Table 1 The comparison results of the EN-MASCA algorithm with some key studies.
From: Enhanced multi agent coordination algorithm for drone swarm patrolling in durian orchards
Study | Method | Application scenario | Key findings | Quantitative results |
|---|---|---|---|---|
Lu et al.11 | Neuromorphic computing technology, exploring potential applications in smart agriculture | Smart agriculture | Provides examples of neuromorphic computing for data processing and task optimization; proposes future development directions | No specific experimental data provided |
Li et al.12 | Highly configurable intelligent agricultural robotic arm (CARA), integrating image acquisition and deep processing modules | Smart agriculture tasks | Improves precision and efficiency in agricultural tasks | Execution efficiency improved by ~ 20%, error rate reduced by 15% |
Zhou et al.13 | Hybrid architecture combining data-driven models and physical models for parameter identification and error characterization | Vehicle lateral dynamics modeling | Achieved accurate and interpretable modeling, providing a reliable model for vehicle motion control | Parameter identification error reduced by 10–15% |
Chen et al.14 | Spatial attention mechanism and feature fusion module, combined with hybrid models of physical and dual attention neural networks | Vehicle dynamics modeling | Improves modeling accuracy in complex dynamic environments with limited data | Prediction accuracy improved by ~ 12–18% |
Meng et al.15 | Mobile navigation combined with visual perception, with path planning and obstacle avoidance | Human-machine interaction | Achieved human-preference-based task assignment and efficient navigation | Target grasp success rate improved by 12%, task completion time reduced by 10% |
Li et al.16 | Centroid Voronoi tessellation (CVT)-based path segmentation method for brain-controlled robot navigation | Brain-controlled robot navigation | Provides efficient path segmentation for generating arbitrary target navigation objectives | Navigation success rate improved by 15%, path planning time reduced by 8% |
Zhou et al.17 | Wavelet decomposition and denoising autoencoder for UAV anomaly detection | UAV fault detection | Improved anomaly detection accuracy in noisy data conditions | Detection accuracy increased by ~ 15–20% |
Chen et al.18 | CSMA/CA-based fair and efficient MAC protocol, supporting multi-user MIMO parallel uplink transmission | Multi-UAV communication | Enhances communication efficiency and optimizes resource allocation | Uplink transmission efficiency improved by ~ 25% |
Yin et al.21 | Deep reinforcement learning algorithm for adaptive UAV navigation | 3D UAV navigation environments | Proposed a new speed constraint loss function, improving UAV speed control capability | Path planning success rate improved by 20%, speed control error reduced by 30% |
Liang et al.24 | Integrated framework combining behavior decision-making, path planning, and motion control modules | High-speed mixed traffic scenarios | Improves safety of autonomous vehicles in mixed traffic environments | Collision rate reduced by 20%, path planning success rate improved by 18% |
This study (EN-MASCA) | Introduced DQN and PPO algorithms with a virtual navigator model for path planning and obstacle avoidance | Complex environments in durian orchards | Significantly improves flight stability, path accuracy, and collaboration of UAV swarms | Path planning deviation reduced by 25% Altitude fluctuation reduced by 30% Task efficiency increased by 20% |
