Fusion of LiDAR and Computer Vision for Autonomous Navigation in Gazebo

Simulating hardware continues to accelerate technology development to either provide initial proof of concept or function as a substitute when either hardware is either unavailable or damaged. Building an efficient simulator only requires basic real-world representations to provide feedback to software development. This paper discusses and tests the application of object detection and autonomous navigation on a Turtlebot3 Waffle platform in a simulated Gazebo environment. Object Detection and mapping are supported using Turtlebot’s Camera, Lidar, and IMU sensor. Testing for the first phase of development demonstrated the feasibility of simulation as applied to the IEEE Hardware Competition and successful detection and navigation between objects placed in a simulated, open environment using the methods proposed in this paper. Follow-on work with this simulator includes posting to the IEEE Southeast-CON 2023 discord site, updating the world parameters to better model the specifications of the IEEE Hardware Competition, and embedding the object detection algorithms directly into the navigation behavior tree.