Er-Ping Li

Dielectric elastomers (DEs) are promising artificial muscles for soft robotics and wearable devices due to their large deformation, fast response, high energy density, and light weight. However, conventional DE films exhibit limited deformation modes and lack multifunctionalities. Passive layers, rigid constraints, or multilayered electrode patterning have been used to extend their deformation but often reduce actuation performance and increase structural complexity. Here, we develop patterned electrophoresis of magnetic additives to program deformations in monolithic DE films, achieving bending, twisting, and complex three-dimensional morphing while maintaining high actuation performance. The additives also endow the DE films with multifunctionalities, including magnetic response and electromagnetic wave shielding. By coupling programmed deformations with these functionalities, we demonstrate intelligent soft devices with new functions and controllability such as electric-magnetic co-driven actuation arrays and deformable electromagnetic metasurfaces. This strategy achieves programmable deformations and multifunctionalities in single-layer DE films, opening possibilities for advanced DE applications.

The continuous scaling of complementary metal oxide semiconductor process technology has necessitated a reduction in operating voltages, thereby diminishing the noise margins of chips. Consequently, the ground bounce induced by electrostatic discharge events is more likely to cause critical power supply fluctuations, leading to soft failures in the chip. To rapidly identify design vulnerabilities, this article developed a power supply noise simulation model and integrated it into the full microcontroller unit (MCU) performance simulation flow to predict noise levels and guide design modifications. A power supply noise measurement method that requires no additional circuitry is further proposed to monitor key signals associated with soft failures on the MCU. Through improvements in down-bond packaging and a redesign of the low-dropout regulator and phase-locked loop circuits, the chip’s soft failure level is enhanced by a factor of more than 10 without increasing die area or degrading chip performance.

Achieving flexible manipulation of specific electromagnetic wave morphology has become an indispensable technique for numerous promising applications, such as wireless communication, energy harvesting, and parameter estimation. However, how to realize efficient and noninvasive electromagnetic wave manipulation for moving objects within complex media remains a significant challenge due to the complicated disordered scattering characteristics. To address this problem, this article proposes a manipulation methodology based on a novel reinforcement-conditioned diffusion model (RC-DM). Specifically, the generalized Wigner–Smith (GWS) operator is leveraged to achieve remarkable target-centric field enhancement or spatial suppression for the desired object inside complex scattering media. Subsequently, to enable noninvasive and accurate wave morphology manipulation for moving objects, the RC-DM is proposed to form the desired electromagnetic field solely through the utilization of the system scattering matrix. Furthermore, an improved sampling process of RC-DM is proposed, offering enhanced efficiency and accuracy to realize a near-real-time response of wave manipulation for moving objects. Finally, the corresponding versatility and performance are comprehensively demonstrated through the numerical validation and experiment under different moving trajectories. The results show that our proposed methodology has excellent effectiveness and robustness to address the challenge of electromagnetic wave morphology manipulation for moving objects inside complex scattering media.

Spike neural networks (SNN) and neuromorphic chips and chiplets have recently drawn increasing attention. However, these systems utilize spike signals and address-event representation protocols, which differ fundamentally from traditional digital circuits. As a result, they may face different signal integrity issues and require special equalization methods. This paper presents a Long Short-Term Memory (LSTM) based signal equalization approach that effectively addresses spike signal equalization of different data rates. This LSTM equalization method exhibits outstanding scalability in restoring spike signals at different data rates using one trained model. By implementing the LSTM equalization approach, high-precision signal equalization for spike signals is achieved, which demonstrates a remarkable advantage over conventional feed-forward equalization (FFE). These results verify the effectiveness of the proposed method for high-precision spike signal equalization and highlight its potential for neuromorphic and spike-based interconnect systems.

The common-mode (CM) noise generated by serializer/deserializer systems causes electromagnetic (EM) interference by radiating EM energy into the surrounding environment. This article focuses on the CM noise challenge in current-mode logic (CML) driver fabricated in TSMC N65 process. The generation of CM noise in the CML driver is analyzed by applying an equivalent circuit model with parasitic capacitors at the 20-Gb/s data rate. Based on the equivalent circuit analysis, an accurate CM noise analysis method is proposed. A low-CM-noise driver design based on cascaded CML drivers is proposed to mitigate the output CM noise. The proposed driver reduces CM noise by 20.5 dB for non-return-to-zero (NRZ) signaling compared with the traditional CML driver.

This article presents a physics-aware convolutional long short-term memory (PC-LSTM) network for efficient and accurate modeling of mutual impedance matrices in dipole antenna arrays. By reinterpreting the Green’s function through a physics-aware neural network and embedding it into an adaptive loss function, the proposed machine learning-based approach achieves enhanced physical interpretability in mutual coupling modeling. Also, an attention mechanism is carefully designed to calibrate complex-valued features by fusing the real and imaginary parts of the Green’s function matrix. These fused representations are subsequently fed into a convolutional long short-term memory network incorporating a physics-aware convolution kernel, and the impedance matrix of the linear antenna array can be finally derived. Validation against five benchmarks underscores the efficacy of the proposed approach, demonstrating accurate impedance extraction with up to a 7× speedup compared to CST Microwave Studio, making it a fast alternative to full-wave simulations for mutual coupling characterization.