ROS 2 | TurtleBot3 Obstacle Avoidance with Sim2Real Transfer [Tutorial]
1.7 هزار بار بازدید -
پارسال
-
This video demonstrates obstacle avoidance
This video demonstrates obstacle avoidance of a TurtleBot3 (a differential-drive robot) in Gazebo Simulator as well as real world using Robot Operating System 2 (ROS 2).
Controller makes the robot avoid obstacles by maintaining a safe distance from them. The robot makes use of de-coupled longitudinal and lateral PID controllers (with FIFO integral anti-windup mechanism) acting on frontal distance to collision (based on mean of frontal laserscan sector of 40°), distance to collision from oblique left and right sectors spanning 70° each, and distance to collision from left and right sectors spanning 55° each respectively for motion control. A pseudo-potential field method is implemented with a hierarchical 3 stage attention mechanism to act on close, fairly far and safely far obstacles. A finite state machine is implemented to account for turning left, turning right, going straight and cautiously rotating on the spot in case of too many obstacle in proximity. The script is setup to wait 4 seconds for proper initialization and then run for eternity to demonstrate dynamically variable obstacle avoidance application.
GitHub Repository: https://github.com/Tinker-Twins/Auton...
TurtleBot3 Packages: https://github.com/Tinker-Twins/Turtl...
ROS 2 Installation: ROS 2 | ROS 2 Foxy Installation [Tuto...
------------------------
Specifications:
------------------------
OS: Ubuntu 20.04
ROS 2: ROS 2 Foxy
Controller makes the robot avoid obstacles by maintaining a safe distance from them. The robot makes use of de-coupled longitudinal and lateral PID controllers (with FIFO integral anti-windup mechanism) acting on frontal distance to collision (based on mean of frontal laserscan sector of 40°), distance to collision from oblique left and right sectors spanning 70° each, and distance to collision from left and right sectors spanning 55° each respectively for motion control. A pseudo-potential field method is implemented with a hierarchical 3 stage attention mechanism to act on close, fairly far and safely far obstacles. A finite state machine is implemented to account for turning left, turning right, going straight and cautiously rotating on the spot in case of too many obstacle in proximity. The script is setup to wait 4 seconds for proper initialization and then run for eternity to demonstrate dynamically variable obstacle avoidance application.
GitHub Repository: https://github.com/Tinker-Twins/Auton...
TurtleBot3 Packages: https://github.com/Tinker-Twins/Turtl...
ROS 2 Installation: ROS 2 | ROS 2 Foxy Installation [Tuto...
------------------------
Specifications:
------------------------
OS: Ubuntu 20.04
ROS 2: ROS 2 Foxy
پارسال
در تاریخ 1402/04/12 منتشر شده
است.
1,700
بـار بازدید شده