Files
Abstract
This thesis develops an advanced simulation framework within the Gazebo environment to enhance the realism and efficiency of modeling blast phenomena. It tackles the challenge of synchronizing visual, acoustic, and pressure data to simulate explosions accurately. By integrating a suite of technologies-including a robotic operating system (ROS), an unmanned aircraft mission planner (QGroundControl), a vehicle autopilot simulator (PX4 Software In The Loop - SITL), and a flight control neural network training environment (OpenAI GymFC)-the framework facilitates precise emulation of real-time events: visual effects at the speed of light, acoustic propagation through seismic and air mediums, and dynamic pressure variations. The implementation utilizes a client-server architecture, enabling real-time adjustments and reducing latency, which enhances the simulation quality. The research introduces specialized plugins to model and manage different aspects of the blast, demonstrating significant improvement in both fidelity and operational efficiency. These enhancements make the simulation tool a valuable asset for training and analytical applications in safety and defense sectors, providing comprehensive insights into blast dynamics.