ARDDQN: Attention Recurrent Double Deep Q-Network for UAV Coverage Path Planning and Data Harvesting

TL;DR

This work tackles UAV coverage path planning and data harvesting in IoT-enabled environments under energy constraints by proposing ARDDQN, which fuses a Double Deep Q-Network with recurrent neural networks and an attention mechanism. A global-local map processing approach and a POMDP formulation enable scalable, temporally aware decision making, with target equations such as $Y(s,a,s') = r(s,a) + gamma Q_bar_phi(s', argmax_a' Q_main(s', a'))$ guiding learning. Empirical results on Manhattan32 and Urban50 show ARDDQN variants outperform baselines in CPP and DH, with attention-based LSTM delivering the strongest gains in coverage, landing, and data collection. The findings highlight the method’s potential for robust, scalable UAV planning in urban IoT contexts and point toward extending to multi-UAV systems for even greater data-harvesting efficiency.

Abstract

Unmanned Aerial Vehicles (UAVs) have gained popularity in data harvesting (DH) and coverage path planning (CPP) to survey a given area efficiently and collect data from aerial perspectives, while data harvesting aims to gather information from various Internet of Things (IoT) sensor devices, coverage path planning guarantees that every location within the designated area is visited with minimal redundancy and maximum efficiency. We propose the ARDDQN (Attention-based Recurrent Double Deep Q Network), which integrates double deep Q-networks (DDQN) with recurrent neural networks (RNNs) and an attention mechanism to generate path coverage choices that maximize data collection from IoT devices and to learn a control scheme for the UAV that generalizes energy restrictions. We employ a structured environment map comprising a compressed global environment map and a local map showing the UAV agent's locate efficiently scaling to large environments. We have compared Long short-term memory (LSTM), Bi-directional long short-term memory (Bi-LSTM), Gated recurrent unit (GRU) and Bidirectional gated recurrent unit (Bi-GRU) as recurrent neural networks (RNN) to the result without RNN We propose integrating the LSTM with the Attention mechanism to the existing DDQN model, which works best on evolution parameters, i.e., data collection, landing, and coverage ratios for the CPP and data harvesting scenarios.

Figures (11)

• Figure 1: Proposed scenario for UAV-assisted coverage path planning and data harvesting
• Figure 2: Figure shows the proposed model for map processing.
• Figure 3: Figure shows the proposed integrated architecture for attention-based recurrent networks with the DDQN Model.
• Figure 4: Coverage path planning on "Manhattan32," Map
• Figure 5: Coverage path planning on "Urban50" Map