How eyantra-robotics-competition/eyrc-25-26_krishi_drone works
Project type and purpose
• This is a ROS-based robotics simulation project (“eYRC25-26 Krishi Drone Theme”) that uses the Gazebo simulator to fly and control a nano-drone in a virtual greenhouse. It comprises:
• swift_pico: Control algorithms (PID/LQR) and ROS launch/scripts
• swift_pico_description: Gazebo world + SDF/XACRO models for the drone, arena, packages
• whycon: Overhead camera localization
• mav_comm / rotors_simulator / controller_tuner: Supporting middleware, simulation of vehicle dynamics, and parameter tuning
File structure
• Top-level modules:
• rotors_simulator: Gazebo plugin for rotor dynamics
• mav_comm: Middleware for MAVLink communication
• controller_tuner: ROS node(s) to tune PID/LQR gains at runtime
• whycon: Vision-based localization package
• swift_pico & swift_pico_description: Main application package (launch, config, src, world, models)
• Look for:
• swift_pico/launch/*.launch.py – orchestration of nodes and Gazebo world
• swift_pico/config/*.yaml – parameters for ROS nodes and map/bridge settings
• swift_pico/src/*.cpp/.py – control algorithms (PID, LQR)
• swift_pico_description/models & worlds – static simulation assets
README for additional insights
• Simulation phase using Gazebo before proceeding to hardware
• The drone must navigate an arena, detect infested plants via overhead vision, pick up and apply pesticides
Detail
• Sequence of launch executions (task_0.launch.py, task_1c.launch.py)
• Call out map_bridge.yaml and swift_pico_bridge.yaml for coordinate transforms
• Where the arena and package models live (swift_pico_description/models)
• Control flows (perception → pose_estimate → control → actuators)