Autonomous Navigation for Unmanned Aerial Vehicles (Uays) Using Machine Learning
S. Ram Prasath, K. Gokulakrishnan, M. Subramanian, S. Mohanap Priya · 2024
This study investigates how Unmanned Aerial Vehicles (UAVs) can use Machine Learning (ML) to facilitate autonomous navigation. The model, which is written in Python and is implemented with TensorFlow and sci-kit-learn, uses supervised learning—more specifically, convolutional neural networks—to procedure images. Model training is made easier by simulated scenarios on UAV platforms, which emphasize flexibility in real-time. The effectiveness of the ML model in enhancing obstacle avoidance as well as spatial awareness is demonstrated by controlled field tests. Its superiority across conventional navigation systems is demonstrated through comparative analyses. Heat maps are one type of visualization that provides qualitative insights into the processes of decision-making. Notwithstanding achievements, a contemplative investigation tackles aberrations, laying the groundwork for a critical assessment. Enhancing adaptability in a range of weather scenarios and researching cooperative autonomy between several UAVs are among the suggestions.