Wearable health monitoring systems have gained considerable interest in recent years owing to their tremendous promise for personal portable health watching and remote medical practices. The sensors with excellent flexibility and stretchability are crucial components that can provide health monitoring systems with the capability of continuously tracking physiological signals of human body without conspicuous uncomfortableness and invasiveness. The signals acquired by these sensors, such as body motion, heart rate, breath, skin temperature and metabolism parameter, are closely associated with personal health conditions. This review attempts to summarize the recent progress in flexible and stretchable sensors, concerning the detected health indicators, sensing mechanisms, functional materials, fabrication strategies, basic and desired features. The potential challenges and future perspectives of wearable health monitoring system are also briefly discussed.

Field programmable gate array (FPGA) is widely considered as a promising platform for convolutional neural network (CNN) acceleration. However, the large numbers of parameters of CNNs cause heavy computing and memory burdens for FPGA-based CNN implementation. To solve this problem, this paper proposes an optimized compression strategy, and realizes an accelerator based on FPGA for CNNs. Firstly, a reversed-pruning strategy is proposed which reduces the number of parameters of AlexNet by a factor of 13× without accuracy loss on the ImageNet dataset. Peak-pruning is further introduced to achieve better compressibility. Moreover, quantization gives another 4× with negligible loss of accuracy. Secondly, an efficient storage technique, which aims for the reduction of the whole overhead cache of the convolutional layer and the fully connected layer, is presented respectively. Finally, the effectiveness of the proposed strategy is verified by an accelerator implemented on a Xilinx ZCU104 evaluation board. By improving existing pruning techniques and the storage format of sparse data, we significantly reduce the size of AlexNet by 28×, from 243 MB to 8.7 MB. In addition, the overall performance of our accelerator achieves 9.73 fps for the compressed AlexNet. Compared with the central processing unit (CPU) and graphics processing unit (GPU) platforms, our implementation achieves 182.3× and 1.1× improvements in latency and throughput, respectively, on the convolutional (CONV) layers of AlexNet, with an 822.0× and 15.8× improvement for energy efficiency, separately. This novel compression strategy provides a reference for other neural network applications, including CNNs, long short-term memory (LSTM), and recurrent neural networks (RNNs).

The field of piezoresistive sensors has been undergoing a significant revolution in terms of design methodology, material technology and micromachining process. However, the temperature dependence of sensor characteristics remains a hurdle to cross. This review focuses on the issues in thermal-performance instability of piezoresistive sensors. Based on the operation fundamental, inducements to the instability are investigated in detail and correspondingly available ameliorative methods are presented. Pros and cons of each improvement approach are also summarized. Though several schemes have been proposed and put into reality with favorable achievements, the schemes featuring simple implementation and excellent compatibility with existing techniques are still emergently demanded to construct a piezoresistive sensor with excellent comprehensive performance.

The longstanding demands for micropressure detection in commercial and industrial applications have led to the rapid development of relevant sensors. As a type of long-term favored device based on microelectromechanical system technology, the piezoresistive micropressure sensor has become a powerful measuring platform owing to its simple operational principle, favorable sensitivity and accuracy, mature fabrication, and low cost. Structural engineering in the sensing diaphragm and piezoresistor serves as a core issue in the construction of the micropressure sensor and undertakes the task of promoting the overall performance for the device. This paper focuses on the representative structural engineering in the development of the piezoresistive micropressure sensor, largely concerning the trade-off between measurement sensitivity and nonlinearity. Functional elements on the top and bottom layers of the diaphragm are summarized, and the influences of the shapes and arrangements of the piezoresistors are also discussed. The addition of new materials endows the research with possible solutions for applications in harsh environments. A prediction for future tends is presented, including emerging advances in materials science and micromachining techniques that will help the sensor become a stronger participant for the upcoming sensor epoch.

Neural network shows great potential in modulation classification because of its excellent accuracy and achievability but overfitting and memorizing data noise often happen in previous researches on automatic digital modulation classifier. To solve this problem, we utilize two neural networks, namely MentorNet and StudentNet, to construct an automatic modulation classifier, which possesses great performance on the test set with −18–20 dB signal-to-noise ratio (SNR). The MentorNet supervises the training of StudentNet according to curriculum learning, and deals with the overfitting problem in StudentNet. The proposed classifier is verified in several test sets containing additive white Gaussian noise (AWGN), Rayleigh fading, carrier frequency offset and phase offset. Experimental results reveal that the overall accuracy of this classifier for common eleven modulation types was up to 99.3% while the inter-class accuracy could be up to 100%, which was much higher than many other classifiers. Besides, in the presence of interferences, the overall accuracy of this novel classifier still could reach 90% at 10 dB SNR indicting its excellent robustness, which makes it suitable for applications like military electronic warfare.

This paper proposes a multidirectional piezoelectric vibration energy harvester based on an improved U-shaped structure with cross-connected beams. Benefitting from the bi-directional capacity of U-shaped beam and additional bending mode induced by cross-connected configuration, the proposed structure can well capture the vibrations in 3D space at the frequencies lower than 15 Hz. These features are further validated by finite element analyses and theorical formulas. The prototype is fabricated and the experiments under different conditions are carried out. The results show that the proposed harvester can generate favorable voltage and power under multidirectional vibrations at a low operating frequency. Practical applications of charging capacitors and powering a wireless sensor node demonstrate the feasibility of this energy harvester in supplying power for engineering devices.