arXiv subjects

Zhi-Li Zhang

Publications and source records attributed to Zhi-Li Zhang.

At least 19 recordsLinked to original sources

arXiv2026

Multi-Agent Video Prediction: Self-Correcting Conditional Frames for Dynamic Scene Forecasting

Transmission latency significantly degrades user quality of experience in real-time interactive perception systems. In remote driving, maintaining reliable visual feedback is critical for safe operation under dynamic network variability. Although video prediction offers a promising approach to compensate for short-term transmission delays and approximate near-zero-latency streaming, prediction-only methods remain vulnerable in highly dynamic scenes, especially when newly emerged objects appear during uplink outages. To address these challenges, we propose a multi-agent video prediction framework that combines continuous edge-side video prediction with lightweight mask-guided conditional frame reconditioning. The framework consists of three role-specialized agents: a continuous prediction agent for low-latency visual continuity, a vehicle-side trigger agent for detecting newly appeared objects, and a conditional reconditioning agent that repairs the predictor conditioning state using sparse mask guidance. This design enables semantic recovery of exogenous scene changes without requiring full-frame retransmission. We validate the proposed framework through extensive experiments on benchmark video data under realistic 5G communication traces. Results show that our method improves semantic recovery of novel objects while preserving perceptual quality and practical runtime efficiency under network-induced disruptions.

cs.NI↗

arXiv2026

Impact of Data Compression on Downstream AI Tasks: A Study using Teleoperated Driving over 5G

Teleoperation, such as remote driving, is considered as a key use case of 5G and Next-Generation (NextG) networks. In this context, robots, autonomous vehicles, or other autonomous agents transmit sensor data over mobile networks to edge or cloud servers, where AI systems collaborate with human operators to provide situational awareness and enable remote control. In the case of teleoperated driving, vehicles are equipped with an array of cameras and LiDAR devices, which can generate 100s Mbps (megabits per second) of data. As shown in existing measurement studies, such data volumes far exceed the \emph{uplink} capacity of currently deployed 5G networks, especially when multiple vehicles compete for radio resources. Data compression is thus imperative. In this paper, we explore the impact of sensor data compression on the performance of downstream AI tasks running in edge/cloud servers, which are crucial to alert human operators for safe teleoperation. Using object recognition and semantic segmentation as two example AI tasks, we study how data compression affects the performance of these two AI tasks using unimodal (video or LiDAR) and multi-modal (video+LiDAR) data. We find that lossy data compression generally decreases the performance of AI tasks. The performances of these AI tasks exhibit differing degrees of sensitivity based on the types of data sources and levels of compression. We also empirically identify an optimal trade-off point for the multi-modal vision tasks.

cs.NI↗

arXiv2026

Case for Vehicle-Edge Collaborative Multi-Sensor Data Fusion for Autonomous Vehicle Teleoperation

Teleoperation provides a critical safety fallback when autonomous vehicles (AVs) encounter scenarios that are outside their operational design domain. In practice, however, remote operators rely primarily on compressed camera streams over 5G, which often lack depth and spatial geometric cues for safe operation in complex dynamic environments. While multi-sensor fusion can enhance situational awareness, directly transmitting raw camera and LiDAR data is impractical due to 5G uplink bandwidth and latency constraints. In this paper, we propose SHARDED, a collaborative camera-LiDAR perception framework that deploys a feature-level fusion pipeline across the vehicle and edge to reduce uplink traffic while preserving 3D detection and depth estimation accuracy. We further design two complementary mechanisms for SHARDED: (i) A network-aware adaptive feature transmission mechanism that reduces data traffic by 50% on average (peaking at over 95%) compared to raw sensor data, and (ii) a latency-aware positional drift compensation mechanism to mitigate cross-modal misalignment induced by unstable network conditions. Evaluations on the nuScenes dataset and real-world 5G measurement traces show that SHARDED achieves competitive perception quality while reducing uplink bandwidth consumption and end-to-end latency.

cs.NI↗

arXiv2026

Multipath Adaptive Video Streaming with Multiple Description Neural Video Codec over 5G Networks

5G networks employ multiple radio channels to meet growing demands for bandwidth and high-resolution video streaming for emerging applications. However, existing multipath video systems are largely designed around monolithic codecs, which require sufficiently complete chunk delivery, or layered codecs, which depend on timely base-layer delivery. Under fast-varying 5G conditions with blockage, handovers, and heterogeneous path capacities, we observe that decoding dependencies in existing codecs make multipath delivery fragile: transient under-delivery of critical video data can directly trigger stalls and degrade QoE. This paper proposes NeuralMDC, a neural multiple-description video codec co-designed with multipath streaming for dynamic 5G networks. NeuralMDC encodes each video chunk into independently decodable and mutually refinable description streams, each spanning the full chunk. This design changes the multipath delivery unit from dependent packets or layers to independent chunk-level streams, so missing streams primarily reduce quality rather than making the chunk undecodable. Built on NeuralMDC, we develop a user-space multipath streaming system that maps description streams to heterogeneous 5G paths with simple yet effective scheduling logic. Across trace-driven emulation and operational 5G experiments, NeuralMDC improves QoE by 26%-44% over existing monolithic, layered, and neural streaming systems, improves video quality by up to 41.8%, and keeps stall ratios below 0.32%.

cs.NI↗

arXiv2026

Spectrum & RAN Sharing: A Measurement-based Case Study of Commercial 5G Networks in Spain

Radio Access Network (RAN) sharing, which often also includes spectrum sharing, is a strategic cooperative agreement among two or more mobile operators, where one operator may use another's RAN infrastructure to provide mobile services to its users. By mutually sharing physical sites, radio elements, licensed spectrum and other parts of the RAN infrastructure, participating operators can significantly reduce the capital (and operational) expenditure in deploying and operating cellular networks, while accelerating coverage expansion -- thereby addressing the spectrum scarcity and infrastructure cost challenges in the 5G era and beyond. While the economic benefits of RAN sharing are well understood, the impact of such resource pooling on user-perceived performance remains underexplored, especially in real-world commercial deployments. We present, to the best of our knowledge, the first empirical measurement study of commercial 5G spectrum and RAN sharing. Our measurement study is unique in that, beyond identifying real-world instances of shared 5G spectrum and RAN deployment "in the wild", we also analyze users' perceived performance and its implication on Quality of Experience (QoE). Our study provides critical insights into resource management (i.e., pooling) and spectrum efficiency, offering a blueprint (and implications) for network evolution in 5G, 6G and beyond.

cs.NI↗

arXiv2026

Demystifying Starlink Network Performance under Vehicular Mobility with Dynamic Beam Switching

In the last few years, considerable research efforts have focused on measuring and improving Starlink network performance, especially for user terminals (UTs) in stationary scenarios. However, the performance of Starlink networks in mobility settings, particularly with frequent changes in the UT's orientation, and the impact of environmental factors, such as transient obstructions, has not been thoroughly studied, leaving gaps in understanding the causes of performance degradation. Recently, researchers have started identifying the communicating satellites to evaluate satellite selection strategies and the impact on network performance. However, existing Starlink satellite identification methods only work in stationary, obstruction-free scenarios, as they do not account for UT mobility, obstructions or detect dynamic beam switching events. In this paper, we reveal that the UT can perform multiple dynamic beam switching attempts to connect to different satellites when the UT-satellite link is degraded. This degradation can occur either due to the loss of line-of-sight (LoS) from changes in the FOV or obstructions, or due to poor signal quality, extending UT-satellite handovers beyond the well-known 15-second regular handover interval. We propose a mobility-aware Starlink satellite identification method that detects dynamic beam switching events, and plausibly explain network performance using UT's diagnostic data and connected satellite information. Our findings demystifies the mobile Starlink network performance degradations, which is crucial to enhance the end-to-end performance of transport layer protocols and in diverse application scenarios.

cs.NI↗

arXiv2026

Examining DOM Coordinate Effectiveness For Page Segmentation

Web pages form a cornerstone of available data for daily human consumption and with the rise of LLM-based search and learning systems a treasure trove of valuable data. The scale of this data and its unstructured format still continue to grow requiring ever more robust automated extraction and retrieval mechanisms. Existing work, leveraging the web pages Document Object Model (DOM), often derives clustering vectors from coordinates informed by the DOM such as visual placement or tree structure. The construction and component value of these vectors often go unexamined. Our work proposes and examines DOM coordinates in a detail to understand their impact on web page segmentation. Our work finds that there is no one-size-fits-all vector, and that visual coordinates under-perform compared to DOM coordinates by about 20-30% on average. This challenges the necessity of including visual coordinates in clustering vectors. Further, our work finds that simple vectors, comprised of single coordinates, fare better than complex vectors constituting 68.2% of the top performing vectors of the pages examined. Finally, we find that if a vector, clustering algorithm, and page are properly matched, one can achieve overall high segmentation accuracy at 74%. This constitutes a 20% improvement over a naive application of vectors. Conclusively, our results challenge the current orthodoxy for segmentation vector creation, opens up the possibility to optimize page segmentation via clustering on DOM coordinates, and highlights the importance of finding mechanisms to match the best approach for web page segmentation.

cs.IR↗

arXiv2025

Teleoperating Autonomous Vehicles over Commercial 5G Networks: Are We There Yet?

Remote driving, or teleoperating Autonomous Vehicles (AVs), is a key application that emerging 5G networks aim to support. In this paper, we conduct a systematic feasibility study of AV teleoperation over commercial 5G networks from both cross-layer and end-to-end (E2E) perspectives. Given the critical importance of timely delivery of sensor data, such as camera and LiDAR data, for AV teleoperation, we focus in particular on the performance of uplink sensor data delivery. We analyze the impacts of Physical Layer (PHY layer) 5G radio network factors, including channel conditions, radio resource allocation, and Handovers (HOs), on E2E latency performance. We also examine the impacts of 5G networks on the performance of upper-layer protocols and E2E application Quality-of-Experience (QoE) adaptation mechanisms used for real-time sensor data delivery, such as Real-Time Streaming Protocol (RTSP) and Web Real Time Communication (WebRTC). Our study reveals the challenges posed by today's 5G networks and the limitations of existing sensor data streaming mechanisms. The insights gained will help inform the co-design of future-generation wireless networks, edge cloud systems, and applications to overcome the low-latency barriers in AV teleoperation.

cs.NI↗

arXiv2024

Robust Multiple Description Neural Video Codec with Masked Transformer for Dynamic and Noisy Networks

Multiple Description Coding (MDC) is a promising error-resilient source coding method that is particularly suitable for dynamic networks with multiple (yet noisy and unreliable) paths. However, conventional MDC video codecs suffer from cumbersome architectures, poor scalability, limited loss resilience, and lower compression efficiency. As a result, MDC has never been widely adopted. Inspired by the potential of neural video codecs, this paper rethinks MDC design. We propose a novel MDC video codec, NeuralMDC, demonstrating how bidirectional transformers trained for masked token prediction can vastly simplify the design of MDC video codec. To compress a video, NeuralMDC starts by tokenizing each frame into its latent representation and then splits the latent tokens to create multiple descriptions containing correlated information. Instead of using motion prediction and warping operations, NeuralMDC trains a bidirectional masked transformer to model the spatial-temporal dependencies of latent representations and predict the distribution of the current representation based on the past. The predicted distribution is used to independently entropy code each description and infer any potentially lost tokens. Extensive experiments demonstrate NeuralMDC achieves state-of-the-art loss resilience with minimal sacrifices in compression efficiency, significantly outperforming the best existing residual-coding-based error-resilient neural video codec.

cs.CV↗

arXiv2024

QUIC is not Quick Enough over Fast Internet

QUIC is expected to be a game-changer in improving web application performance. In this paper, we conduct a systematic examination of QUIC's performance over high-speed networks. We find that over fast Internet, the UDP+QUIC+HTTP/3 stack suffers a data rate reduction of up to 45.2% compared to the TCP+TLS+HTTP/2 counterpart. Moreover, the performance gap between QUIC and HTTP/2 grows as the underlying bandwidth increases. We observe this issue on lightweight data transfer clients and major web browsers (Chrome, Edge, Firefox, Opera), on different hosts (desktop, mobile), and over diverse networks (wired broadband, cellular). It affects not only file transfers, but also various applications such as video streaming (up to 9.8% video bitrate reduction) and web browsing. Through rigorous packet trace analysis and kernel- and user-space profiling, we identify the root cause to be high receiver-side processing overhead, in particular, excessive data packets and QUIC's user-space ACKs. We make concrete recommendations for mitigating the observed performance issues.

cs.NI↗

arXiv2024

HERL: Tiered Federated Learning with Adaptive Homomorphic Encryption using Reinforcement Learning

Federated Learning is a well-researched approach for collaboratively training machine learning models across decentralized data while preserving privacy. However, integrating Homomorphic Encryption to ensure data confidentiality introduces significant computational and communication overheads, particularly in heterogeneous environments where clients have varying computational capacities and security needs. In this paper, we propose HERL, a Reinforcement Learning-based approach that uses Q-Learning to dynamically optimize encryption parameters, specifically the polynomial modulus degree, $N$, and the coefficient modulus, $q$, across different client tiers. Our proposed method involves first profiling and tiering clients according to the chosen clustering approach, followed by dynamically selecting the most suitable encryption parameters using an RL-agent. Experimental results demonstrate that our approach significantly reduces the computational overhead while maintaining utility and a high level of security. Empirical results show that HERL improves utility by 17%, reduces the convergence time by up to 24%, and increases convergence efficiency by up to 30%, with minimal security loss.

cs.CR↗

arXiv2024

A Comprehensive Real-World Evaluation of 5G Improvements over 4G in Low- and Mid-Bands

As discussions around 6G begin, it is important to carefully quantify the spectral efficiency gains actually realized by deployed 5G networks as compared to 4G through various enhancements such as higher modulation, beamforming, and MIMO. This will inform the design of future cellular systems, especially in the mid-bands, which provide a good balance between bandwidth and propagation. Similar to 4G, 5G also utilizes low-band (<1 GHz) and mid-band spectrum (1 to 6 GHz), and hence comparing the performance of 4G and 5G in these bands will provide insights into how further improvements can be attained. In this work, we address a crucial question: is the performance boost in 5G compared to 4G primarily a result of increased bandwidth, or do the other enhancements play significant roles, and if so, under what circumstances? Hence, we conduct city-wide measurements of 4G and 5G cellular networks deployed in low- and mid-bands in Chicago and Minneapolis, and carefully quantify the contributions of different aspects of 5G advancements to its improved throughput performance. Our analyses show that (i) compared to 4G, the throughput improvement in 5G today is mainly influenced by the wider channel bandwidth, both from single channels and channel aggregation, (ii) in addition to wider channels, improved 5G throughput requires better signal conditions, which can be delivered by denser deployment and/or use of beamforming in mid-bands, (iii) the channel rank in real-world environments rarely supports the full 4 layers of 4x4 MIMO and (iv) advanced features such as MU-MIMO and higher order modulation such as 1024-QAM have yet to be widely deployed. These observations and conclusions lead one to consider designing the next generation of cellular systems to have wider channels, perhaps with improved channel aggregation, dense deployment with more beams.

cs.NI↗

arXiv2023

Examining Interplay of Compression and Encryption and Applicability to 5G Teleoperations

Modern IoT and networked systems rely on fast and secure delivery of time-critical information. Use cases such as teleoperations require fast data delivery over mobile networks, which despite improvements in 5G are still quite constrained. Algorithms for encryption and compression provide security and data size efficiency, but come with time and data size trade-offs. The impact of these trade-offs is related to the order in which these operations are applied, and as such necessitates a robust exploration from a performance perspective. In this paper, we assess several compression and encryption algorithms, combinations of their execution order, timings and size changes from such order, and the implications of such changes on 5G teleoperations. From our assessments we have three major takeaways: (1) Compression-First is faster and more compressed, except for certain circumstances. (2) In these specific circumstances, the compression against a raw file leads to a lengthier time than if applied to an encrypted file first. (3) Applying both encryption and compression on data samples larger than 10MB is impractical for real time transmission due to the incurred delay.

cs.NI↗

arXiv2023

Dissecting IoT Device Provisioning Process

We examine in detail the provisioning process used by many common, consumer-grade Internet of Things (IoT) devices. We find that this provisioning process involves the IoT device, the vendor's cloud-based server, and a vendor-provided mobile app. In order to better understand this process, we develop two toolkits. IoT-Dissect I enables us to decrypt and examine the messages exchanged between the IoT device and the vendor's server, and between the vendor's server and a vendor-provided mobile app. IoT-Dissect II permits us to reverse engineer the vendor's mobile app and observe its operation in detail. We find several potential security issues with the provisioning process and recommend ways to mitigate these potential problems. Further, based on these observations, we conclude that it is likely feasible to construct a vendor-agnostic IoT home gateway that will automate this largely manual provisioning process, isolate IoT devices on their own network, and perhaps open the tight association between an IoT device and the vendor's server.

cs.NI↗

arXiv2023

Mid-Band 5G: A Measurement Study in Europe and US

Fifth Generation (5G) mobile networks mark a significant shift from previous generations of networks. By introducing a flexible design, 5G networks support highly diverse application requirements. Currently, the landscape of previous measurement studies does not shed light on 5G network configuration and the inherent implications to application performance. In this paper, we precisely fill this gap and report our in-depth multi-country measurement study on 5G deployed at mid-bands. This is the common playground for U.S. and European carriers. Our findings reveal key aspects on how carriers configure their network, including spectrum utilization, frame configuration, resource allocation and their implication on the application performance.

cs.NI↗

arXiv2023

CageCoach: Sharing-Oriented Redaction-Capable Distributed Cryptographic File System

The modern data economy is built on sharing data. However, sharing data can be an expensive and risky endeavour. Existing sharing systems like Distributed File Systems provide full read, write, and execute Role-based Access Control (RBAC) for sharing data, but can be expensive and difficult to scale. Likewise such systems operate on a binary access model for their data, either a user can read all the data or read none of the data. This approach is not necessary for a more read-only oriented data landscape, and one where data contains many dimensions that represent a risk if overshared. In order to encourage users to share data and smooth out the process of accessing such data a new approach is needed. This new approach must simplify the RBAC of older DFS approaches to something more read-only and something that integrates redaction for user protections. To accomplish this we present CageCoach, a simple sharing-oriented Distributed Cryptographic File System (DCFS). CageCoach leverages the simplicity and speed of basic HTTP, linked data concepts, and automatic redaction systems to facilitate safe and easy sharing of user data. The implementation of CageCoach is available at https://github.umn.edu/CARPE415/CageCoach.

cs.CR↗

arXiv2022

A Multiple-Entanglement Routing Framework for Quantum Networks

Quantum networks are gaining momentum in finding applications in a wide range of domains. However, little research has investigated the potential of a quantum network framework to enable highly reliable communications. The goal of this work is to investigate and design the multiple-entanglement routing framework, namely k-entangled routing. In particular, the $k$-entangled routing will enable k paths connecting all demands (source-destination pairs) in the network. To design the $k$-entangled routing, we propose two algorithms that are called Sequential Multi-path Scheduling Algorithm and Min-Cut-based Multi-path Scheduling Algorithm. In addition, we evaluate the performance of the proposed algorithms and models through a realistic quantum network simulator, NetSquid, that models the stochastic processes underlying quantum communications. The results show that the proposed algorithms (SMPSA and MCSA) largely enhance the network's traffic flexibility. The proposed paradigms would lay the foundation for further research on the area of entanglement routing.

cs.NI↗

arXiv2021

CLARA: A Constrained Reinforcement Learning Based Resource Allocation Framework for Network Slicing

As mobile networks proliferate, we are experiencing a strong diversification of services, which requires greater flexibility from the existing network. Network slicing is proposed as a promising solution for resource utilization in 5G and future networks to address this dire need. In network slicing, dynamic resource orchestration and network slice management are crucial for maximizing resource utilization. Unfortunately, this process is too complex for traditional approaches to be effective due to a lack of accurate models and dynamic hidden structures. We formulate the problem as a Constrained Markov Decision Process (CMDP) without knowing models and hidden structures. Additionally, we propose to solve the problem using CLARA, a Constrained reinforcement LeArning based Resource Allocation algorithm. In particular, we analyze cumulative and instantaneous constraints using adaptive interior-point policy optimization and projection layer, respectively. Evaluations show that CLARA clearly outperforms baselines in resource allocation with service demand guarantees.

cs.NI↗