Lishuo Deng
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Lishuo Deng has a total of 21 co-authors. Here are the top co-authors by number of shared publications:
- Weiwei Shan
Southeast University
7 shared publications - Junyi Qian
School of Integrated Circuits, Southeast University, Nanjing, China
5 shared publications - Cai Li
School of Integrated Circuits, Southeast University, Nanjing, China
4 shared publications - Zhangrui Qian
National ASIC Center, School of Integrated Circuit, Southeast University, Nanjing, China
3 shared publications Zhihao Yan
School of Integrated Circuits, Southeast University, Nanjing, China
3 shared publications- Yuxuan Du
National ASIC Center, Southeast University, Nanjing, China
2 shared publications
Latest Publications
- 14 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Solid-State Circuits
- Vol. PP (00189200) , 1-12
A BEV Perception Transformer Accelerator With Saliency-Driven Image/Point Cloud Fusion and Phase-Linked Dataflow in 28 nm CMOS
Deploying advanced Transformer-based models for real-time, high-accuracy multimodal bird’s-eye-view (BEV) perception in autonomous driving imposes substantial hardware demands. To address this, we propose a low-cost, low-power image/point-cloud fusion Transformer accelerator that supports two modes: high-performance driving and ultra-low-power sentry operation. We first propose a cross-modal saliency evaluation mechanism that prunes redundant computation. We then propose a column-wise computing array with a three-level adaptive-precision strategy to enable saliency-driven fusion. Furthermore, we propose a phase-linked dataflow that optimizes on-chip data reuse and an on-chip data-forwarding method that reduces external memory access (EMA). Fabricated in 28nm CMOS, this chip achieves the best perception accuracy on the nuScenes dataset, surpassing prior state-of-the-art (SOTA) accelerators. This is the first BEV perception Transformer accelerator to support a dual-mode architecture of high-performance driving and low-power sentry mode, sustaining an active power of 26.6mW and a long-term average power of only 1.68mW in sparse-activity scenarios. In driving mode, it delivers a peak frame rate of 15.89FPS and a peak energy efficiency of 6.11TOPS/W. Overall, this chip outperforms SOTA accelerators in accuracy, frame rate, and energy efficiency.
- 14 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Solid-State Circuits
- Vol. PP (00189200) , 1-12
A BEV Perception Transformer Accelerator With Saliency-Driven Image/Point Cloud Fusion and Phase-Linked Dataflow in 28 nm CMOS
Deploying advanced Transformer-based models for real-time, high-accuracy multimodal bird’s-eye-view (BEV) perception in autonomous driving imposes substantial hardware demands. To address this, we propose a low-cost, low-power image/point-cloud fusion Transformer accelerator that supports two modes: high-performance driving and ultra-low-power sentry operation. We first propose a cross-modal saliency evaluation mechanism that prunes redundant computation. We then propose a column-wise computing array with a three-level adaptive-precision strategy to enable saliency-driven fusion. Furthermore, we propose a phase-linked dataflow that optimizes on-chip data reuse and an on-chip data-forwarding method that reduces external memory access (EMA). Fabricated in 28nm CMOS, this chip achieves the best perception accuracy on the nuScenes dataset, surpassing prior state-of-the-art (SOTA) accelerators. This is the first BEV perception Transformer accelerator to support a dual-mode architecture of high-performance driving and low-power sentry mode, sustaining an active power of 26.6mW and a long-term average power of only 1.68mW in sparse-activity scenarios. In driving mode, it delivers a peak frame rate of 15.89FPS and a peak energy efficiency of 6.11TOPS/W. Overall, this chip outperforms SOTA accelerators in accuracy, frame rate, and energy efficiency.
- 16 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. PP (15498328) , 1-13
A 2.2- μ W Keyword Spotting System With Filterbank Learning and Attention-Aware Soft-Threshold Denoising in 28-nm CMOS
Keyword spotting (KWS) is crucial for speech interactions in mobile devices, yet its widespread adoption is constrained by the trade-off between low power and high accuracy in diverse noisy environments. Here, we present an ultra-low-power and robust KWS chip to address this challenge through architecture and circuit co-optimization. First, a learnable Mel filter bank is proposed and co-designed with a group-based time-multiplexed architecture to perform adaptive spectral reshaping in a log-Mel feature extractor, reducing the downstream network computations by 25.8% while preserving feature representation capability. Second, a broadcasted residual network (BC-ResNet) is designed to support both KWS and speaker verification (SV) classification tasks, which reduces additions by 43.5% and parameters by 5% compared to a lightweight depthwise separable CNN (DSCNN). Third, an attention-aware soft-threshold denoising strategy is developed to dynamically filter noisy features, yielding up to a 5.8% increase in accuracy over the baseline without denoising at 0dB signal-to-noise ratio (SNR). Finally, a multiplier-less processing element (PE) array combined with stacked-transistor-based ultra-low-leakage custom memory is designed, reducing hardware area by 31.3% and leakage power by 20.4%. Fabricated in a 28nm CMOS process, the proposed 12-class KWS chip consumes only $2.2\mu $ W at 0.42V supply. It achieves 87.6%-93.5% accuracy on GSCD across 0dB to 20dB SNR levels, demonstrating high robustness for diverse scenarios.
- 16 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. PP (15498328) , 1-13
A 2.2- μ W Keyword Spotting System With Filterbank Learning and Attention-Aware Soft-Threshold Denoising in 28-nm CMOS
Keyword spotting (KWS) is crucial for speech interactions in mobile devices, yet its widespread adoption is constrained by the trade-off between low power and high accuracy in diverse noisy environments. Here, we present an ultra-low-power and robust KWS chip to address this challenge through architecture and circuit co-optimization. First, a learnable Mel filter bank is proposed and co-designed with a group-based time-multiplexed architecture to perform adaptive spectral reshaping in a log-Mel feature extractor, reducing the downstream network computations by 25.8% while preserving feature representation capability. Second, a broadcasted residual network (BC-ResNet) is designed to support both KWS and speaker verification (SV) classification tasks, which reduces additions by 43.5% and parameters by 5% compared to a lightweight depthwise separable CNN (DSCNN). Third, an attention-aware soft-threshold denoising strategy is developed to dynamically filter noisy features, yielding up to a 5.8% increase in accuracy over the baseline without denoising at 0dB signal-to-noise ratio (SNR). Finally, a multiplier-less processing element (PE) array combined with stacked-transistor-based ultra-low-leakage custom memory is designed, reducing hardware area by 31.3% and leakage power by 20.4%. Fabricated in a 28nm CMOS process, the proposed 12-class KWS chip consumes only $2.2\mu $ W at 0.42V supply. It achieves 87.6%-93.5% accuracy on GSCD across 0dB to 20dB SNR levels, demonstrating high robustness for diverse scenarios.
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Solid-State Circuits
- p. 1-13
VITAMIN: A Variation-Tolerant Clock Duty-Cycle Control System to Mitigate Aging-Induced Clock Duty-Cycle Distortion and V min Degradation With Monitoring-Enabled Adjuster and Path Replica
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Solid-State Circuits
- p. 1-13
VITAMIN: A Variation-Tolerant Clock Duty-Cycle Control System to Mitigate Aging-Induced Clock Duty-Cycle Distortion and V min Degradation With Monitoring-Enabled Adjuster and Path Replica
- 17 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Solid-State Circuits Letters
- Vol. 9, 25-28
Actiniaria: Distributed Dynamic-IR-Drop-Aware Timing Monitor for AVFS With Lightweight Tentacles
Advances in integrated circuit (IC) technology have amplified the effects of process, voltage, and temperature (PVT) variations, particularly dynamic IR drop, which severely affects timing. Post-silicon IR drop monitoring circuits are lacking, forcing designers to reserve substantial static guard bands for worst-case scenarios, compromising energy efficiency. Inspired by biomimetics, this paper proposes an Actiniaria-shaped structure for distributed monitoring of multiple IR drop hotspots without invading critical paths (CPs). It comprises “lightweight tentacles + central monitor”. The tentacles use voltage-sensitive cells to sense the impact of IR drop on timing, while also supporting adaptive tuning for global variations. By using on-chip PVT sensors, Actiniaria tracks the timing characteristics of the longest CP in real time under PVT variations. A two-stage pre-warning timing monitor captures timing margins and employs adaptive voltage/frequency scaling (AVFS) strategies to compress redundant guard bands. When applied to an open-source multi-core RISC-V processor fabricated in 22-nm CMOS, Actiniaria achieves a 40.6% reduction in power consumption through non-invasive dynamic IR drop monitoring while having only 0.08% area overhead.
- 17 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Solid-State Circuits Letters
- Vol. 9, 25-28
Actiniaria: Distributed Dynamic-IR-Drop-Aware Timing Monitor for AVFS With Lightweight Tentacles
Advances in integrated circuit (IC) technology have amplified the effects of process, voltage, and temperature (PVT) variations, particularly dynamic IR drop, which severely affects timing. Post-silicon IR drop monitoring circuits are lacking, forcing designers to reserve substantial static guard bands for worst-case scenarios, compromising energy efficiency. Inspired by biomimetics, this paper proposes an Actiniaria-shaped structure for distributed monitoring of multiple IR drop hotspots without invading critical paths (CPs). It comprises “lightweight tentacles + central monitor”. The tentacles use voltage-sensitive cells to sense the impact of IR drop on timing, while also supporting adaptive tuning for global variations. By using on-chip PVT sensors, Actiniaria tracks the timing characteristics of the longest CP in real time under PVT variations. A two-stage pre-warning timing monitor captures timing margins and employs adaptive voltage/frequency scaling (AVFS) strategies to compress redundant guard bands. When applied to an open-source multi-core RISC-V processor fabricated in 22-nm CMOS, Actiniaria achieves a 40.6% reduction in power consumption through non-invasive dynamic IR drop monitoring while having only 0.08% area overhead.
- 18 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
- Vol. 33 (10) , 2772-2782
A Compound Timing Detection of Both Data Transition and Path Activation for Reliable In Situ Error Detection and Correction
Timing error detection and correction (EDAC) in resilient circuits helps eliminate excess timing margins. However, it faces misdetection risks when critical paths (CPs) remain inactive. We propose a 15-transistor timing error and path-activation detector (TEPD) capable of detecting both timing violations in late-arriving signals and CP activation, with robust operation down to 0.33 V. Regarding circuit-level error correction, we introduce an error-correcting flip-flop (ECFF), leveraging time-borrowing for zero-cycle response latency without requiring pipeline refresh. The custom-optimized ECFF adds only six transistors, increasing delay, dynamic power, and static power by 15%, 11%, and 14%, respectively, compared with a standard flip-flop, ensuring efficient error correction with minimal cost. A system-level voltage tuning strategy is further developed to handle continuous timing errors, ensuring robust adaptive voltage scaling (AVS) operation. Implemented on a neural network (NN) accelerator in the 28-nm CMOS, the system operates across a wide voltage range from 0.54 to 0.9 V. It achieves up to 52% power gain or 123% frequency gain at the near-threshold region, with negligible area and power overhead compared with the margined baseline.
- 18 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
- Vol. 33 (10) , 2772-2782
A Compound Timing Detection of Both Data Transition and Path Activation for Reliable In Situ Error Detection and Correction
Timing error detection and correction (EDAC) in resilient circuits helps eliminate excess timing margins. However, it faces misdetection risks when critical paths (CPs) remain inactive. We propose a 15-transistor timing error and path-activation detector (TEPD) capable of detecting both timing violations in late-arriving signals and CP activation, with robust operation down to 0.33 V. Regarding circuit-level error correction, we introduce an error-correcting flip-flop (ECFF), leveraging time-borrowing for zero-cycle response latency without requiring pipeline refresh. The custom-optimized ECFF adds only six transistors, increasing delay, dynamic power, and static power by 15%, 11%, and 14%, respectively, compared with a standard flip-flop, ensuring efficient error correction with minimal cost. A system-level voltage tuning strategy is further developed to handle continuous timing errors, ensuring robust adaptive voltage scaling (AVS) operation. Implemented on a neural network (NN) accelerator in the 28-nm CMOS, the system operates across a wide voltage range from 0.54 to 0.9 V. It achieves up to 52% power gain or 123% frequency gain at the near-threshold region, with negligible area and power overhead compared with the margined baseline.