科研成果 by Type: Conference Proceedings

This work presents a detailed study on the high-frequency performance of 22FDX ® FDSOI for 5G front-end power amplifiers. The following report focuses on the S-parameters and large-signal figure-of-merits such as output power, gain and power-added efficiency for an insightful and correct assessment on the device capability. DC characteristics of the test transistors are firstly investigated to determine the optimum operating point. Small-signal characterization is performed up to 110 GHz using a state-of-the-art mm-Wave measurement setup. An overall MSG/MAG of 16 ± 4 dB is recorded in the frequency range 10 - 80 GHz. On the other hand, large-signal performance on non-50 Ohm impedance environment is evaluated thoroughly through vector-receiver load-pull measurement up to 24 GHz. The measured output power and efficiency indicate that the DUTs perform well in the sub-6 GHz band and even in K-band. The outstanding experimental results emphasize the applicability and suitability of the 22FDX ® FDSOI technology platform for 5G low-power transmitters.

Personalized recommendation has important implications in raising online shopping efficiency and increasing product sales. There has been wide interest in finding ways to provide more efficient personalized recommendations. Most existing studies focus on how to improve the accuracy of the recommendation algorithms, or are more concerned on ways to increase consumer satisfaction. Unlike these studies, our study focuses on the process of decision-making, using long tail theory as a basis, to reveal the mechanisms involved in consumers' adoption of recommendations. This paper analyzes the effect of personalized recommendations from two angles: product sales and ratings, and tries to point out differences in consumer preferences between mainstream products and niche products, high rating products and low rating products, search products and experience products. The study verifies that consumers demand diversity in the recommended content, and also provides suggestions on how to better plan and operate a personalized recommendation system. © 2019 IEEE Computer Society. All rights reserved.

Stretchable and movable 3D structure is a great choice for sensing stretching and bending. This paper reports a novel cross-shaped 3D buckling strain sensor based on polydimethylsiloxane (PDMS) substrate for detecting stretching and bending. Using pre-stretched PDMS, cross-shaped Polyimide (PI) film with conductive silver paint on its top surface as a 2D precursor can pop up as a dynamic 3D structure and possesses capacitive effect and triboelectric effect under different stretching and bending, which can detect stretching directions, strain value, bending axis direction and radius of curvature simultaneously, showing great potential in human and robot applications.

This paper presents a 60 GHz phase shifter, based on a coplanar waveguide (CPW) transmission line, loaded with ferroelectric hafnium zirconium oxide (HZO) variable metal-insulator-metal (MIM) varactors, developed for the back-end-of-line (BEoL) on-chip integration. Using the measured data of capacitance-voltage (C-V) characteristics of HZO and implementing the method-of-moments simulation, it was shown, that by changing the bias voltage between 0.95 and -3 V, the device shows a phase shift of 111° and a minimum insertion loss of -5.84 dB at 60 GHz. The chip area of the device is 0.206 mm 2 , making it the smallest among non-CMOS phase shifters.

Task type may have influence on driving behavior in disasters, while personality trait may have moderating effects on driving behavior in different task types, there are few studies focus on it. In this paper, the research method of laboratory experiment and questionnaires is used to investigate the impact of altruistic task and self-interest task on driving behavior. And the personality trait is taken as the moderating variable to research the moderating effect of personality trait on driving behavior in different task types. The results show that, (1) Compared with self-interest task, altruistic task has critically significant influence on average acceleration, which indicates driver will improve the driving performance in altruistic task. (2) Personality trait of emotional stability has a moderating effect in the influence mechanism of task type on driving behavior. The results of this paper may provide scientific support to traffic evacuation and emergency management in disasters. © 2019 ACM.

This paper reports a novel self-powered three-dimension motion sensor capable of independently detecting contact trajectory, pressure and velocity based on triboelectrification and electrostatic induction synchronously. Motion trajectories in the full plane can be identified by using a unique net-cross electrodes configuration design. In addition, the patterned silver nanowires (AgNWs) electrodes are sprayed onto the polydimethylsilane (PDMS) substrate to achieve good transparency and stretchability. By attaching the 3D motion sensor on human skin or robot surface directly, the 3D motion information of the object could be acquired including pressure, velocity and trajectory. The self-powered 3D motion sensor is a promising candidate in terms of human-computer interaction, anti-counterfeiting signatures, etc.

In this paper, a comprehensive analysis on small-signal modeling of mm-wave transistor in 22nm FDSOI technology is presented. The model is constructed based on experimental S-parameters up to 110 GHz of a 22FDX® thick-oxide n-MOSFET and analytical parameter extraction approach. The non-quasi static effect is addressed thoroughly in the equivalent circuit model for high frequency validity. The bias-dependent series source and drain resistances are considered to account for the overlap regions between the gate and the highly doped source/drain regions. In addition, a simple RC network is included at the output to model the innegligible substrate coupling at mm-wave frequencies. Excellent agreements between model prediction and measurement are observed in the interested bandwidth for various bias conditions.

In this paper, we present a scalable and general fabrication for micro-supercapacitors (MSCs) among various flexible substrates assisted by the stamp, which combines the conductive polymer composites with gravure printing process. Compared with the traditional transferring techniques, this method greatly simplifies the process and mitigates the mechanical damage during the preparation. Profiting from the composites of carbon nanotubes (CNTs) and polydimethylsiloxane (PDMS) as the printing inks, the MSCs exhibit elegant areal capacitance (10.491 μF/cm2) on the paper substrate. Meanwhile, optimizing the ratio of matrix and curing agent of PDMS, the interaction between ink and substrate is effectively enhanced. Therefore, such novel fabrication technology significantly improves the production efficiency as well as broadens the applications.

This paper presents a tunable 60 GHz band-pass filter, based on a coplanar waveguide (CPW) transmission line, periodically loaded with ferroelectric Hafnium Zirconium Oxide (HZO) variable metal-ferroelectric-metal (MIM) capacitors (varactors), developed for back-end-of-line (BEoL) integration. Derived from the nonlinear capacitance of hafnium zirconium oxide and implementing the method-of-moments simulation, it was shown, that with changing the bias voltage between 0.95 and -3 V, the filter’s center frequency can be tuned between 60.5 and 69,7 GHz, respectively. Hereby, a minimum insertion loss of -3.3 dB is realized. The chip area of the filter is only 0.062 mm 2 , making it the smallest among tunable V-band filters.

Emojis have become more and more popular in text-based online communication to express emotions. This indicates a potential to utilize emojis in sentiment analysis and emotion measurements. However, many factors could affect people’s emoji usage and need to be examined. Among them, age, gender, and relationship types may result in different interpretations of the same emoji due to the ambiguity of the iconic expression. In this paper, we aim to explore how these factors may affect the frequency, type, and sentiment of people’s emoji usage in communications. After analyzing 6,821 Wechat chatting messages from 158 participants, we found people between 26–35 had lowest frequency of emoji usage; younger and elder groups showed different sentiment levels for the same emojis; people chose emoji types based on relationships. These findings shed light on how people use emojis as a communication tool.

This paper for the first time prototypes and compares the homodyne and heterodyne terahertz dielectric sensors for lab-on-chip applications. The homodyne sensor consists of a multiplier chain, a balun-based power divider, an on-chip transducer, and IQ mixers. Differently, the heterodyne sensor requires an additional multiplier chain; however, it waives one mixer and a power divider, leading to reduced losses and alleviated power consumption. Fabricated using 0.13 µm SiGe BiCMOS technology, the homodyne and heterodyne sensors take 4 mm 2 and 5.2 mm 2 , and consume 400 mW and 499 mW, respectively. By experiments, both designed homodyne and heterodyne sensors can effectively sense the dielectric parameters of the samples. Moreover, the heterodyne sensor can address the DC offset issues with merely 99 mW additional power.

This paper presents two scalable resonator-based transducers (RBTs) at terahertz (THz) frequency range to realise THz spectroscopy for dielectric sensing. First, the design of 0.24 THz RBT is described by scaled a 0.12 THz sensing structures which utilises a wavelength-long closed-ring resonator to place inside of the Coplanar stripline (CPS) to make a high-selective bandpass response and combines with short-ended strips to create the bandstop behavior. Its scattering parameter can have a very large magnitude change and resonance frequency shift for the loaded samples. Next, a ring structure is also presented to implement 0.48 THz sensing by scaled a low frequency RBT, which employs ring resonator with an asymmetrically loaded stubs to perform a high analytic sensitivity and selectivity for loaded samples. Both presented scalable transducers, possessing the high integration capability of silicon circuits, are proved to be the promising employments in THz spectroscopy.