11 December 2024 in IEEE Transactions on Aerospace and Electronic Systems
Designing a radiation-tolerant SRAM-based system poses a significant challenge for space facilities operating in radiation environments. Accurate evaluation of SEE sensitivity, along with the design of radiation-hardened memory cells and power management circuits is crucial, particularly for advanced nanotechnologies. In this paper, a radiation-tolerant FDSOI SRAM with two switchable bandgap reference circuits was designed. An additional SEE-hardened selection circuit was used to enable the targeted bandgap reference circuits. The ∼3×10 -14 cm 2 /bit saturated SEU cross section for the hardened SRAM device was measured by our heavy ion irradiation test. The bipolar bandgap reference circuit induced SET cross sections were confirmed by comparing different test modes. Besides, the features of 181 Ta-induced transients in the bipolar bandgap reference circuit were captured during our experiments. The designed two reference circuits exhibit SEE tolerance, although a few transients can be observed in the bipolar circuit, leading to a ∼30% increase in output voltage. The MOS bandgap reference circuit exhibits strong radiation resistance but is slightly inferior at controlling temperature drift and maintaining the stability of the reference voltage. The SET measurement results for the bipolar bandgap reference circuit reveal specific radiation sensitivity and potential risks for high-reliable SRAM systems, whereas the full MOS structure can further enhance the radiation tolerance of the entire system.
14 November 2024 in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
Fully homomorphic encryption (FHE) enables calculations on encrypted data and is a crucial foundation for achieving privacy computing. However, the high computation overhead restricts its widespread application. Even after algorithm and software optimization, its processing speed remains low. This article proposes the first practical system-level multicore Brakerski-Gentry-Vaikuntanathan (BGV) hardware acceleration scheme based on field-programmable gate array (FPGA). By analyzing the bottleneck of system acceleration, a hierarchical storage structure is introduced to reduce data movement. A novel 4-2 mixed-radix number theoretic transform (NTT) algorithm is proposed, allowing flexible switching between radix-4 and radix-2, with the ability to reuse twiddle factors. In addition, a reconfigurable processing element (PE) is proposed that supports all homomorphic operations of BGV. The design of this article is evaluated on Xilinx Virtex7 series FPGA, achieving a throughput of NTT/inverse NTT (INTT) up to $14\times $ higher than previous designs. Compared with simple encrypted arithmetic library (SEAL), the full system performances of homomorphic encryption (ENC), decryption (DEC), and homomorphic multiplication achieve improvements of $13.9\times $ , $7.07\times $ , and $16.6\times $ , respectively.
4 August 2025 in IEEE Transactions on Nuclear Science
With the growing reliance on graph data processing in safety-critical applications, ensuring the reliability of graph convolutional network (GCN) hardware systems has become paramount, especially in radiation-prone environments where single-event upsets (SEUs) pose significant risks. This article presents a comprehensive design framework for a fault-tolerant GCN system, addressing the unique challenges of SEU susceptibility through a novel fault-aware centrality measure and a dynamic hardware defense (DHD) strategy. Our approach begins with the development of a fault-aware centrality measure to precisely model the distribution of critical nodes within graph data. Leveraging this measure, we design a DHD strategy that integrates partial circuit reinforcement and high-centrality node marking, which dynamically route the dataflow of the most influential nodes in the graph to reinforced circuit units. The proposed system architecture is rigorously validated through extensive experiments, demonstrating significant improvements in fault tolerance. Specifically, the DHD strategy achieves improvements in hardening efficiency of $2.53\times $ , $2.70\times $ , and $2.85\times $ , respectively, compared with the traditional full triple modular redundancy (FTMR) strategy. Neutron radiation testing further validates the robustness of the system, showing effective mitigation of fault propagation under extreme conditions. By maintaining minimal hardware overhead and offering dynamic, cost-effective protection, this design framework provides a reliable solution for deploying GCNs in safety-critical applications.
1 September 2026 in IEEE Transactions on Device and Materials Reliability
In recent years, Convolutional Neural Networks (CNNs) have emerged as a cornerstone technology for image processing tasks in safety-critical domains. Although advanced Artificial Intelligence (AI) acceleration devices are ideal for space applications due to their high computational efficiency and flexibility, they remain highly susceptible to Single Event Upsets (SEUs) and are still constrained by limited resources. Thus, it becomes imperative to design CNN implementations for Field- Programmable Gate Array (FPGA) devices that achieve a delicate balance between low cost and high reliability. However, traditional hardening strategies often neglect the sensitivity of internal nodes, relying instead on global or coarse-grained module-level hardening strategies that significantly increase resource overhead and degrade timing performance. To address this issue, we propose a basic unit partitioning strategy that divides CNN accelerators into functionally independent circuit nodes. Within our fault analysis framework, we map fault locations to specific nodes using address mapping, enabling sensitivity analysis. Additionally, we implement fine-grained hardening on critical nodes identified through sensitivity analysis and compare this strategy against module-level coarse-grained hardening and full triple modular redundancy (TMR) strategies. Fault analysis results reveal that our strategy delivers excellent hardening effectiveness at a reduced cost. Our strategy achieves a 31.9% reduction in resource overhead and a 29.9% improvement in timing performance compared to Full_TMR, with comparable fault tolerance. Compared to module-level hardening strategies, our approach achieves higher hardening efficiency. The irradiation experiment results further validate the effectiveness of our strategy, achieving a 10.6× improvement in Mean Time To Failure (MTTF) compared to the unhardened baseline, which is comparable to that of full TMR. The proposed reliability evaluation method achieves high accuracy, with errors within ±4.3%. Overall, the proposed hardening strategy and reliability evaluation method present a significant advancement in designing robust and cost-effective systems, holding vast potential for adoption in safety-critical fields such as aerospace and autonomous driving.
8 December 2025 in IEEE Transactions on Device and Materials Reliability
A novel radiation-hardened 14T (NRH-14T) SRAM cell based on 16nm FinFET process is designed in this paper. The NRH-14T cell can recover from errors induced by single-event upsets (SEUs) and is capable of resisting multiple-bit upsets caused by charge-sharing effects. The NRH-14T cell employs an interlocking structure to protect against SEUs. To address the write-failure issue caused by the interlocking structure, four access transistors are utilized in NRH-14T cell. Due to the small size of FinFET devices, the charge-sharing effect is more likely to induce multiple-bit upsets in the circuit, thereby leading to hardening failure. To reduce the impact of charge sharing and recover from multiple-bit upsets with minimal area cost, an optimized NRH-14T cell layout is designed and verified. The single-event transient (SET) pulse waveform generated by particle strikes on FinFET devices, distinct from that of planar devices, is extracted and employed to verify the error-tolerance performance of the NRH-14T cell. Verification results indicate that the NRH-14T cell, with an acceptable increase in area cost, achieves 30.89ns read speed and 49.32ns write speed, can resist SEUs under all conditions with a critical charge of 13fC, and can recover from multiple-bit upsets.
18 May 2026 in IEEE Journal of the Electron Devices Society
To address the unclear voltage-dependent electrical characteristics and long-range array disturb mechanism in high-reliability symmetric split-gate memory, a systematic evaluation methodology is proposed.This approach integrates a comprehensive test metric covering a wide temperature range (-40 to 125 °C) ,along with a test sequence designed to precisely isolate potential disturb sources,for reliability characterization.Through experimental testing,the temperature dependence of read current under various conditions is further revealed.TCAD simulations and key parameter extraction are employed to investigate the underlying physical mechanisms behind these phenomena.Furthermore,by analyzing the current distribution across the array,estimating the threshold voltage range based on terminal voltage configurations of the memory cell,and considering the overall array architecture,the dominant mechanisms for both erase disturb and program disturb are elucidated.This work provides critical experimental foundations for the design and modeling of high-reliability NOR Flash memory.