A Survey of Methods and Architectures for Deep Convolutional U-Shape Network with Jump Attention-Based Vision Transformer for Integrated Sequence Scheduling and Trajectory Planning with Obstacle Avoidance in Wireless Rechargeable Sensor Networks

Abstract

Wireless Rechargeable Sensor Networks (WRSNs) have emerged as a critical advancement in addressing the energy limitations of conventional Wireless Sensor Networks (WSNs). By incorporating mobile chargers and intelligent scheduling mechanisms, WRSNs enhance network longevity and operational efficiency. However, optimizing sequence scheduling, trajectory planning, and obstacle avoidance remains a complex challenge due to dynamic environments, energy constraints, and multi-objective optimization requirements. Recent developments in deep learning, particularly Convolutional Neural Networks (CNNs), U-shaped architectures (U-Net), and Vision Transformers (ViTs), have demonstrated significant potential in addressing these challenges. CNN-based U-shaped architectures are highly effective in spatial feature extraction and segmentation tasks, while transformers provide superior capability in modelling long-range dependencies using self-attention mechanisms. Hybrid models such as TransUNet and Swin-Unet integrate convolutional and transformer architectures to capture both local and global contextual information, significantly improving decision-making performance in trajectory optimization tasks. This survey explores state-of-the-art deep learning approaches for integrated sequence scheduling and trajectory planning with obstacle avoidance in WRSNs. The study reviews recent literature from 2020–2023, categorizing methods based on CNN-based models, transformer-based models, hybrid architectures, and reinforcement learning techniques. Additionally, it analyses optimization strategies such as multi-objective algorithms and attention-based mechanisms for improving efficiency and adaptability. The survey identifies key challenges, including computational complexity, scalability, real-time deployment constraints, and data dependency. Finally, future research directions are outlined, focusing on lightweight architectures, edge computing integration, and hybrid intelligent optimization frameworks.