In this paper, the decentralized optimal control problem is addressed for a class of large-scale systems subject to injection attacks. All subsystem matrices are considered to be unavailable to the designer. A model-free decentralized sliding mode control (SMC) scheme for each subsystem is designed via just utilizing its own state information and the known bounds of the interconnections and the injection attacks. Moreover, the adaptive dynamic programming (ADP) approach is incorporated to deal with the infinite horizon optimal control problem for the sliding mode dynamics, which is equivalent to the solution of a set of parallel algebraic Riccati equations. Furthermore, a novel parallel policy iteration algorithm is developed to implement the proposed decentralized SMC scheme without using all subsystems dynamics matrices. Specifically, it is shown that during the whole policy iteration steps, the reachability of each sliding variable and the stability of each sliding mode dynamics are guaranteed simultaneously by the online updating decentralized SMC scheme. Finally, the applicability of the proposed novel ADP-based decentralized SMC strategy is illustrated by a two-machine power system subject to three different injection attacks.

Incremental learning of semantic segmentation has emerged as a promising strategy for visual scene interpretation in the open-world setting. However, it remains challenging to acquire novel classes in an online fashion for the segmentation task, mainly due to its continuously-evolving semantic label space, partial pixelwise ground-truth annotations, and constrained data availability. To address this, we propose an incremental learning strategy that can fast adapt deep segmentation models without catastrophic forgetting, using a streaming input data with pixel annotations on the novel classes only. To this end, we develop a unified learning strategy based on the Expectation-Maximization (EM) framework, which integrates an iterative relabeling strategy that fills in the missing labels and a rehearsal-based incremental learning step that balances the stability-plasticity of the model. Moreover, our EM algorithm adopts an adaptive sampling method to select informative training data and a class-balancing training strategy in the incremental model updates, both improving the efficacy of model learning. We validate our approach on the PASCAL VOC 2012 and ADE20K datasets, and the results demonstrate its superior performance over the existing incremental methods.

Digital microfluidic (DMF) has been a unique tool for manipulating micro-droplets with high flexibility and accuracy. To extend the application of DMF for automatic and in-site detection, it is promising to introduce colorimetric sensing based on gold nanoparticles (AuNPs), which have advantages including high sensitivity, label-free, biocompatibility, and easy surface modification. However, there is still a lack of studies for investigating the movement and stability of AuNPs for in-site detection on the electrowetting-based digital microfluidics. Herein, to demonstrate the ability of DMF for colorimetric sensing with AuNPs, we investigated the electrowetting property of the AuNPs droplets on the hydrophobic interface of the DMF chip and examined the stability of the AuNPs on DMF as well as the influence of evaporation to the colorimetric sensing. As a result, we found that the electrowetting of AuNPs fits to a modified Young–Lippmann equation, which suggests that a higher voltage is required to actuate AuNPs droplets compared with actuating water droplets. Moreover, the stability of AuNPs was maintained during the processing of electrowetting. We also proved that the evaporation of droplets has a limited influence on the detections that last several minutes. Finally, a model experiment for the detection of Hg²⁺ was carried out with similar results to the detections in bulk solution. The proposed method can be further extended to a wide range of AuNPs-based detection for label-free, automatic, and low-cost detection of small molecules, biomarkers, and metal ions.

Indium tin oxide (ITO)-based digital microfluidics (DMF) with unique optical and electrical properties are promising in the development of integrated, automatic and portable analytical systems. The fabrication technique using laser direct etching (LDE) on ITO glass has the advantages of being rapid, low cost and convenient. However, the fabrication resolution of LDE limits the minimum line width for patterns on ITO glasses, leading to a related wider lead wire for the actuating electrodes of DMF compared with photolithography. Therefore, the lead wire of electrodes could affect the droplet motion on the digital microfluidic chip due to the increased contact line with the droplet. Herein, we developed a finite element model of a DMF with improved efficiency to investigate the effect of the lead wire. An optimized electrode pattern was then designed based on a theoretical analysis and validated by a simulation, which significantly decreased the deformation of the droplets down to 0.012 mm. The performance of the optimized electrode was also verified in an experiment. The proposed simulation method could be further extended to other DMF systems or applications to provide an efficient approach for the design and optimization of DMF chips.

Digital microfluidics is a novel technique for manipulating discrete droplets with the advantages of programmability, small device size, low cost, and easy integration. The development of droplet sensing methods advances the automation control of digital microfluidics. Impedance measurement emerges as a promising technique for droplet localization and characterization due to its non-invasive nature, high sensitivity, simplicity, and cost-effectiveness. However, traditional impedance measurement approaches in digital microfluidics based on the high-voltage actuating signal are limited in sensing accuracy in practical applications. In this paper, we propose a novel droplet impedance sensing system for digital microfluidics by introducing a low-voltage and addressable measurement circuit, which enables impedance measurement over a wide frequency range. The proposed measurement system has also been used for detecting the droplet composition, size, and position in a digital microfluidic chip. The improved impedance sensing method can also promote the applications of the digital microfluidic, which requires high accuracy, real-time, and contactless sensing with automatic sample pretreatment.

In recent years, deep learning has been widely utilized in the fields of biomedical image segmentation and cellular image analysis. Supervised deep neural networks trained on annotated data have demonstrated good performance in tasks related to cell and nucleus segmentation. However, the use of supervised models necessitates carefully constructed training data and a substantial amount of ground truth information. Unfortunately, high-quality annotated data for cellular images are scarce. To address the issue of limited datasets, we propose a cell and nucleus segmentation method based on deep learning multimodal fusion. The proposed method includes three modules: a segmentation fundamental module, a multimodal prompter module, and an object output module. This comprehensive approach enables cell and nucleus segmentation tasks to be performed without the need for retraining on new data. The segmentation fundamental module is the core of the framework, as it provides essential segmentation capabilities. By leveraging preexisting models trained on natural imagery, this module effectively performs cell segmentation by incorporating prior knowledge. The multimodal prompter module, a pretrained model, aids in combining image and textual information. It employs a data fusion technique for multiple modalities to deliver prompts that steer the network's output, thereby avoiding the constraints inherent to single-modality approaches. The object output module combines the inputs from the preceding modules to generate the final segmentation output. The experimental validation confirms the superiority of the proposed method, which outperforms comparative methods in cell and nucleus segmentation tasks and has promise for future applications in cell tracking.