Fengyuan Zhu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Fengyuan Zhu has a total of 60 co-authors. Here are the top co-authors by number of shared publications:
- Xiaohua Tian
Shanghai Jiao Tong University
22 shared publications - Bingbing Wang
Shanghai Jiao Tong University, Shanghai, China
15 shared publications - Linling Zhong
Shanghai Jiao Tong University, Shanghai, China
11 shared publications - Meng Jin
School of Artificial Intelligence, Shanghai Jiao Tong University, Shanghai, China
10 shared publications - Wenhui Li
Shanghai Jiao Tong University, Shanghai, China
10 shared publications - Zeming Yang
Shanghai Jiao Tong University, Shanghai, China
8 shared publications
Latest Publications
- Open Access
- 26 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 250-264
ORNet: OFDMA-Enabled mmWave Radar ISAC Networks
- 26 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 250-264
Open AccessORNet: OFDMA-Enabled mmWave Radar ISAC Networks
- Open Access
- 26 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 1625-1628
WiP: Orthogonal Time Frequency Space Backscatter for ISAC in High-Mobility Scenarios
- 26 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 1625-1628
Open AccessWiP: Orthogonal Time Frequency Space Backscatter for ISAC in High-Mobility Scenarios
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 25 (10) , 16459-16475
Toward Beamform Charging Without Chicken-or-Egg Dilemma
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 25 (10) , 16459-16475
Toward Beamform Charging Without Chicken-or-Egg Dilemma
- 1 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Networking
- Vol. 34, 5444-5459
Ultra-Low-Power Backscatter for Large-Scale and High-Rate Sensing
This paper presents μTag, an ultra-low-power backscatter sensor that supports high-rate sensing of a large number of targets simultaneously. The core of μTag is an RF “gene editing” technique that embeds both the identity of the sensor and the real-time motion or vibration state of the attached target intensively in the transient features of the sensor’s RF signal, in a collision-resilient manner. We provide practical techniques which i) generate such “genetic signal” with purely analog and extremely simple circuits; and ii) separate the signals from a large scale of sensors reliably. Our experimental results show that our design can support concurrent tracking of 150 targets with a 12 kHz per-tag sampling rate. We also demonstrate with multiple sensing applications that μTag can achieve high-speed and large-scale motion tracking, rotation frequency sensing, and contactless sensing. The PCB power consumption of μTag is 38∼107 μW, according to the operating frequency of the tag. Our ASIC simulation based on the 40 nm CMOS process shows that the power consumption can be further reduced to 0.13∼0.52 μW.
- 1 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Networking
- Vol. 34, 5444-5459
Ultra-Low-Power Backscatter for Large-Scale and High-Rate Sensing
This paper presents μTag, an ultra-low-power backscatter sensor that supports high-rate sensing of a large number of targets simultaneously. The core of μTag is an RF “gene editing” technique that embeds both the identity of the sensor and the real-time motion or vibration state of the attached target intensively in the transient features of the sensor’s RF signal, in a collision-resilient manner. We provide practical techniques which i) generate such “genetic signal” with purely analog and extremely simple circuits; and ii) separate the signals from a large scale of sensors reliably. Our experimental results show that our design can support concurrent tracking of 150 targets with a 12 kHz per-tag sampling rate. We also demonstrate with multiple sensing applications that μTag can achieve high-speed and large-scale motion tracking, rotation frequency sensing, and contactless sensing. The PCB power consumption of μTag is 38∼107 μW, according to the operating frequency of the tag. Our ASIC simulation based on the 40 nm CMOS process shows that the power consumption can be further reduced to 0.13∼0.52 μW.
- 26 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. PP (15361233) , 1-17
DDScatter: Enabling Delay-Doppler Modulation for Mmwave Radar-Backscatter Communication
Millimeter-wave radar backscatter communication (Rad-BackCom) is emerging as a low-power, low-cost enabler for integrated sensing and communication (ISAC) in 6 G ambient IoT. Yet, existing mmWave Rad-BackCom designs face a hard trade-off: Delay-domain modulation offers higher rates but short range, while Doppler-domain modulation extends range but at extremely low throughput. We introduce DDScatter, a dual-function Rad-BackCom system that alleviates this trade-off by jointly exploiting Delay- and Doppler-domain spectral resources for two-dimensional Delay– Doppler (DD) modulation. The transmitted waveform is generated via 2D IDFT mapping, and demodulation is seamlessly performed using the native 2D DFT processing pipeline of mmWave radars. To address severe interference from environmental sensing echoes, we propose an inter-symbol matching filter (ISMF) algorithm that dynamically separates sensing and modulated components, enabling robust demodulation while doubling the SNR of sensing components. A prototype using commercial mmWave radars and COTS tags shows that DDScatter extends range by $4.5\times$ over Delay-domain modulation and more than $2\times$ data rate compared to Doppler-domain modulation, while preserving and enhancing sensing performance. DDScatter opens a path toward scalable, long-range, and high-efficiency ISAC for 6 G ambient IoT.
- 26 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. PP (15361233) , 1-17
DDScatter: Enabling Delay-Doppler Modulation for Mmwave Radar-Backscatter Communication
Millimeter-wave radar backscatter communication (Rad-BackCom) is emerging as a low-power, low-cost enabler for integrated sensing and communication (ISAC) in 6 G ambient IoT. Yet, existing mmWave Rad-BackCom designs face a hard trade-off: Delay-domain modulation offers higher rates but short range, while Doppler-domain modulation extends range but at extremely low throughput. We introduce DDScatter, a dual-function Rad-BackCom system that alleviates this trade-off by jointly exploiting Delay- and Doppler-domain spectral resources for two-dimensional Delay– Doppler (DD) modulation. The transmitted waveform is generated via 2D IDFT mapping, and demodulation is seamlessly performed using the native 2D DFT processing pipeline of mmWave radars. To address severe interference from environmental sensing echoes, we propose an inter-symbol matching filter (ISMF) algorithm that dynamically separates sensing and modulated components, enabling robust demodulation while doubling the SNR of sensing components. A prototype using commercial mmWave radars and COTS tags shows that DDScatter extends range by $4.5\times$ over Delay-domain modulation and more than $2\times$ data rate compared to Doppler-domain modulation, while preserving and enhancing sensing performance. DDScatter opens a path toward scalable, long-range, and high-efficiency ISAC for 6 G ambient IoT.
- Open Access
- 10 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 158-171
Enabling Active Sensing with Zero-Power Components
- 10 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 158-171
Open AccessEnabling Active Sensing with Zero-Power Components
- 22 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. PP (15361233) , 1-15
Hi-MAC: An Intelligent MAC for Heterogeneous Wireless Network
When Internet of Things (IoTs) devices operate in the dense 2.4 GHz band, severe cross-interference with incumbent networks like Wi-Fi degrades total throughput. While cognitive radio (CR) can replace short-term contention with predictive access to improve long-term performance, existing designs ignore system delays, causing decision-execution mismatch and impractical deployment. We propose Hi-MAC, a gateway-assisted spectrum sharing system that uses machine learning for long-term channel access decisions. It generates an opportunity times list for future devices' transmissions, allowing devices to skip outdated slots and use valid ones despite delays. Key contributions include a low-latency end-to-end neural network with incremental learning, a signaling synchronization mechanism, and coordinated end-edge operations to mitigate hidden/exposed nodes. Implemented on software-defined radio-based gateways and IoT devices, Hi-MAC realizes at least 15% higher packet reception rate and 1.42-7.4$times$ better delivery-loss ratio than other access solutions, demonstrating fewer transmission attempts and less interference to coexisting networks. It not only delivers near-optimal own throughput for IoT but also minimizes its impact on the throughput of incumbent networks. In two evaluated cases, the total throughput achieved by Hi-MAC is 131% and 121% of that of other methods with comparable performance. Besides, the prototype performs fast inference only in 1.93 ms, with $\gt $95% time utilization per inference cycle.
- 22 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. PP (15361233) , 1-15
Hi-MAC: An Intelligent MAC for Heterogeneous Wireless Network
When Internet of Things (IoTs) devices operate in the dense 2.4 GHz band, severe cross-interference with incumbent networks like Wi-Fi degrades total throughput. While cognitive radio (CR) can replace short-term contention with predictive access to improve long-term performance, existing designs ignore system delays, causing decision-execution mismatch and impractical deployment. We propose Hi-MAC, a gateway-assisted spectrum sharing system that uses machine learning for long-term channel access decisions. It generates an opportunity times list for future devices' transmissions, allowing devices to skip outdated slots and use valid ones despite delays. Key contributions include a low-latency end-to-end neural network with incremental learning, a signaling synchronization mechanism, and coordinated end-edge operations to mitigate hidden/exposed nodes. Implemented on software-defined radio-based gateways and IoT devices, Hi-MAC realizes at least 15% higher packet reception rate and 1.42-7.4$times$ better delivery-loss ratio than other access solutions, demonstrating fewer transmission attempts and less interference to coexisting networks. It not only delivers near-optimal own throughput for IoT but also minimizes its impact on the throughput of incumbent networks. In two evaluated cases, the total throughput achieved by Hi-MAC is 131% and 121% of that of other methods with comparable performance. Besides, the prototype performs fast inference only in 1.93 ms, with $\gt $95% time utilization per inference cycle.
- 12 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. PP (15361233) , 1-16
Tiga: Autonomous Hybrid Active-Passive Communication for Ambient IoT System
Power consumption continues to pose a fundamental challenge for large-scale IoT tag deployment. While conventional active communication strategies struggle to mitigate peak power demands due to energy-intensive components, passive alternatives achieve microwatt-level consumption but suffer from limited operational range. This paper introduces Tiga, an innovative hybrid active-passive communication system for Ambient IoT that autonomously selects optimal transmission modes. Recognizing the ambiguity inherent in the distance-based mode switching boundary, we instead adopt energy level as our boundary criterion. Tiga employs passive communication when channel energy suffices and seamlessly transitions to active transmission when required. Our design features a novel microwatt-scale dual-function detector that performs concurrent downlink signal demodulation and RSSI computation. We further propose an adaptive four-phase blind boundary search algorithm capable of efficient execution on any low-power controller. We prototype Tiga on a four-layer PCB and evaluate its performance across various environments. Experimental results demonstrate that compared to duty-cycling active communication, Tiga reduces dynamic power consumption by 1.85× to 9.42× across various distances from the gateway.
- 12 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. PP (15361233) , 1-16
Tiga: Autonomous Hybrid Active-Passive Communication for Ambient IoT System
Power consumption continues to pose a fundamental challenge for large-scale IoT tag deployment. While conventional active communication strategies struggle to mitigate peak power demands due to energy-intensive components, passive alternatives achieve microwatt-level consumption but suffer from limited operational range. This paper introduces Tiga, an innovative hybrid active-passive communication system for Ambient IoT that autonomously selects optimal transmission modes. Recognizing the ambiguity inherent in the distance-based mode switching boundary, we instead adopt energy level as our boundary criterion. Tiga employs passive communication when channel energy suffices and seamlessly transitions to active transmission when required. Our design features a novel microwatt-scale dual-function detector that performs concurrent downlink signal demodulation and RSSI computation. We further propose an adaptive four-phase blind boundary search algorithm capable of efficient execution on any low-power controller. We prototype Tiga on a four-layer PCB and evaluate its performance across various environments. Experimental results demonstrate that compared to duty-cycling active communication, Tiga reduces dynamic power consumption by 1.85× to 9.42× across various distances from the gateway.
- 26 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Wireless Communications
- Vol. 25 (15361276) , 12438-12452
Enabling Symbol-Level mmWave Radar-Backscatter Communication
This paper presents mmDFRBC, a symbol-level millimeter-wave (mmWave) backscatter communication system that reuses commercial mmWave FMCW radar infrastructure as the dual-function access point (AP) without hardware modification. We propose M-ary Frequency Shift Modulation (MFSM), a lightweight encoding scheme that enables the tag to modulate information using orthogonal frequency blocks over adjacent chirps, achieving symbol-level modulation with data rates exceeding kbps. To robustly extract the modulation signals from strong radar sensing clutter, we introduce a Coherent Cancellation Demodulation (CCD) method that exploits the coherence difference between sensing signals and modulated reflections. We develop a soft synchronization strategy for operating asynchronously and requiring no synchronization or downconversion circuits at the tag, supporting kbps-level data rates with low power and cost. We implement mmDFRBC using a commercial mmWave radar and verify its performance across static and mobile scenarios, achieving BERs below $10^{-3}$ over 4 meters, with strong resilience to radar clutter interference. mmDFRBC supports data rates up to 50 kbps, highlighting DFRBC’s potential to enable high-performance DFRC systems using existing infrastructure.
- 26 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Wireless Communications
- Vol. 25 (15361276) , 12438-12452
Enabling Symbol-Level mmWave Radar-Backscatter Communication
This paper presents mmDFRBC, a symbol-level millimeter-wave (mmWave) backscatter communication system that reuses commercial mmWave FMCW radar infrastructure as the dual-function access point (AP) without hardware modification. We propose M-ary Frequency Shift Modulation (MFSM), a lightweight encoding scheme that enables the tag to modulate information using orthogonal frequency blocks over adjacent chirps, achieving symbol-level modulation with data rates exceeding kbps. To robustly extract the modulation signals from strong radar sensing clutter, we introduce a Coherent Cancellation Demodulation (CCD) method that exploits the coherence difference between sensing signals and modulated reflections. We develop a soft synchronization strategy for operating asynchronously and requiring no synchronization or downconversion circuits at the tag, supporting kbps-level data rates with low power and cost. We implement mmDFRBC using a commercial mmWave radar and verify its performance across static and mobile scenarios, achieving BERs below $10^{-3}$ over 4 meters, with strong resilience to radar clutter interference. mmDFRBC supports data rates up to 50 kbps, highlighting DFRBC’s potential to enable high-performance DFRC systems using existing infrastructure.