Bo Lin, Huanming Zhang, Yuhua Jiang, Yucong Wang, Tengyu Zhang, Shaoqiang Yan, Hongyao Li, Yihong Liu, Feifei Gao. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Leveraging cameras and LiDARs has been proved as an effective way to achieve beam management without training overhead for millimeter wave (mmWave) communications. However, existing methods place sensors at base station (BS) and/or mobile station (MS), which may induce privacy concerns and augment communications system expenses. In this paper, we propose a novel third-party camera aided mmWave beam management framework and self-build an RFSoC prototype for validation. Specifically, we design third-party camera aided beam alignment and blockage prediction algorithms that could work with random third-party perspective. Then, we leverage the proposed beam alignment to design initial access and beam recovery, and utilize the proposed blockage prediction to realize failure prediction. Hand-off from mmWave to sub-6G communications is conducted once the blockage is predicted. To validate the proposed framework, we develop an RFSoC prototype from scratch independently. The real-world real-time experimental results show that the beam alignment achieves over 98% in top-5 accuracy and reduces the time consumption to below$1/50$of that incurred by exhaustive beam sweeping. Meanwhile, the prototype maintains 410 MHz dynamic mmWave communications, and can seamlessly switch to sub-6G communications before a mmWave blockage happens.

Datacenters need to reduce embodied carbon emissions, particularly for flash, which accounts for 40% of embodied carbon in servers. However, decreasing flash’s embodied emissions is challenging due to flash’s limited write endurance, which more than halves with each generation of denser flash. Reducing embodied emissions requires extending flash lifetime, stressing its limited write endurance even further. The legacy Logical Block-Addressable Device (LBAD) interface exacerbates the problem by forcing devices to perform garbage collection, leading to even more writes. Flash-based caches in particular write frequently, limiting the lifetimes and densities of the devices they use. These flash caches illustrate the need to break away from LBAD and switch to the new Write-Read-Erase iNterfaces (WREN) now coming to market. WREN affords applications control over data placement and garbage collection. We present Fairy Wren , 1 a flash cache designed for WREN. Fairy Wren reduces writes by co-designing caching policies and flash garbage collection. Fairy Wren provides a 12.5× write reduction over state-of-the-art LBAD caches. This decrease in writes allows flash devices to last longer, decreasing flash cost by 35% and flash carbon emissions by 33%.

In this paper, we design a radio frequency system on chip (RFSoC) based scalable integrated sensing and communication (ISAC) prototyping platform with orthogonal time frequency space (OTFS) waveform. Within the ISAC prototype, we conceive an efficient frame structure tailored to streamline the communication and sensing schemes. The prototype utilizes the low-complexity channel equalization algorithm in the time-delay domain and senses the targets using the synchronization preamble and the peak pilot. The RFSoC based platform employs a hybrid architecture, which can be scaled to accommodate di-verse OTFS waveform configurations, facilitating the verification of a wide range of OTFS algorithms. We test the prototyping platform under both static and dynamic scenarios, and the experimental outcomes affirm the prototype's competence in accurately demodulating transmitted data. Moreover, the sensing function attains a velocity resolution of up to 1.65$m$/$S$and a range resolution of up to 0.5 m.

In this paper, we propose a practical integrated sensing and communications (ISAC) framework to sense dynamic targets from clutter environment while ensuring users communications quality. We design multiple communications beams that can communicate with users while one rotating sensing beam can scan entire space, and then we propose the supporting beam-forming design and power allocation strategies for such design. Unlike most existing ISAC studies that ignore the interference of static environmental clutter on target sensing, we construct a mixed sensing channel that includes both static environment and dynamic targets. When base station receives echo signals, we first provide a practical clutter filtering method to filter out static environmental clutter. Then dynamic target detection and angle estimation are realized through angle-Doppler spectrum estimation (ADSE) and joint detection over multiple subcarriers (MSJD), while distance and velocity estimation are realized through the extended subspace algorithm. Simulation results are provided to demonstrate the effectiveness of the proposed scheme.

Leveraging various sensors, i.e., cameras and Li-DARs, is deemed as a promising way to realize fast millimeter wave (mmWave) beam alignment with no frequency overhead. However, previous methods that place extra sensors at base station (BS) and mobile station (MS) may augment communication system expenses and also induce privacy concerns. In this paper, we propose a novel beam alignment framework that utilizes images taken by camera placed at third-party perspective. We design a deep neural network that extracts and fuses the position and orientation of mobile user in the third-party images to infer the optimal beam pair, and we establish real-world datasets to train and evaluate the proposed network. Specifically, we design a mmWave light-weight beam sweeping system to reduce time expenses and system complexity during dataset collection. To calculate the time consumption of the third-party camera aided beam alignment framework, we independently implement a mmWave communication prototype with 410MHz bandwidth OFDM baseband according to IEEE 802.11 protocol on Xilinx RFSoC. The experimental results show the third-party camera aided beam alignment approach achieves over 98% in top-5 accuracy, and reduces the mmWave beam alignment time consumption to below 1/50 of that incurred by exhaustive beam sweeping.

In this paper, we propose a practical integrated sensing and communications (ISAC) framework to sense dynamic targets from clutter environment while ensuring users communications quality. To implement communications function and sensing function simultaneously, we design multiple communications beams that can communicate with the users as well as one sensing beam that can rotate and scan the entire space. To minimize the interference of sensing beam on existing communications systems, we divide the service area intosensing beam for sensing (S4S) sectorandcommunications beam for sensing (C4S) sector, and provide beamforming design and power allocation optimization strategies for each type sector. Unlike most existing ISAC studies that ignore the interference of static environmental clutter on target sensing, we construct a mixed sensing channel model that includes both static environment and dynamic targets. When base station receives the echo signals, it first filters out the interference from static environmental clutter and extracts the effective dynamic target echoes. Then a complete and practical dynamic target sensing scheme is designed to detect the presence of dynamic targets and to estimate their angles, distances, and velocities. In particular, dynamic target detection and angle estimation are realized through angle-Doppler spectrum estimation (ADSE) and joint detection over multiple subcarriers (MSJD), while distance and velocity estimation are realized through the extended subspace algorithm. Simulation results demonstrate the effectiveness of the proposed scheme and its superiority over the existing methods that ignore environmental clutter.

In this paper, we propose a novel computer vision-based approach to aid reconfigurable intelligent surface (RIS) for dynamic beam tracking and implement the corresponding prototype verification system. A camera is attached at the RIS to obtain the visual information about the surroundings, with which RIS identifies the incident beam direction and the desired reflected beam direction, and then adjusts the reflection coefficients according to the pre-designed codebooks. We build a 20-by-20 RIS running at 5.4 GHz and develop a high-speed control board to ensure the real-time refresh of the reflection coefficients. Meanwhile we implement an independent peer-to-peer communications system to simulate the physical link between the base station and the user equipment. The vision-aided RIS prototype system is tested in two mobile scenarios: RIS works in near-field conditions as a passive array antenna of the base station; RIS works in far-field conditions to assist the communication between the base station and the user equipment. The experimental results show that RIS can quickly adjust the reflection coefficients for dynamic beam tracking with the help of visual information.

With the increasing demand for real-time system monitoring and tracking in various contexts, the amount of time-stamped event data grows at an astonishing rate. Analytics on time-stamped events must be real time and the aggregated results need to be accurate even when data arrives out of order. Unfortunately, frequent occurrences of out-of-order data will significantly slow down the processing, and cause a large delay in the query response. Timon is a timestamped event database that aims to support aggregations and handle late arrivals both correctly (i.e., upholding the exactly-once semantics) and efficiently. Our insight is that a broad range of applications can be implemented with data structures and corresponding operators that satisfy associative and commutative properties. Records arriving after the low watermark are appended to Timon directly, allowing aggregations to be performed lazily. To improve query efficiency, Timon maintains a TS-LSM-Tree, which keeps the most recent data in memory and contains a time-partitioning tree on disk for high-volume data accumulated over long time span. Besides, Timon supports materialized aggregation views and correlation analysis across multiple streams. Timon has been successfully deployed at Alibaba Cloud and is a critical building block for Alibaba cloud's continuous monitoring and anomaly analysis infrastructure.