Trajectory Planning for UAV-Based Smart Farming Using Imitation-Based Triple Deep Q-Learning
Wencan Mao, Quanxi Zhou, Tomás Couso Coddou, Manabu Tsukada, Liu Yunling, Yusheng Ji
AI summary
Problem
Planning UAV trajectories for smart farming is hindered by environmental uncertainty, partial field observations, and strict battery limits, which degrade traditional optimization and standard multi-agent reinforcement learning methods.
Approach
The authors model the task as a Markov decision process and introduce ITDQN, a multi-agent reinforcement learning algorithm that uses elite imitation to cut exploration costs and a mediator Q-network to stabilize and accelerate training.
Key results
- Formulation of UAV smart farming trajectory planning as a multi-agent Markov decision process
- Development of ITDQN combining elite imitation and a mediator Q-network for stable MARL training
- 4.43% higher weed recognition rate compared to DDQN in simulated environments
- 6.94% higher data collection rate compared to DDQN validated in both simulation and real-world tests
Why it matters
Provides a scalable, battery-aware planning framework that enables more reliable and precise autonomous UAV operations for modern precision agriculture.
Abstract
Unmanned aerial vehicles (UAVs) have emerged as a promising auxiliary platform for smart agriculture, capa- ble of simultaneously performing weed detection, recognition, and data collection from wireless sensors. However, trajectory planning for UAV-based smart agriculture is challenging due to the high uncertainty of the environment, partial observations, and limited battery capacity of UAVs. To address these issues, we formulate the trajectory planning problem as a Markov decision process (MDP) and leverage multi-agent reinforcement learning (MARL) to solve it. Furthermore, we propose a novel imitation-based triple deep Q-network (ITDQN) algorithm, which employs an elite imitation mechanism to reduce explo- ration costs and utilizes a mediator Q-network over a double deep Q-network (DDQN) to accelerate and stabilize training and improve performance. Experimental results in both simulated and real-world environments demonstrate the effectiveness of our solution. Moreover, our proposed ITDQN outperforms DDQN by 4.43% in weed recognition rate and 6.94% in data collection rate.