In the Internet of Autonomous Vehicles (IoAV), task offloading is a method to address computationally intensive tasks and ensure safe operation vehicles. However, under extreme weather conditions, number these significantly increases, posing higher risks challenges. Therefore, mitigate vehicles, it crucial make quick effective decisions during process. Currently, most methods in this domain utilize Deep Reinforcement Learning (DRL). large parameters deep networks results characteristics long...

Vehicular fog computing (VFC) could provide fast task processing services for vehicles. To make vehicles/fog nodes willing to buy/sell resources, a double auction mechanism considering the interests of all parties is needed. However, few works study issue in VFC. Different from existing edge-related which only considers price, some nonprice attributes (location, reputation, and power) are also important providing fair resource allocation In this article, we propose multiattribute-based VFC,...

Overtaking and lane-changing are essential maneuvers in autonomous driving. A correct timely operation can reduce the probability of collisions, enable vehicles to avoid potential hazards on road, enhance overall safety. In addition, they also play a vital role ensuring efficient flow traffic arrival at destinations. Currently, dominant research this problem is based Deep Q-Network (DQN) approach some its improved variants solve it. However, limited by deep networks, there inevitably...

The development of communication technologies which support traffic-intensive applications presents new challenges in designing a real-time traffic analysis architecture and an accurate method that suitable for wide variety types. Current methods are executed on the cloud, needs to upload data. Fog computing is more promising way save bandwidth resources by offloading these tasks fog nodes. However, models based traditional machine learning need retrain all data when updating trained model,...

In the upcoming 6G era, edge artificial intelligence (AI), as a key technology, will be able to deliver AI processes anytime and anywhere by deploying of models on devices. As hot issue in public safety, person re-identification (Re-ID) also needs its urgently deployed devices realize real-time accurate recognition. However, due complex scenarios other practical reasons, performance model is poor practice. This especially case places, where most people have similar characteristics, there are...

Through offloading tasks to surrounding edge nodes, the edge-assistant vehicular network (EAVN) could provide faster and more efficient services. To promote deployment of EAVN, we need a reasonable resource allocation scheme. Nowadays, there are few auction mechanisms in EAVN scenario. Moreover, mechanism other computing scenario has following limitations. Firstly, current works only focus on static/offline do not consider that users will join leave system at any time EAVN. Secondly,...

Pedestrian re-identification (Re-ID) leverages cross-camera data acquired by the Internet of Things (IoT) devices and sensors to identify, monitor, analyze pedestrians, allowing IoT applications provide more intelligent, secure, tailored services. Current pedestrian Re-ID research faces many challenges, such as low image resolution, perspective changes, posture light occlusions, resulting in models trained on other datasets being unable be directly applied showing poor generalization...

The advancement of the Internet Autonomous Vehicles has facilitated development and deployment numerous onboard applications. However, delay-sensitive tasks generated by these applications present enormous challenges for vehicles with limited computing resources. Moreover, are often interdependent, preventing parallel computation severely prolonging completion times, which results in substantial energy consumption. Task-offloading technology offers an effective solution to mitigate...

Autonomous driving has brought about a growing interest in enhancing traffic efficiency and ensuring road safety. One of the fundamental functions autonomous technology is lane-keeping, which become popular research topic driving. However, current deep reinforcement learning (DRL)-based algorithms used for solving lane-keeping problems have limitations, such as low sample utilization high time cost complex scenarios. To address this, we propose transfer Double Deep Q Network (LK-TDDQN)...

Accurate detection of objects in the three-dimensional world is a core problem for autonomous vehicles and ensures safety vehicles. Nevertheless, complex real has different degrees occlusion night driving conditions, lack detail images road scenes low-light conditions may increase risk collisions And many previous unimodal methods do not solve such problems well. In this paper, we propose PointrendPainting, which fuses image point cloud data early processing stage. The method exploits...

Partial differential equations (PDEs) usually apply for modeling complex physical phenomena in the real world, and corresponding solution is key to interpreting these problems. Generally, traditional solving methods suffer from inefficiency time consumption. At same time, current rise machine learning algorithms, represented by Deep Operator Network (DeepONet), could compensate shortcomings effectively predict solutions of PDEs operators data. The deep learning-based focus on one-dimensional...

Recently, vehicle classification is becoming increasingly important with the development of automated driving technology. In particular, it can provide basis and prerequisites for autonomous vehicles to make decisions in terms improving safety. However, current mainstream methods are deep learning algorithms based on Convolutional Neural Networks (CNN), which mainly focused cloud, these have complex models large training parameters. addition, computationally intensive urgent tasks, poor...