Recent Advances in 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: A Systematic Review

Abstract

Wireless Rechargeable Sensor Networks (WRSNs) have emerged as an advanced solution to extend the lifetime of traditional Wireless Sensor Networks by enabling energy replenishment through mobile chargers. However, challenges such as efficient sequence scheduling, trajectory planning, and obstacle avoidance persist due to dynamic environments, energy constraints, and real-time decision-making requirements. Recent advancements in deep learning, particularly Deep Convolutional U-Shape Networks (U-Net) and Vision Transformers (ViTs), have shown strong potential in addressing these issues. U-Net architectures excel in feature extraction through their encoder–decoder structure with skip connections, providing precise spatial and contextual representations. However, to overcome limitations in capturing long-range dependencies, attention-based transformer mechanisms are integrated, enhancing global context modelling. Hybrid models such as TransUNet and UNetFormer combine convolutional feature extraction with transformer-based attention, improving performance in path planning, obstacle detection, and scheduling optimization. Additionally, transformer-based models effectively handle temporal dependencies for trajectory prediction. The integration of deep reinforcement learning and attention-guided frameworks further enables adaptive and energy-efficient charging strategies, improving system responsiveness, network lifetime, and overall reliability in WRSNs.