Cellular communication in the sky will become increasingly important in the near future with the rise of airborne services such as drone taxis. While polarization has traditionally been a secondary concern in terrestrial communication due to multipath and reflections, the dominance of line-of-sight paths in aerial environment makes polarization alignment critical. This study is the first to conduct both experimental and mathematical analyses of polarization alignment to estimate the unknown polarization characteristics of commercial cellular base stations from the aerial device's perspective. We perform real measurements to evaluate these estimations, verifying the base station's practical polarization profile and validating our model. Using this model, we then propose PACMAN, a method that identifies the optimal polarization for each drone position, thereby enhancing alignment and maximizing throughput. By combining polarization prediction with antenna steering, PACMAN improves signal-tonoise- ratio and throughput by 20.7% on average (up to 67% in peak scenarios), offering a novel approach to optimizing cellularbased aerial communication systems.

has gained significant attention for Internet of Things (IoT) applications that require low-power and long-range communication such as animal tracking. However, LoRa's spreading factor (SF) selection mechanism and 1-hop topology lack effective per-link adaptation to dynamics of the communication channel nor provide comprehensive coverage of moving animals. To address this problem, we present AURORA, an adaptive and distributed SF control scheme for low-power multihop LoRa networks. AURORA exploits the key tradeoff of SF-receiver sensitivity vs. data rate-with a focus on energy efficiency. AURORA accurately predicts the packet delivery ratio (PDR) for each SF using efficient probing and model fitting techniques, enabling rapid per-link SF adaptation to improve both PDR and duty cycle in multihop networks. Through real-world experiments conducted on both indoor and outdoor testbeds, we demonstrate that AURORA effectively reduces energy consumption while ensuring reliable communication and extending network coverage. Compared to state-of-the-art approaches such as ADR+, AURORA improves the PDR by 14% and reduces the duty cycle by 18%.

Millimeter-wave (mmWave) communication using multiple-input multiple-output (MIMO) transmission is a key technology for increasing fifth-generation (5G) network capacity. This has the advantages of greater bandwidth and spatial diversity, but brings problems of line-of-sight blocking and high propagation loss. Dense deployment of base stations (BSs) has emerged as a promising architecture to address these challenges, but it often causes serious and complex interference problems due to the large number of mmWave beams. In this work, we develop low-complexity beam alignment schemes in 5G indoor mmWave ultra-dense networks (UDNs). Based on the centralized radio access network (C-RAN) architecture, we design practical non-iterative schemes using predefined analog beamforming training (ABT). In particular, we minimize performance degradation due to new connections by taking into account possible interference from future incoming connections. Through extensive experiments on a ray tracing simulator, we confirm that our proposed schemes achieve significant performance gains over existing schemes.

AI agents are enabling a new paradigm of agent-augmented real-time communication (RTC), where humans focus on high-level collaboration, while agents autonomously retrieve, analyze, and generate information in real time to support their interactions. These apps enable new experiences across various domains: for example, when corporate employees co-author a legal document, their agents can discuss and draft on their behalf, sparing them the burden of manually reviewing each other's work. As existing cloud-based agents suffer from privacy risks and unscalable server costs, on-device agent-augmented RTC offers a promising alternative. However, this on-device paradigm introduces a new networking challenge: contention between concurrent traffic flows generated by humans (for live video streaming) and agents (for sending context files for analysis). We design HFS, a framework to ensure both high live video quality and low agent response latency in agent-augmented RTC apps. We achieve the goal through an app-guided multi-flow transport approach, where a unified app-layer orchestrator jointly controls the sending rates of live video and agent context flows based on their heterogeneous app requirements. Our prototype built atop WebRTC and llama.cpp demonstrates that HAFS outperforms baselines, achieving 1.5x higher video quality while reducing agent response time by 31%.

High-quality wireless communication is paramount to coordinating flights and missions effectively on a micro-UAV (a.k.a. drone). However, most cellular antennas are optimized for users on the ground and have not been planned for aerial devices. Many have studied cellular communication quality, but few works explore the impact of altitude during a flight. Through real-world experiments using an actual drone, we demonstrate significant connectivity dynamics in the air for both 4G LTE and 5G NR, and reveal that fixed-coordinate flights cannot maintain high quality connectivity in response to those dynamics. To address this problem, we present ASCEND, a reinforcement learning-based 3D altitude selection scheme that maintains high-quality connectivity during a flight over a planned 2D path without requiring prior training. We evaluate ASCEND at multiple real-world locations to demonstrate a notable increase in expected throughput and a reduction in the proportion of low-quality legs during a flight mission.

Emerging drone delivery systems face challenges in last-meter package deployment due to positioning inaccuracies of GPS. Ultra-wideband (UWB) technology offers high-precision ranging, and when combined with IMU-based dead reckoning, it can enhance drone navigation to enable precise arrival at the destination. However, drone-to-tag UWB localization presents several challenges, including flip ambiguity, IMU drift, and horizontal insensitivity due to high-altitude flights. To address the challenges, we propose NOVA, a navigation optimization scheme via UWB-assisted iterative path planning. With NOVA, the drone follows a trajectory that mitigates flip ambiguity while an iterative joint optimization process simultaneously estimates the UWB tag’s position and refines the drone’s trajectory. At the vicinity of the tag, the drone performs subtle 2D corrective maneuvers during descent to ensure precise last-meter positioning. We evaluate NOVA via real experiments and simulations, demonstrating an average navigation error of 0.83 m with 25.45% trajectory overhead.

With mobile video calls now ubiquitous, ensuring seamless video-based real-time communications (RTC) remains a critical challenge for 5G operators. Despite abundant 5G bandwidth, video calls frequently experience low quality and unacceptable latency during channel fluctuations—not due to bandwidth limitations or congestion, but because of non-congestive delays in the radio access network (RAN). These delays stem from general-purpose RAN transmission procedures that prioritize radio resource efficiency over application latency through reactive scheduling and static timer-based retransmissions. Existing solutions largely address congestion-induced delays or sacrifice spectral efficiency to mitigate non-congestive delays.To address this, we present PAVE, a novel RAN-side solution that breaks the fundamental tradeoff between spectral efficiency and RTC latency. Our key insight is that RTC traffic exhibits distinct characteristics—periodic traffic generation and deadline-driven urgency—that can be leveraged to optimize RTC quality of experience (QoE) without sacrificing efficiency. For practical deployment, PAVE extracts these characteristics at the RAN, enables resource preallocation that goes beyond strictly periodic traffic patterns, and incorporates a selective retransmission and skip mechanism that maintains high spectral efficiency while strategically leveraging application-layer recovery. Implemented in an Open-RAN compliant RAN Intelligent Controller, PAVE improves tail frame rates by 1.8× and reduces video stalls by 94% in real-world evaluations.

Wi-Fi 7 introduces multi-link operation (MLO) that allows devices to operate simultaneously over multiple frequency bands, improving user-perceived latency and transmission opportunities. However, the actual benefits of multi-link devices (MLDs) under network saturation remain unclear due to practical constraints, such as in-device interference and non-simultaneous transmit-and-receive (NSTR) limitations. More critically, prior studies have assumed unequal radio frequency (RF) availability—assigning more antenna chains to MLDs than to single-link devices (SLDs)—which leads to overestimated gains from MLO. In this work, we revisit the saturation performance of MLDs by developing a unified framework that models standard-compliant channel access behavior across various device types and operation modes. Our analysis and simulations reveal that current MLD designs, regardless of STR/NSTR, have no throughput advantage over coexisting SLDs or alternative single-link operation (SLO), given an equal number of RF chains per device. Their inherent resource under-utilization even degrades overall network capacity despite increased deployment costs and complexity. Motivated by this, we identify enhanced multi-link single-/multi-radio (EMLXR) as the only effective throughput-maximizing mode, which can combine the benefits of opportunistic multi-link access and spatial multiplexing in theory. For practical viability against hardware-induced latency, we further present two extension strategies to enable adaptive transitions between MLO and SLO. Our design achieves up to 2× per-device throughput over legacy SLDs (static SLO), offering practical guidelines on how MLO-capable Wi-Fi 7 devices should behave to unlock their performance potential.

Non-linear chirp modulation has been empirically shown to improve interference resilience in dense LoRa networks, yet its theoretical origin remains underexplored. To close the gap, this letter presents a general interference model for polynomial-phase chirps and establishes the analytical basis of energy scattering effect (ESE). Introducing ESE score as a novel metric to quantify the spectral energy dispersion across FFT bins, we prove that asynchronous non-linear chirps yield consistently higher ESE score and hence suppress interference peaks more effectively than conventional linear chirps. Our numerical and simulation results further reveal a clear inverse correlation between ESE score and symbol error rate, validating the practical performance benefit of non-linear designs with stronger interference resilience.

Conventional capacitive touchscreens in mobile devices often malfunction in the presence of moisture, dust, or gloved hands, leading to unreliable user interactions. To overcome these limitations, we present TouchAI , a multimodal touch detection and localization system that leverages built-in inertial sensors and audio interfaces. Running as a background process, TouchAI constantly monitors inertial measurements for potential touch input and then activates acoustic sensing only upon detected events, thus reducing power consumption and privacy concerns associated with continuous audio recording. We introduce a novel event identification method that distinguishes touch-start and touch-end instances, providing precise timing for data sampling. For fine-grained touch localization, TouchAI employs a lightweight Transformer-based classification model to capture spatiotemporal features from combined acoustic-inertial signals. Experiments on commercial smartphones demonstrate true positive rates of 95.3% and 99.1% for touch detection and event identification, respectively, at false positive rates of less than 3%. Ultimately, TouchAI achieves up to 97.0% and 86.5% localization accuracy on 4 & times;2 and 6 & times;3 touch input grids, respectively, confirming its practicality and effectiveness as an alternative interface under touchscreen malfunction scenarios.

Accurate channel state information (CSI) at the base station (BS) is crucial for achieving high beamforming gains in massive multiple-input multiple-output (MIMO) systems. In frequency division duplex (FDD) systems, the BS relies on feedback from the users (UEs) to obtain CSI. However, the aggressive CSI compression and quantization at the UE limits the achievement of the optimal beamforming gains. To resolve this problem, we propose an autoencoder network that learns a compact, discrete representation of the channel for variable feedback rates. We train and evaluate our variable-rate autoencoder using a large synthetic dataset, generated through advanced 3D modeling tools and ray-tracing simulators. Simulation results demonstrate that the proposed network achieves more accurate channel reconstruction compared to competitive methods.

The increasing use of unmanned aerial vehicles (UAVs) across multiple domains presents significant opportunities but also raises critical security threats, including unauthorized surveillance and direct attacks. Conventional detection systems such as radar, vision, and acoustic methods face considerable limitations, especially against UAVs flying rapidly at low altitudes. LiDAR technology, offering reliable detection under adverse weather and low-light conditions, has emerged as a promising alternative. However, detecting fast-moving UAVs using LiDAR is challenging due to the sparse and indistinct point clouds generated by high-speed objects. In this study, we experimentally evaluate LiDAR point clouds from UAVs, identify key detection difficulties, and suggest directions for future research toward effective UAV surveillance solutions.

There have been long efforts to refine Wi-Fi carrier sensing (CS) for more aggressive channel access, in pursuit of enhanced network performance. To this end, the recent 802.11ax amendment introduced a preamble detection (PD)-based spatial reuse, allowing concurrent transmissions between adjacent links via adjustable sensitivity levels. Against these conventional ideas, this paper presents a different perspective: Wi-Fi devices already have excessive transmission (TX) opportunities in practice, even without detecting each other under certain scenarios. We shed light on CS anomalies relevant to undetected preambles, which not only cause adjacent devices to transmit concurrently but are also triggered by the new PD-based mechanism, ultimately disrupting its intended operations. Our testbed experiments and in-depth scrutiny reveal the dominant impact of these anomalies on overall network behaviors. Based on these insights, we present two comprehensive frameworks, REFRAIN and AdOPT , to fully exploit TX opportunities enabled by the anomalies and PD-based mechanism respectively, for practical spatial reuse. Prototypes using commercial Wi-Fi devices and NI USRP show the feasibility and effectiveness of our approaches. Extensive simulation results further demonstrate that REFRAIN and AdOPT achieve up to 1.94 x and 1.61 x higher average throughput, only with reduced transmission attempts by half, highlighting their potential to elevate network capacity and efficiency in practical Wi-Fi networks.

Recent advances in electric vehicles, robots, and renewable energy have renewed interest in direct current (DC) power line communication (PLC) technology. However, existing DC power line communication (DC-PLC) systems face several limitations, including low datarates, lack of support for duplex communication, and the absence of an effective medium access control (MAC) mechanism. To overcome these challenges, we propose D-2-PLC, a novel DC-PLC system that features a redesigned physical layer, enabling high-speed duplex communication over a single pair of wires supporting simultaneous power and data transmission. D-2-PLC introduces voltage polarity modulation (VPM) and current amplitude modulation (CAM) for downlink and uplink communication, respectively. In addition, we develop a custom data link layer and MAC protocols to coordinate communication in a bus topology where multiple slave nodes interact with a single master node (the power source), minimizing the risk of collisions. We implement a fully functional prototype and evaluate on a 5-node testbed as well as via 256-node simulations. Results demonstrate that D-2-PLC achieves a maximum datarate of similar to 100 Kb/s, 260% improvement over existing solutions, while maintaining 99+% reliability. These findings highlight D-2-PLC's potential to reduce the cost and weight of battery-powered systems, such as electric vehicles.

Millimeter-wave (mmWave) frequency-modulated continuous-wave (FMCW) radar is increasingly used for sensing in indoor and in-cabin environments. We investigate its ability to detect objects located behind transparent obstacles such as acrylic boards. Experiments were conducted under three scenarios: only the acrylic board, only the target object, and target behind acrylic board. We processed the received signals in MATLAB by applying range FFT, angle FFT, and background subtraction using reference data without objects. The resulting range–angle maps revealed reflections not only from the acrylic board but also from the target placed behind it, confirming that the radar waveform propagated through the transparent material. These findings indicate that mmWave FMCW radar can detect objects beyond transparent obstacles, implying that unintended objects outside the region of interest may also appear in radar images. This effect should be considered in future designs and applications of mmWave radar sensing systems.

Recent advancements in Integrated Sensing and Communication (ISAC) have enabled human activity recognition (HAR) based on channel state information (CSI) from Wi-Fi signals, as a privacy-conscious and cost-effective alternative to conventional vision-based applications. With smart TV platforms increasingly crucial for home automation, Wi-Fi CSI-based HAR offers seamless integration into existing infrastructure. However, in real-world home environments, the close proximity between Wi-Fi AP and smart TV often leads to dominant Line-of-Sight (LOS) signal path, limiting multipath channel diversity and thus the HAR accuracy. To address this, we present HActiFi, an advanced HAR framework that adaptively controls CSI sampling to ensure reliable sensing without overloading the network. Our smartphone-assisted CSI injection method effectively restores recognition accuracy, and adaptive sampling dynamically optimizes throughput, ensuring robust HAR functionality while preserving streaming quality. These findings confirm that Wi-Fi CSI sensing is a scalable and privacy-conscious alternative to vision-based HAR, making it viable for real-world smart home environments.

In this study, we investigate the latency characteristics of commercial 5G networks through empirical measurements and analyze their implications for TCP Cubic congestion control. Our measurement results reveal that 5G latency exhibits variable round-trip times and sudden delay spikes, creating irregular time intervals that may indirectly affect congestion control dynamics. While TCP CUBIC, the default algorithm in Linux, was originally optimized for stable high-bandwidth networks, the observed 5G latency behavior suggests potential challenges such as premature congestion window reductions and throughput inefficiency. Our study highlights how the temporal patterns of 5G latency can influence transport-layer performance, thereby motivating the need for congestion control mechanisms specifically tailored to the variability of 5G networks.

Delay-based congestion control algorithms (CCAs) have been proposed to tackle the bufferbloat problem of traditional loss-based CCAs. However, existing delay-based CCAs either fail to adequately consider the non-congestive delay caused by scheduling characteristics of modern cellular networks leading to improper congestion control, or face practical deployment issues, which results in an inability to fully utilize the high bandwidth and low latency that recent cellular systems provide. To resolve this problem, we propose Cesar, a cellular resource scheduling-aware congestion control with only sender-side modification. Cesar estimates scheduling unit through TCP ACK interval patterns to deduce the scheduling characteristics of the current cellular link, and adjusts the congestion window size based on scheduling unit in a step-wise manner to minimize the impact of the scheduling delay on congestion control. Experimental results on 5G and LTE cellular networks of three different mobile carriers show that Cesar outperforms other state-of-the-art CCAs. Results show that throughput-over-latency performance improves by up to 2.89x, 10.09x, 1.39x, and 5.65x compared to ExLL, PropRate, BBR, and Cubic, respectively.

Concurrent transmission (CTX) has enabled lowpower and low-latency data collection over multihop, demonstrating high efficiency across various wireless technologies such as Zigbee, BLE, and ultra-wideband (UWB) radio. However, most existing CTX-based protocols use logical hop-distancebased scheduling without considering the physical distances or the unique phenomenon of concurrent transmission, which may lead to frequent retransmissions. To further enhance the speed and reliability of data collection, we propose Cupid, an ultra-fast and reliable convergecast for multihop UWB wireless network that leverages in-band ranging, intra-hop grouping, and role differentiation within its collection topology to achieve its goals. Extensive evaluation via real experiments on a 36-node UWB testbed demonstrate that Cupid outperforms Weaver, a state-ofthe-art UWB-based many-to-one data collection protocol, reducing latency by up to similar to 29% while providing higher reliability and throughput across diverse application scenarios.