[Project] Astryin: A Post-Mortem Analysis & Visualization Toolkit for ROS 2 Navigation

Robotics developer Kehan Zhou has released Astryin, an open-source Python toolkit designed to solve a persistent pain point in ROS 2 navigation development. Currently in MVP stage, the tool automates post-mortem analysis of robot navigation performance by processing standard ROS 2 bag files. It addresses the common trial-and-error approach to tuning Nav2 parameters—such as those for the DWB (Dynamic Window Approach) or MPPI (Model Predictive Path Integral) controllers—by replacing subjective visual assessments in RViz with concrete, quantitative data.

Astryin's technical core features dynamic TF compensation to ensure accurate time-alignment of map-to-odom transformations, and automated 'Motion Windowing' to filter out stationary periods and focus analysis on active navigation segments. For a sample TurtleBot3 run, it calculated a motion duration of 18.75 seconds, a path length of 3.56 meters, and identified a maximum tracking error of 0.242 meters. The toolkit generates both numerical summaries and visualizations, overlaying global plans, local plans, and actual odometry to help developers correlate a controller's intended actions with the robot's physical execution.

The project emerged from community discussions around the lack of standardized metrics for evaluating navigation stack performance. By providing command-line tools (`astryin analyze` and `astryin plot`), Zhou aims to give developers a shared framework for benchmarking improvements. This data-driven approach could significantly reduce the time spent diagnosing issues like robot oscillation or unexpected behavior, making the navigation tuning process more systematic and reproducible across different robotic platforms and environments.