implications for indoor robotic navigation and benchmarking of planning algorithms.
Abstract
We present a tool AUTONAV that automates the mapping, localization, and path-planning tasks for autonomous navigation of robots. The modular architecture allows easy integration of various algorithms for these tasks for comparison. We present the generated maps and path-plans by AUTONAV in indoor simulation scenarios.