The ROS Navigation Stack

The ROS navigation stack is a toolkit designed for autonomous vehicle navigation, providing a comprehensive set of functionalities for localization, path planning, and map utilization. This stack enables mobile vehicles to autonomously navigate through complex environments, avoid obstacles, and reach target locations based on predefined paths.

The core modules of the ROS navigation stack include

1. AMCL (Adaptive Monte Carlo Localization): used for vehicle localization on a known map, AMCL employs a particle filter algorithm to adaptively estimate the vehicle’s pose based on sensor data.
2. Move Base: this is the central module of navigation, responsible for coordinating path planning, obstacle avoidance, and publishing control commands to the vehicle. Move Base integrates both global and local path planners, supporting dynamic environmental perception and real-time response.
3. Global planner & Local planner: the global planner (e.g., the A* or D* algorithm) calculates the optimal route from the current location to the target. The local planner (e.g., the TEB algorithm) adjusts the vehicle’s path in real-time to avoid obstacles encountered during navigation.

By configuring the parameters and aligning data from these core modules, the ROS navigation stack can be applied to ISSPA PAVS and PAVL vehicles for waypoint navigation. Additionally, it offers convenient plugin support for algorithm integration, such as the PSO algorithm available in ISSPA.