Publications
2026
- Diffusion2: Turning 3D Environments into Radio Frequency HeatmapsIn Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Findings Track, Jun 2026
Modeling radio frequency (RF) signal propagation is essential for understanding the environment, as RF signals offer valuable insights beyond the capabilities of RGB cameras, which are limited by the visible-light spectrum, lens coverage, and occlusions. It is also useful for supporting wireless diagnosis, deployment, and optimization. However, accurately predicting RF signals in complex environments remains a challenge due to interactions with obstacles such as absorption and reflection. We introduce Diffusion2, a diffusion-based approach that uses 3D point clouds to model the propagation of RF signals across a wide range of frequencies, from Wi-Fi to millimeter waves. To effectively capture RF-related features from 3D data, we present the RF-3D Encoder, which encapsulates the complexities of 3D geometry along with signal-specific details. These features undergo multi-scale embedding to simulate the actual RF signal dissemination process. Our evaluation, based on synthetic and real-world measurements, demonstrates that Diffusion2 accurately estimates the behavior of RF signals in various frequency bands and environmental conditions, with an error margin of just 1.9 dB and 27x faster than existing methods, marking a significant advancement in the field.
@inproceedings{XX.XXXX/XXXXXXXXXX, author = {}, title = {Diffusion2: Turning 3D Environments into Radio Frequency Heatmaps}, year = {2026}, isbn = {}, publisher = {}, address = {}, url = {}, doi = {}, booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Findings Track}, month = jun, pages = {}, numpages = {11}, location = {Denver, Colorado, USA}, keywords = {Machine Learning, Computer Vision}, series = {CVPR '26 Findings Track}, } - CIPAT: A Two-stage Configuration Impact Prediction Analysis Toolkit for Cellular NetworksIn ACM Transactions on Modeling and Performance Evaluation of Computing Systems (ACM TOMPECS), Jun 2026
Improving network performance and user experience by tuning network configurations is crucial to cellular service providers (CSPs). However, predicting the performance impacts of configuration changes is non-trivial. The large scale, diversity and complexity of configuration parameters and base station deployments, and more importantly, the uncontrollable external factors (e.g., weather, called latents), lead to confounding effects between configurations and performance metrics. In this paper, we show that the effects of latents can be properly mitigated by considering intermediates, called MAT metrics, which separate the configurations and latents from performance metrics. Then, we propose the CIPAT, a novel two-stage toolkit, driven by a large real-world dataset from live LTE and 5G networks. Our extensive evaluation shows that CIPAT enables network operators to confidently predict the performance impact of candidate configuration settings with an accuracy of up to 86% and an efficacy of up to 85%.
@inproceedings{XX.XXXX/XXXXXXXXXY, author = {Patel, Kartik and Ge, Changhan and Mahimkar, Ajay and Shakkottai, Sanjay and Shaqalle, Yusef}, title = {CIPAT: A Two-stage Configuration Impact Prediction Analysis Toolkit for Cellular Networks}, year = {2026}, isbn = {0000000000000}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, url = {}, doi = {}, booktitle = {ACM Transactions on Modeling and Performance Evaluation of Computing Systems (ACM TOMPECS)}, pages = {1-11}, numpages = {11}, keywords = {Cellular Network, Machine Learning Application}, series = {TOMPECS}, } - Joint Optimization of Handoff and Video Rate in LEO Satellite NetworksIn Proceedings of the 2026 IEEE International Conference on Computer Communications (IEEE INFOCOM ’26), Jun 2026
Low Earth Orbit (LEO) satellite communication presents a promising solution for delivering Internet access to users in remote regions. Given that video content is expected to dominate network traffic in LEO satellite systems, this study presents a new video-aware mobility management framework specifically designed for such networks. By combining simulation models with real-world datasets, we highlight the critical role of handoff strategies and throughput prediction algorithms in both single-user and multi-user video streaming scenarios. Building on these insights, we introduce a suite of innovative algorithms that jointly determine satellite selection and video bitrate to enhance users’ quality of experience (QoE). Initially, we design model predictive control (MPC) and reinforcement learning (RL) based methods for individual users, then extend the approach to manage multiple users sharing a satellite. Notably, we incorporate centralized training with distributed inference in our RL design to develop distributed policies informed by a global view. The effectiveness of our approach is validated through trace-driven simulations and testbed experiments.
@inproceedings{XX.XXXX/XXXXXXXXXZ, author = {Park, Kyoungjun and He, Zhiyuan and Luo, Cheng and Xu, Yi and Qiu, Lili and Ge, Changhan and Muaz, Muhammad and Yang, Yuqing}, title = {Joint Optimization of Handoff and Video Rate in LEO Satellite Networks}, year = {2026}, isbn = {0000000000000}, publisher = {Institute of Electrical and Electronics Engineers}, address = {Piscataway, NJ, USA}, doi = {10.48550/arXiv.2504.04586}, booktitle = {Proceedings of the 2026 IEEE International Conference on Computer Communications (IEEE INFOCOM '26)}, pages = {1-11}, numpages = {11}, keywords = {Networking and Internet Architecture, Satellite Communication}, location = {Tokyo, Japan}, series = {IEEE INFOCOM '26}, }
2024
- Predicting the Performance of Cellular Networks: A Latent-resilient ApproachIn Proceedings of the 30th Annual International Conference on Mobile Computing and Networking (ACM MobiCom ’24), Jun 20243rd Place Winner of ACM Student Research Competition (Graduate Track) at MobiCom ’24, shared with Kartik Patel
Cellular service providers (CSPs) require predicting the network performance for various reasons such as analyzing the impact of planned configuration changes and large-scale events on the network. Although network configurations are widely considered as key predictors of performance, we claim that they are insufficient for accurately predicting cellular network performance. The cellular networks are impacted by unmeasured external factors (e.g., weather, called latents), therefore, the performance prediction based solely on configurations may result in confounding effects. We show that the MAT metrics should be considered in addition as network performance predictors. Using a large dataset collected from a live cellular network, we validate the claim and show the benefit of using MAT metrics for accurate performance prediction.
@inproceedings{10.1145/3636534.3697425, author = {Patel, Kartik and Ge, Changhan and Mahimkar, Ajay and Shakkottai, Sanjay and Shaqalle, Yusef}, title = {Predicting the Performance of Cellular Networks: A Latent-resilient Approach}, year = {2024}, isbn = {979840070489}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, url = {https://doi.org/10.1145/3636534.3697425}, doi = {10.1145/3636534.3697425}, booktitle = {Proceedings of the 30th Annual International Conference on Mobile Computing and Networking (ACM MobiCom '24)}, pages = {7-9}, numpages = {3}, keywords = {Network performance modeling, Network management}, location = {Washington, DC, USA}, series = {ACM MobiCom '24}, } - CIPAT: Latent-resilient Toolkit for Performance Impact Prediction due to Configuration TuningIn Proceedings of the 1st ACM Workshop on Machine Learning for NextG Networks (ACM MLNextG ’24), Jun 2024
Cellular service providers (CSPs) aim to optimize network performance and enhance user experience by tuning network configurations. However, this process often requires continuous live network testing, which incurs significant operational costs. In this paper, we focus on predicting the impact of configuration changes using historical data, thereby reducing the need for live network tests. A key challenge in developing such a model is accounting for unobserved external factors (e.g., weather, referred to as latents) that can introduce confounding effects between configurations and performance metrics. To address this, we employ intermediate network metrics, called MAT metrics, which are influenced by both configurations and latents, and in turn, affect performance metrics. We introduce CIPAT, a novel two-stage toolkit developed using a comprehensive real-world dataset from live LTE networks. Our evaluation demonstrates that CIPAT enables CSPs to predict the performance impact of proposed configuration changes with up to 86% accuracy and 85% efficacy, thereby reducing the operational costs associated with configuration tuning.
@inproceedings{10.1145/3636534.3698246, author = {Patel, Kartik and Ge, Changhan and Mahimkar, Ajay and Shakkottai, Sanjay and Shaqalle, Yusef}, title = {CIPAT: Latent-resilient Toolkit for Performance Impact Prediction due to Configuration Tuning}, year = {2024}, isbn = {979840070489}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, url = {https://doi.org/10.1145/3636534.3698246}, doi = {10.1145/3636534.3698246}, booktitle = {Proceedings of the 1st ACM Workshop on Machine Learning for NextG Networks (ACM MLNextG '24)}, pages = {19-24}, numpages = {6}, keywords = {Cellular Network, Machine Learning Application}, location = {Washington, DC, USA}, series = {ACM MLNextG '24}, } - Optimized Live 4K Video Multicast Streaming on Commodity WiGig DevicesZhaoyuan He, Changhan Ge, Wangyang Li, Lili Qiu, Peijie Li, and Ghufran BaigIn Proceedings of the 44th International Conference on Distributed Computing Systems (IEEE ICDCS ’24), Jun 2024
The popularity of 4K videos is on the rise. However, streaming such high-quality videos over mmWave to several users presents significant challenges due to directional communication, fluctuating channels, and high bandwidth demands. To address these challenges, this paper introduces an innovative 4K layered video multicast streaming system. We (i) develop a video quality model tailored for layered video coding, (ii) optimize resource allocation, scheduling, and beamforming based on the channel conditions of different users, and (iii) design a streaming strategy that integrates fountain code to eliminate redundancy in multicast groups, coupled with a Leaky-Bucket approach for congestion control. We implement our system on commodity-off-the-shelf (COTS) WiGig devices and demonstrate its effectiveness through comprehensive testbed and emulation experiments.
@inproceedings{10631023, author = {He, Zhaoyuan and Ge, Changhan and Li, Wangyang and Qiu, Lili and Li, Peijie and Baig, Ghufran}, title = {Optimized Live 4K Video Multicast Streaming on Commodity WiGig Devices}, year = {2024}, isbn = {9798350386059}, publisher = {IEEE}, address = {Piscataway, NJ, USA}, url = {https://doi.org/10.1109/ICDCS60910.2024.00108}, doi = {10.1109/ICDCS60910.2024.00108}, booktitle = {Proceedings of the 44th International Conference on Distributed Computing Systems (IEEE ICDCS '24)}, pages = {1131-1142}, numpages = {12}, keywords = {Video streaming, Millimeter-wave, Multicast}, location = {Jersey City, New Jersey, USA}, series = {IEEE ICDCS '24}, }
2023
- 2ACE: Spectral Profile-Driven Multi-Resolutional Compressive Sensing for MmWave Channel EstimationYiwen Song, Changhan Ge, Lili Qiu, and Yin ZhangIn Proceedings of the 24th International Symposium on Theory, Algorithmic Foundations, and Protocol Design for Mobile Networks and Mobile Computing (ACM MobiHoc ’23), Jun 2023
Channel estimation is critical to millimeter-wave capability. Unlike sub-6 GHz WiFi, commercial-off-the-shelf 60 GHz WiFi devices adopt a single RF-chain and can only report the combined received signal strength (RSS) instead of the antenna-wise channel state information (CSI). Therefore, recovering the CSI using a limited number of RSS measurements is important but faces the following challenges: (i) solving a non-convex objective is hard and computationally heavy, (ii) the estimation error is high with insufficient RSS measurements, and (iii) channel fluctuates dynamically. To jointly tackle them, we propose 2ACE, an Accelerated and Accurate Channel Estimation approach using spectral profile-driven multiresolutional compressive sensing. Our thorough experiments show that 2ACE yields 2–8 dB reduction in CSI estimation error, 1–5 dB improvement in beamforming performance, and 5° - 10° reduction in angle-of-departure estimation error over the existing schemes.
@inproceedings{10.1145/3565287.3610252, author = {Song, Yiwen and Ge, Changhan and Qiu, Lili and Zhang, Yin}, title = {2ACE: Spectral Profile-Driven Multi-Resolutional Compressive Sensing for MmWave Channel Estimation}, year = {2023}, isbn = {9781450399265}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, url = {https://doi.org/10.1145/3565287.3610252}, doi = {10.1145/3565287.3610252}, booktitle = {Proceedings of the 24th International Symposium on Theory, Algorithmic Foundations, and Protocol Design for Mobile Networks and Mobile Computing (ACM MobiHoc '23)}, pages = {41–50}, numpages = {10}, keywords = {wireless sensing, channel estimation, millimeter-wave}, location = {Washington, DC, USA}, series = {ACM MobiHoc '23}, } - Chroma: Learning and Using Network Contexts to Reinforce Performance Improving ConfigurationsIn Proceedings of the 29th Annual International Conference on Mobile Computing and Networking (ACM MobiCom ’23), Jun 2023
Managing network configuration and improving service experience effectively is essential for cellular service providers (CSPs). This is challenging because of cellular networks’ large scale and complexity, the wide variety of configuration parameters, and the performance impact tradeoffs resulting across multiple metrics and geographical locations. This paper focuses on learning and using network contexts to recommend performance-improving configurations. While learning contexts, one must carefully account for the configuration parameter dependency, performance impact confusion that can arise due to co-occurring unrelated changes, and uneven change deployment distribution across locations. We present a new solution Chroma that addresses the above challenges. Using real-world data collected from a large operational LTE and 5G cellular service provider, we thoroughly evaluate and demonstrate the efficacy of Chroma. We successfully trial Chroma on an operational cellular network and highlight its benefits in practical settings.
@inproceedings{10.1145/3570361.3613256, author = {Ge, Changhan and Ge, Zihui and Liu, Xuan and Mahimkar, Ajay and Shaqalle, Yusef and Xiang, Yu and Pathak, Shomik}, title = {Chroma: Learning and Using Network Contexts to Reinforce Performance Improving Configurations}, year = {2023}, isbn = {9781450399906}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, url = {https://doi.org/10.1145/3570361.3613256}, doi = {10.1145/3570361.3613256}, booktitle = {Proceedings of the 29th Annual International Conference on Mobile Computing and Networking (ACM MobiCom '23)}, articleno = {42}, numpages = {16}, pages = {628-643}, keywords = {performance impact analysis, context learning, network configuration recommendation, configuration change clustering}, location = {Madrid, Spain}, series = {ACM MobiCom '23}, }
2022
- Extracting and Predicting Multipath Profiles under High MobilityGhufran Baig, Changhan Ge, Lili Qiu, Yuanjie Li, Wangyang Li, Wei Sun, Jian He, Zhehui Zhang, and Songwu LuIn Proceedings of the 23rd International Symposium on Theory, Algorithmic Foundations, and Protocol Design for Mobile Networks and Mobile Computing (ACM MobiHoc ’22), Jun 2022
The wireless signal propagates via multipath arising from different reflections and penetration between a transmitter and receiver. Extracting multipath profiles (e.g., delay and Doppler along each path) from received signals enables many important applications, such as channel prediction and crossband channel estimation (i.e., estimating the channel on a different frequency). The benefit of multipath estimation further increases with mobility since the channel in that case is less stable and more important to track. Yet high-speed mobility poses significant challenges to multipath estimation. In this paper, instead of using time-frequency domain channel representation, we leverage the delay-Doppler domain representation to accurately extract and predict multipath properties. Specifically, we use impulses in the delay-Doppler domain as pilots to estimate the multipath parameters and apply the multipath information to predicting wireless channels as an example application. Our design rationale is that mobility is more predictable than the wireless channel since mobility has inertial while the wireless channel is the outcome of a complicated interaction between mobility, multipath, and noise. We evaluate our approach via both acoustic and RF experiments, including vehicular experiments using USRP. Our results show that the estimated multipath matches the ground truth, and the resulting channel prediction is more accurate than the traditional channel prediction schemes.
@inproceedings{10.1145/3492866.3549710, author = {Baig, Ghufran and Ge, Changhan and Qiu, Lili and Li, Yuanjie and Li, Wangyang and Sun, Wei and He, Jian and Zhang, Zhehui and Lu, Songwu}, title = {Extracting and Predicting Multipath Profiles under High Mobility}, year = {2022}, isbn = {9781450391658}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, url = {https://doi.org/10.1145/3492866.3549710}, doi = {10.1145/3492866.3549710}, booktitle = {Proceedings of the 23rd International Symposium on Theory, Algorithmic Foundations, and Protocol Design for Mobile Networks and Mobile Computing (ACM MobiHoc '22)}, pages = {181–190}, numpages = {10}, location = {Seoul, Republic of Korea}, series = {ACM MobiHoc '22}, }
2019
- Millimeter-wave Architectures for Automated Vehicles: An Experiment-Driven ExplorationChanghan Ge, Xiao Sai, Andrew Yoo, Zhuolun Zhou, and Xinyu ZhangIn Proprietary Technical Report of Sony Focused Research Project 2017, Jun 2019