Bingcong Lu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Bingcong Lu has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Li Song
Shanghai Jiao Tong University, Shanghai, China
7 shared publications - Wenjun Zhang
Shanghai Jiao Tong University
5 shared publications - Jun Xu
School of Information Science and Electronic Engineering, Shanghai Jiao Tong University, Shanghai, China
4 shared publications - Rong Xie
School of Information Science and Electronic Engineering, Shanghai Jiao Tong University, Shanghai, China
4 shared publications - Zhengxue Cheng
School of Information Science and Electronic Engineering, Shanghai Jiao Tong University, Shanghai, China
3 shared publications - Keyu Wang
SHANGHAI JIAOTONG UNIVERSITY, Shanghai, China
2 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 19-28
DeformableGrasp: Efficient Dexterous Grasping of Deformable Objects via Reinforcement Learning
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 19-28
Open AccessDeformableGrasp: Efficient Dexterous Grasping of Deformable Objects via Reinforcement Learning
- Open Access
- 11 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 730-742
Adaptive Bitrate Live Streaming over HTTP-FLV: A Practical System Perspective
- 11 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 730-742
Open AccessAdaptive Bitrate Live Streaming over HTTP-FLV: A Practical System Perspective
- 6 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Broadcasting
- Vol. PP (00189316) , 1-15
GenRTC: Generative Real-Time Video Conferencing via Joint Adaptive Coding and Bandwidth Estimation
Real-time video broadcasting has become increasingly prevalent, enabling wide-scale distribution of video content to diverse audiences. However, network conditions often remain volatile, particularly in bandwidth-constrained environments such as subways or congested public areas, where both broadcasting and real-time communication (RTC) systems struggle to maintain optimal performance. In this paper, we propose GenRTC, a generative video conferencing system designed to mitigate bandwidth constraints by integrating joint scalable generative face coding and network-aware adaptation. This approach significantly improves the system’s robustness against network volatility, guaranteeing low-latency and high-fidelity video conferencing across a broad spectrum of bandwidth scenarios. GenRTC integrates conventional video codecs and generative face video encoding within a latency–quality joint adaptive codec controller, which dynamically selects the most appropriate encoding scheme to maximize perceptual quality under the latency constraints of real-time applications across diverse network conditions. In addition, GenRTC incorporates a bandwidth prediction module that delivers accurate and temporally stable low-bandwidth estimation. Experimental results demonstrate that GenRTC achieves a good latency-quality trade-off. GenRTC reduces frame timeouts from 60.4% to 5.4% on average under 200 Kbps. At a similar missrate, it improves PSNR by 7.2 dB. Specifically, at 20 Kbps, GenRTC achieves an 8% missrate and 32dB PSNR, while traditional RTC systems cannot operate at such ultra-low bandwidth. To facilitate a more intuitive comparison of perceptual quality, demo videos are provided at https://github.com/IreneLu12/GenRTC
- 6 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Broadcasting
- Vol. PP (00189316) , 1-15
GenRTC: Generative Real-Time Video Conferencing via Joint Adaptive Coding and Bandwidth Estimation
Real-time video broadcasting has become increasingly prevalent, enabling wide-scale distribution of video content to diverse audiences. However, network conditions often remain volatile, particularly in bandwidth-constrained environments such as subways or congested public areas, where both broadcasting and real-time communication (RTC) systems struggle to maintain optimal performance. In this paper, we propose GenRTC, a generative video conferencing system designed to mitigate bandwidth constraints by integrating joint scalable generative face coding and network-aware adaptation. This approach significantly improves the system’s robustness against network volatility, guaranteeing low-latency and high-fidelity video conferencing across a broad spectrum of bandwidth scenarios. GenRTC integrates conventional video codecs and generative face video encoding within a latency–quality joint adaptive codec controller, which dynamically selects the most appropriate encoding scheme to maximize perceptual quality under the latency constraints of real-time applications across diverse network conditions. In addition, GenRTC incorporates a bandwidth prediction module that delivers accurate and temporally stable low-bandwidth estimation. Experimental results demonstrate that GenRTC achieves a good latency-quality trade-off. GenRTC reduces frame timeouts from 60.4% to 5.4% on average under 200 Kbps. At a similar missrate, it improves PSNR by 7.2 dB. Specifically, at 20 Kbps, GenRTC achieves an 8% missrate and 32dB PSNR, while traditional RTC systems cannot operate at such ultra-low bandwidth. To facilitate a more intuitive comparison of perceptual quality, demo videos are provided at https://github.com/IreneLu12/GenRTC
- Open Access
- 15 April 2024
- Published by Association for Computing Machinery (ACM)
- p. 306-312
Pioneer
For Real-time Communication (RTC), Bandwidth Estimation (BWE) is crucial for enhancing user Quality of Experience (QoE) by ensuring efficient bandwidth utilization and low latency. Recent advancements have shifted towards machine learning based algorithms, particularly online reinforcement leanring (RL), to dynamically infer future bandwidth using statistical data. However, challenges such as dependency on training settings, the necessity for extensive trial and error, and instability in complex state spaces hinder their efficacy. To address these limitations, we propose Pioneer, a novel offline RL framework for BWE in RTC systems. Unlike its predecessors, Pioneer eliminates the need for real-time environment interaction during training and achieves good performance through lightweight training. Our framework consists of a Trajectory Sampler for state information preprocessing and a Bandwidth Estimator based on offline RL model. Our test results on offline datasets show that Pioneer can achieve better performance than expert algorithms. We also tested Pioneer on online simulation platforms, and Pioneer can improve QoE by 9% compared to other offline algorithm, demonstrating good robustness.
- 15 April 2024
- Published by Association for Computing Machinery (ACM)
- p. 306-312
Open AccessPioneer
For Real-time Communication (RTC), Bandwidth Estimation (BWE) is crucial for enhancing user Quality of Experience (QoE) by ensuring efficient bandwidth utilization and low latency. Recent advancements have shifted towards machine learning based algorithms, particularly online reinforcement leanring (RL), to dynamically infer future bandwidth using statistical data. However, challenges such as dependency on training settings, the necessity for extensive trial and error, and instability in complex state spaces hinder their efficacy. To address these limitations, we propose Pioneer, a novel offline RL framework for BWE in RTC systems. Unlike its predecessors, Pioneer eliminates the need for real-time environment interaction during training and achieves good performance through lightweight training. Our framework consists of a Trajectory Sampler for state information preprocessing and a Bandwidth Estimator based on offline RL model. Our test results on offline datasets show that Pioneer can achieve better performance than expert algorithms. We also tested Pioneer on online simulation platforms, and Pioneer can improve QoE by 9% compared to other offline algorithm, demonstrating good robustness.
- Open Access
- 18 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-6
SAFR: A Real-Time Communication System with Adaptive Frame Rate
Real-Time Communication (RTC) systems tend to drop some frames during transmission to cope with network congestion. Though frame-dropping is helpful to low-delay performance, it may cause sudden increase in frame loss and noticeable Quality of Experience (QoE) degradation. In this paper, we propose a novel RTC system, SAFR (System of Adaptive Frame Rate), to jointly optimize all the key indicators of QoE performance. Firstly, a Frame Rate Controller (FRC) is designed for making reasonable frame-rate alteration decisions. Then, a reinforcement learning-based bandwidth prediction module, named Target Rate Controller (TRC), is established to improve the bandwidth utilization by cooperating with FRC. Finally, an AI-based frame interpolation module is applied on the receiver side to mitigate the frame-rate degradation. Compared to traditional RTC systems, SAFR achieves 16.3%-26.7% frame-delay reductions, 61.8%-69.9% frame loss rate decreases, 6.4%-11.3% QoE improvements, with up to 25.3% savings of bandwidth at the same time. Especially, SAFR shows greater superiority under worse network conditions. The supplementary materials[10] further discuss the settings of hyper-parameters and pseudo-codes of FRC. Our test demos are available at: https://github.com/xiaosayin/SAFR.
- 18 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-6
Open AccessSAFR: A Real-Time Communication System with Adaptive Frame Rate
Real-Time Communication (RTC) systems tend to drop some frames during transmission to cope with network congestion. Though frame-dropping is helpful to low-delay performance, it may cause sudden increase in frame loss and noticeable Quality of Experience (QoE) degradation. In this paper, we propose a novel RTC system, SAFR (System of Adaptive Frame Rate), to jointly optimize all the key indicators of QoE performance. Firstly, a Frame Rate Controller (FRC) is designed for making reasonable frame-rate alteration decisions. Then, a reinforcement learning-based bandwidth prediction module, named Target Rate Controller (TRC), is established to improve the bandwidth utilization by cooperating with FRC. Finally, an AI-based frame interpolation module is applied on the receiver side to mitigate the frame-rate degradation. Compared to traditional RTC systems, SAFR achieves 16.3%-26.7% frame-delay reductions, 61.8%-69.9% frame loss rate decreases, 6.4%-11.3% QoE improvements, with up to 25.3% savings of bandwidth at the same time. Especially, SAFR shows greater superiority under worse network conditions. The supplementary materials[10] further discuss the settings of hyper-parameters and pseudo-codes of FRC. Our test demos are available at: https://github.com/xiaosayin/SAFR.
- 4 August 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Configurable Low Delay Congestion Control Scheme for Cellular Networks
Loss-based TCP congestion control schemes suffer from high self-inflicted delay or low bandwidth utilization in cellular networks. To solve the problem, many delay-based algorithms are proposed to realize high throughput with the lowest possible delay. However, there is no need to achieve both two goals, as most delay-sensitive applications only need to maximize throughput while keeping the average delay below the predefined target delay to provide a high quality of experience (QoE). Thus, we propose CLDCC, a configurable low delay congestion control scheme. CLDCC estimates the RTT and RTT gradient in cellular networks as the congestion signals. Then it regulates congestion window size dynamically to keep its average delay below the target value. We test CLDCC in the trace-driven environment and compare it with many state-of-the-art schemes such as LEDBAT and BBR. The results show that CLDCC reduces 8.2% delay compared with LEDBAT, which is also designed for keeping average delay below the target value. And CLDCC can reduce 12.6 % delay with only 2.8 % throughput degradation in contrast with BBR.
- 4 August 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Configurable Low Delay Congestion Control Scheme for Cellular Networks
Loss-based TCP congestion control schemes suffer from high self-inflicted delay or low bandwidth utilization in cellular networks. To solve the problem, many delay-based algorithms are proposed to realize high throughput with the lowest possible delay. However, there is no need to achieve both two goals, as most delay-sensitive applications only need to maximize throughput while keeping the average delay below the predefined target delay to provide a high quality of experience (QoE). Thus, we propose CLDCC, a configurable low delay congestion control scheme. CLDCC estimates the RTT and RTT gradient in cellular networks as the congestion signals. Then it regulates congestion window size dynamically to keep its average delay below the target value. We test CLDCC in the trace-driven environment and compare it with many state-of-the-art schemes such as LEDBAT and BBR. The results show that CLDCC reduces 8.2% delay compared with LEDBAT, which is also designed for keeping average delay below the target value. And CLDCC can reduce 12.6 % delay with only 2.8 % throughput degradation in contrast with BBR.