Automatic Path Planning for Spraying Drones Based on Deep Q-Learning

<p>The reduction of the agricultural workforce due to rapid development technology has resulted in labor shortages. Agricultural mechanization, such as drone use for pesticide spraying, can solve this problem. However, terrain, culture, and operational limitations mountainous orchards Taiwan make spraying challenging. By combining reinforcement learning with deep neural networks, we propose train drones avoid obstacles find optimal paths that reduce difficulties, costs, battery consumption. We experimented different reward mechanisms, network depths, flight direction granularities, environments devise a plan suitable sloping orchards. Reinforcement is more effective than traditional algorithms solving path planning complex environments.</p> <p> </p>