SMBU-PolarBear-Robotics-Team/pb2025_sentry_nav

What it solves

This project provides a comprehensive navigation and simulation package for "sentry" robots, specifically designed for the 2025 season of the RoboMaster competition. It enables robots to perform autonomous navigation, mapping (SLAM), and relocalization using LiDAR data in both simulated and real-world environments.

How it works

The system is built on the ROS 2 Humble and NAV2 navigation framework. It integrates several key components to handle movement and positioning:

• Odometry and Relocalization: It uses point_lio for odometry and small_gicp for relocalization, with a custom loam_interface to manage coordinate transformations between the LiDAR and the chassis.
• Path Planning: It utilizes the default NAV2 Global Planner for high-level routing and a custom pb_omni_pid_pursuit_controller for precise path tracking.
• Environment Analysis: It includes tools for terrain analysis to identify obstacles and their heights within and beyond a 4-meter range of the robot.
• Sensor Processing: It uses a Livox mid360 LiDAR and includes a tool (ign_sim_pointcloud_tool) to fix missing data fields in simulation point clouds to ensure the odometry system works correctly.

Who it’s for

This software is intended for robotics engineers and students participating in the RoboMaster competition or those developing autonomous navigation systems for omnidirectional robots using ROS 2 and LiDAR.

Highlights

• Dual-Mode Support: Fully compatible with both Gazebo simulation and real-world robot deployment.
• Optimized Coordinate Logic: Specifically handles the implicit transformations between the LiDAR origin and the chassis origin.
• Multi-Robot Ready: Implements a namespace design to allow the system to be extended for multiple robots.
• Integrated Tooling: Includes a full suite of drivers, controllers, and terrain analysis tools in a single repository.