Driven by the evolutionary development of automobile industry and cellular technologies, dependable vehicular connectivity has become essential to realize future intelligent transportation systems (ITS). In this paper, we investigate how achieve content distribution in device-to-device (D2D)-based cooperative networks combining big data-based vehicle trajectory prediction with coalition formation game-based resource allocation. First, is predicted based on global positioning system...

Vehicular edge computing is essential to support future emerging multimedia-rich and delay-sensitive applications in vehicular networks. However, the massive deployment of infrastructures induces new problems including energy consumption carbon pollution. This motivates us develop BEGIN (Big data enabled EnerGy-efficient vehIcular computiNg), a programmable, scalable, flexible framework for integrating big analytics with computing. In this article, we first present comprehensive literature...

Robotic grasping is one of the most fundamental robotic manipulation tasks and has been subject extensive research. However, swiftly teaching a robot to grasp novel target object in clutter remains challenging. This paper attempts address challenge by leveraging attributes that facilitate recognition, grasping, rapid adaptation new domains. In this work, we present an end-to-end encoder-decoder network learn attribute-based with data-efficient capability. We first pre-train model variety...

Robotic grasping is one of the most fundamental robotic manipulation tasks and has been subject extensive research. However, swiftly teaching a robot to grasp novel target object in clutter remains challenging. This paper attempts address challenge by leveraging attributes that facilitate recognition, grasping, rapid adaptation new domains. In this work, we present an end-to-end encoder-decoder network learn attribute-based with data-efficient capability. We first pre-train model variety...

Vehicle-to-vehicle (V2V) communication which supports ubiquitous information exchange and content sharing among vehicles with little or no human intervention becomes a key enabler for intelligent transportation industry. In this paper, we adopt two-hop relay transmission mode to maximize the total spectrum efficiency of device-to-device (D2D) based vehicular cooperative network while ensuring quality service (QoS) provisions both V2V links cellular links. We propose an auction-matching joint...

In this paper, we investigate how to achieve reliable content distribution in device-to-device (D2D) based cooperative vehicular networks by combining big data vehicle trajectory prediction with coalition formation game resource allocation. Firstly, is predicted on global positioning system (GPS) and geographic information (GIS) data, which critical for finding longlasting connections. Then, the determination of groups different lifetimes formulated as a game. We model utility function...

The language-guided robot grasping task requires a agent to integrate multimodal information from both visual and linguistic inputs predict actions for target-driven grasping. While recent approaches utilizing Multimodal Large Language Models (MLLMs) have shown promising results, their extensive computation data demands limit the feasibility of local deployment customization. To address this, we propose novel CLIP-based parameter-efficient tuning (PET) framework designed three object...

When robots retrieve specific objects from cluttered scenes, such as home and warehouse environments, the target are often partially occluded or completely hidden. Robots thus required to search, identify a object, successfully grasp it. Preceding works have relied on pre-trained object recognition segmentation models find object. However, methods require laborious manual annotations train even fail novel objects. In this paper, we propose an Image-driven Object Searching Grasping (IOSG)...

Adversarial object rearrangement in the real world (e.g., previously unseen or oversized items kitchens and stores) could benefit from understanding task scenes, which inherently entail heterogeneous components such as current objects, goal environmental constraints. The semantic relationships among these are distinct each other crucial for multi-skilled robots to perform efficiently everyday scenarios. We propose a hierarchical robotic manipulation system that learns underlying maximizes...

When robots retrieve specific objects from cluttered scenes, such as home and warehouse environments, the target are often partially occluded or completely hidden. Robots thus required to search, identify a object, successfully grasp it. Preceding works have relied on pre-trained object recognition segmentation models find object. However, methods require laborious manual annotations train even fail novel objects. In this paper, we propose an Image-driven Object Searching Grasping (IOSG)...

Adversarial object rearrangement in the real world (e.g., previously unseen or oversized items kitchens and stores) could benefit from understanding task scenes, which inherently entail heterogeneous components such as current objects, goal environmental constraints. The semantic relationships among these are distinct each other crucial for multi-skilled robots to perform efficiently everyday scenarios. We propose a hierarchical robotic manipulation system that learns underlying maximizes...

Instance segmentation with unseen objects is a challenging problem in unstructured environments. To solve this problem, we propose robot learning approach to actively interact novel and collect each object's training label for further fine-tuning improve the model performance, while avoiding time-consuming process of manually labeling dataset. The Singulation-and-Grasping (SaG) policy trained through end-to-end reinforcement learning. Given cluttered pile objects, our chooses pushing...