Optimized Real-Time Path Planning and Obstacle Avoidance for Dynamic Objects Using a Velocity-Based Approach

This research presents a novel real-time trajectory replanning and obstacle avoidance strategy for autonomous drones using depth camera technology. It uniquely combines VFH* and 3DVFH+ algorithms, configured for 3D spaces, a new application area for these traditionally 2D-based methods. This integration enhances UAV navigation by ensuring smoother, more reliable flight paths. This method aligns vertical resolution in 3DVFH+ with VFH*'s horizontal resolution and synchronizes histogram resolutions, obstacle detection thresholds, and safety distances. The approach dynamically replans trajectories based on depth data and cost analysis, improving safety and efficiency. This enhanced trajectory planning framework significantly advances real-time decision-making and system autonomy. The effectiveness of this method has been validated through realworld and simulated testing in ROS Gazebo, demonstrating its superiority over existing methods. This approach significantly advances navigating complex 3D environments, increasing drone autonomy and adaptability.