With the rapid development of the computer and sensor field, inertial sensor data have been widely used in human activity recognition. At present, most relevant studies divide human activities into basic actions and transitional actions, in which basic actions are classified by unified features, while transitional actions usually use context information to determine the category. For the existing single method that cannot well realize human activity recognition, this paper proposes a human activity classification and recognition model based on smartphone inertial sensor data. The model fully considers the feature differences of different properties of actions, uses a fixed sliding window to segment the human activity data of inertial sensors with different attributes and, finally, extracts the features and recognizes them on different classifiers. The experimental results show that dynamic and transitional actions could obtain the best recognition performance on support vector machines, while static actions could obtain better classification effects on ensemble classifiers; as for feature selection, the frequency-domain feature used in dynamic action had a high recognition rate, up to 99.35%. When time-domain features were used for static and transitional actions, higher recognition rates were obtained, 98.40% and 91.98%, respectively.

Imaging past the diffraction limit is of significance to an optical system. Fourier ptychography (FP) is a novel coherent imaging technique that can achieve this goal and it is widely used in microscopic imaging. Most phase retrieval algorithms for FP reconstruction are based on Gaussian measurements which cannot extend straightforwardly to long range, sub-diffraction imaging setup because of laser speckle noise corruption. In this work, a new FP reconstruction framework is proposed for macroscopic visible imaging. When compared with existing research, the reweighted amplitude flow algorithm is adopted for better signal modeling, and the Regularization by Denoising (RED) scheme is introduced to reduce the effects of speckle. Experiments demonstrate that the proposed method can obtain state-of-the-art recovered results on both visual and quantitative metrics without increasing computation cost, and it is flexible for real imaging applications.

Data-driven fault diagnosis methods have received extensive attention, and the existing diagnostic methods usually require sufficient supervised data. However, developing effective diagnostic methods with insufficient training data remains challenging, which is highly demanding in real industrial scenarios since collecting high-quality fault data is often difficult and expensive. Considering the potential similarity of rotating machinery, collecting bearing fault data from different but related equipment may benefit the diagnostic performance of the target machinery. This paper proposes a novel multi-source domain adaptation framework for bearing fault diagnosis method based on adversarial network and feature enhancement. This method reduces the feature distribution discrepancy between the target domain and each source domain through domain adversarial training, and transfers the fault diagnosis knowledge learned from multiple labeled source domains to a single unlabeled target domain. Considering the problem of scarce fault data, this method uses multi-linear mapping to fuse the class information with high-level features, and introduces AlignMixup to enhance the real fault signal features. To evaluate the model, experimental validation was performed on the CWRU and KAT datasets. The results show that the proposed method is promising to address the bearing fault diagnosis tasks from different places of machines, further improving the applicability of data-driven methods in real industries.

As an important error in star centroid location estimation, the systematic error greatly restricts the accuracy of the three-axis attitude supplied by a star sensor. In this paper, an analytical study about the behavior of the systematic error in the center of mass (CoM) centroid estimation method under different Gaussian widths of starlight energy distribution is presented by means of frequency field analysis and numerical simulations. Subsequently, an optimized extreme learning machine (ELM) based on the bat algorithm (BA) is adopted to predict the systematic error of the actual star centroid position and then compensate the systematic error from the CoM method. In the BA-ELM model, the input weights matrix and hidden layer biases parameters are encoded as microbat’s locations and optimized by utilizing the strong global search capacity of BA, which significantly improves the performance of ELM in terms of prediction accuracy. The simulation result indicates that our method can reduce the systematic error to less than 3.0 × 10⁻⁷ pixels, and its compensation accuracy is two or three orders of magnitude higher than that of other methods for estimating a star centroid location under a 3 × 3 pixel sampling window.

In recent years, much research has been conducted on time series based human activity recognition (HAR) using wearable sensors. Most existing work for HAR is based on the manual labeling. However, the complete time serial signals not only contain different types of activities, but also include many transition and atypical ones. Thus, effectively filtering out these activities has become a significant problem. In this paper, a novel machine learning based segmentation scheme with a multi-probability threshold is proposed for HAR. Threshold segmentation (TS) and slope-area (SA) approaches are employed according to the characteristics of small fluctuation of static activity signals and typical peaks and troughs of periodic-like ones. In addition, a multi-label weighted probability (MLWP) model is proposed to estimate the probability of each activity. The HAR error can be significantly decreased, as the proposed model can solve the problem that the fixed window usually contains multiple kinds of activities, while the unknown activities can be accurately rejected to reduce their impacts. Compared with other existing schemes, computer simulation reveals that the proposed model maintains high performance using the UCI and PAMAP2 datasets. The average HAR accuracies are able to reach 97.71% and 95.93%, respectively.

Imaging correlography, an effective method for long-distance imaging, recovers an object using only the knowledge of the Fourier modulus, without needing phase information. It is not sensitive to atmospheric turbulence or optical imperfections. However, the unreliability of traditional phase retrieval algorithms in imaging correlography has hindered their development. In this work, we join imaging correlography and ptychography together to overcome such obstacles. Instead of detecting the whole object, the object is measured part-by-part with a probe moving in a ptychographic way. A flexible optimization framework is proposed to reconstruct the object rapidly and reliably within a few iterations. In addition, novel image space denoising regularization is plugged into the loss function to reduce the effects of input noise and improve the perceptual quality of the recovered image. Experiments demonstrate that four-fold resolution gains are achievable for the proposed imaging method. We can obtain satisfactory results for both visual and quantitative metrics with one-sixth of the measurements in the conventional imaging correlography. Therefore, the proposed imaging technique is more suitable for long-range practical applications.