Real-Time Collaborative Route Planning in Autonomous Vehicles by Leveraging Optimized Quantum Neural Networks

In this paper, we present an approach to enhance the intelligence of autonomous vehicles by focusing on real-time collaborative route planning. This aspect is critical not only for individual vehicles but also for ensuring effective interaction among multiple autonomous systems. The challenge is to optimize route planning algorithms for real-time operation without compromising accuracy. To address this, we leverage quantum computing techniques, integrating them with CNN architectures to develop Quantum Convolutional Neural Networks (QCNNs) and the Optimized QCNN. Neural networks play a key role in tasks such as predicting paths, estimating traffic flow, and selecting optimal routes, ensuring efficient and adaptive planning. The integration of quantum computing with neural networks significantly enhances real-time performance and the Optimized QCNN, developed using the Circuit Depth Reduction strategy, further improves accuracy and convergence rate. Our evaluation, conducted on the Cityscapes Dataset, demonstrates that QCNNs comparing with traditional CNNs achieve faster convergence rate and lower loss value. Furthermore, Optimized QCNNs surpass standard QCNNs in both convergence rate and loss, while also achieving higher accuracy even than traditional CNNs, making them a powerful model for real-time collaborative route planning in autonomous vehicles.