15 November 2022 in IEEE Internet of Things Journal
Structural health monitoring (SHM) is essential for modern large buildings and infrastructure. Radio frequency identification (RFID) widen a new paradigm for SHM in the Internet of Things (IOT) era. This paper introduces a low-cost intelligent RFID monitoring system for future IoT applications. Through adding a multi-parameter sensing strategy to the passive RFID tags that can be deployed on a large scale to form a sensor network, which expands the coverage of IOT in key positions health monitoring of buildings. In this work, a novel chipless RFID strain sensing tag is designed to characterize the magnitude and direction of metal surface strain, and a low-cost detection method suitable for large mechanical structures is proposed. The traditional strain antennas focus on identifying the strain characteristics without the function of encoding, and there are limitations such as the rigid measurement methods and the expensive measurement instruments. The tag designed in this paper integrates both strain sensing and encoding functions, and has the advantages of small size, high data capacity, and information reading is not easily affected by environmental noise. This paper proposes an economical and flexible tag spectrum extraction method, which realizes the intelligent collection and transmission of tag data by connecting lightweight vector network analyzer (VNA) with microcontroller. Combined with the IOT, this method can be well applied to the security assessment and damage detection of large infrastructure structures, which provides a new approach to modern building health monitoring applications from integrated tag design to intelligent detection and risk assessment schemes.
5 December 2022 in IEEE Access
As people’s daily behavioral activities become more data-based, how to protect personal information security is a key consideration for the whole society. Finger vein recognition is becoming an important means of identification because of its uniqueness, live detection, security and many other advantages. Although deep learning can make finger vein recognition have good effect, but the number of samples needed to build deep network model is too large, and the current authoritative finger vein database cannot reach the minimum number of samples required. The emergence of Muti-Grained Cascade Forest provides a solution to the problem of insufficient sample data and long training time, which can give a new research avenue in feature extraction. In order to obtain higher accuracy, the deep forest algorithm is introduced in this paper to process the finger vein images. Firstly, the image data in the finger vein image database is pre-processed to prepare for the subsequent feature extraction and matching. Then, the deep forest algorithm is used to find the feature points and ORB algorithm is used to match the features to obtain the angular information of each matched pair, and the final identity is determined according to the sparse distribution of angles. The accuracy of finger vein recognition based on deep forest algorithm is 98.40%. By comparing with other machine learning methods for finger vein recognition, the proposed method has a higher accuracy rate.
3 March 2025 in IEEE Transactions on Instrumentation and Measurement
This article proposes a multi-parameter radio frequency identification (RFID) sensor with two controllable coding structures. The sensor integrates the ability of strain and humidity detection and is suitable for the field of structural health monitoring (SHM). The sensor is equipped with three S-shaped structures and one rectangular structure, two of S-shaped structures are used as encoding units, one S-shaped structure is used as a humidity detection unit, and the rectangular structure is used as a strain unit. These structures are made on a RogersRT3003 dielectric substrate. Controllable coding is achieved by changing the capacitance value of the varactor diode. A single coding unit can achieve a coding value of more than 2 bits. Compared with the traditional coding structure, the coding information capacity is improved, which is more suitable for distributed SHM scenarios that require multiple sensors. The digital control circuit is designed on the flexible printed circuit (FPC), which has the characteristics of bendability. It is mainly composed of microcontroller, Bluetooth, varactor diodes and other components, and can realize the wireless control function of the coding unit. A broadband antenna is further designed to match the working frequency band of the sensor. Experimental verification shows that the micro-strain of the strain unit is linearly related to its resonant frequency, and the humidity sensitivity is 193.92 kHz/%RH at a relative humidity (RH) of 33.7%–81.2%. The two encoding units can be controlled wirelessly to achieve 25-bit coding, while having the advantages of small size and low cost. In the future, this method can be combined with self-powered devices and internet of thing (IoT) gateways to achieve long-term state control of multiple sensors, providing a new method for SHM applications.
24 June 2022 in IEEE Transactions on Instrumentation and Measurement
Aiming at the problem of whether the color difference exceeds the threshold when the RGB-LED is lit and how to calibrate the deviation, this article proposes an RGB LED chromaticity parameter prediction and calibration algorithm. This article is divided into two parts for the intelligent prediction and calibration method of RGB LED chromaticity parameters. The first is the color difference prediction part. First, combining the ideas of random forest and XGBoost, this article designs a classification algorithm and named it RF-XGBoost. Use RF-XGBoost to predict whether the color difference exceeds 0.0055 when the RGB lamp lights up 30 colors. Compared with decision tree, random forest, and XGBoost, RF-XGBoost performs better in recall rate through comparative experiment analysis. The second is the color correction part. Aiming at the problem of excessive color difference when RGB-LED is lit, this article proposes the cor-calibration color calibration algorithm. The function relation between chromaticity coordinate and error model is established. The current error compensation model, automatic selection of calibration step, and closed-loop calibration are used to calibrate the colors beyond the error. Compared with the existing calibration algorithm, after using the cor-calibration algorithm, the color pass rate in the color gamut increased from 91.25% to 99.79%. Experiments show that the RF-XGBoost algorithm and the cor-calibration algorithm can achieve precise screening and high-precision calibration of RGB-LEDs.