Pre-built Library Research Articles

Model-based learning of coordinators of the decentralized multi-zone objects control systems

Decentralized control systems are gaining more and more expansion, which is due to the increase in the availability and power of microcontrollers. Decentralized control of multi-zone objects is associated with the need to coordinate the local control systems of zones state. Learning systems are preferred for implementation of the coordination methods, as they are able for flexibly adjust to the specifics of control of each zone. However, the training of coordinators is complicated task by the absence at the stage of a system creating of marked datasets for controlled multi-zonal objects. This article considers the creation of a dataset based on a simulation of a decentralized system and four scenarios for training neural coordinators. A model for simulation of a decentralized system was been created on the Scilab/Xcos platform using a pre-built library of blocks for simulating decentralized systems. The scenarios differ depending on the structure of the neural coordinators: a segmented network according to the structure of the coordinator simulation model or an integrated one, as well as on the training strategy: train all the coordinators of the decentralized system in parallel or only one coordinator and then clone the results. Experimental studies of the proposed method of training neural network coordinators, implemented on Python TensorFlow, were conducted. The study showed greater effectiveness of segmented coordinators parallel training. However, in the course of the study, the last step of the scenarios – fine tuning on a real physical object, was not performed. A preliminary evaluation suggests that after such additional training, the advantages of mono-neural coordinators will become more visible, since such additional training will correct the shortcomings of imitation.

AutoBPS-Prototype: A web-based toolkit to automatically generate prototype building energy models with customizable efficiency values in China

Prototype building energy models are significant in building energy performance evaluation and energy efficiency potential analysis. However, their establishment process is time-consuming and error-prone. This study aims to develop a user-friendly toolkit, AutoBPS-Prototype, to automatically establish prototype building energy models in China. A web-based graphical user interface of AutoBPS-Prototype was designed and implemented. AutoBPS-Prototype only needs to define the climate zone, building type, and standard vintage to generate a default prototype building energy model based on a pre-built library. In addition, it can generate user-defined models through customized important efficiency values. The OpenStudio Standards used to create the U.S. Department of Energy prototype building models based on the ASHRAE 90.1 standards was extended to cover more prototype building energy models in China, including 20 sub-climate zones, 22 building types and four code and standard vintages. The Chinese building standards were used for the efficiency settings. Besides, case studies were conducted. Firstly, 20 prototype building models of four building types in five cities were generated. The credibility of the model results simulated by EnergyPlus was verified through comparison with measured data. Then, the impact of spatiotemporal differences on building energy consumption was analyzed through simulation of the models using different standard vintages in different climate zones. Finally, a baseline and four upgrade scenarios were designed to explore the impact of various upgrades on building energy consumption. This toolkit aims to assist users in conducting rapid research on building energy efficiency improvement. More toolkit features are under development for retrofit analysis, demand response analysis, and photovoltaics system generation estimation.

Knowledge-informed data-driven modeling for sparse identification of governing equations for microbial inactivation processes in food

Prevention of the growth of harmful microorganisms in food products is an important requirement for ensuring food safety and quality. Mathematical models to predict the quantitative changes in microbial populations in food to the variations of environmental conditions are useful tools in this regard. While equations for microbial inactivation have typically been formulated based on polynomial functions, empirical choice of the model order and terms not only results in over- or underfitting, but also makes it difficult to identify key factors governing the target variable. To address this issue, we present a data-driven modeling pipeline that enables 1) automatic discovery of model equations through parsimonious selection of relevant terms from a pre-built library and 2) subsequent evaluation of the impacts of individual terms on the model output. Through case studies using literature data, we evaluated the effectiveness of our pipeline in predicting the D-value (i.e., the time taken to reduce microbial population to 10% of the initial level) as a function of multiple factors including temperature, pH, water activity, NaCl content, and phosphate level. In doing this, we determined basic functional forms of input and output variables based on their pre-known relationships, e.g., by accounting for the Arrhenius dependence of D-value on temperature. Incorporation of such theoretical knowledge into the pipeline improved model accuracy. Using the Akaike information criterion, we optimally determined hyperparameters that control a trade-off between model accuracy and sparsity. We found the literature models benchmarked in this study to be over- or under-determined and consequently proposed better structured and more accurate equations. The subsequent global sensitivity analysis allowed us to evaluate the context-dependent impacts of key factors on the D-value. The pipeline presented in this work is readily applicable to many other related non-linear systems without being limited to microbial inactivation datasets.

Spectral partition correlation based on Voigt function for Raman spectral library search

Raman spectral library search revolves around matching unknown query spectra with reference spectra in library based on similarity metrics, which plays an important role in the field of rapid detection. However, in spectral library search, query spectra and reference spectra are normally obtained by different spectrometers, so peaks within the same range always suffer inconsistent intensities, which exerts a negative influence on calculating spectral similarity metrics. In this paper, a new method named spectral partition correlation based on Voigt function (SPCV) is proposed to advance Raman spectral library search. After normal spectral preprocessing, Voigt function is utilized to fit and partition each unknown query spectrum and its corresponding reference spectrum into segments according to the distribution of Raman peaks and the spectral regions not containing any peaks are ignored. Subsequently, univariate linear regression is applied to correct each pair of segments to decrease the inconsistent peak intensity. The spectra of hexane measured by two different Raman spectrometers are selected as a case study to demonstrate that SPCV can remarkably and effectively increase the values of Pearson's correlation coefficient and ameliorate the performance of spectral matching. Moreover, 252 Raman query spectra and the pre-built library with 14033 reference spectra are also used to further verify the performance of SPCV. As a result, the matching accuracy rate of SPCV is the highest compared with hit quality index (HQI) and weighted segmental hit quality index (SHQI). Thus, SPCV can be a promising candidate for enhancing the performance of library-based Raman spectral matching.

Traffic surveillance video coding with libraries of vehicles and background

This paper presents a video coding scheme tailored for traffic surveillance videos, which features a pre-built library that is utilized in both encoder and decoder to pursue higher compression efficiency. We are motivated by the observation that, in traffic surveillance videos, not only the background is steady for a long while, but also the foreground (e.g. vehicles) contains high redundancy. For example, in the video taken by a traffic monitoring camera, we can observe that the passing-through vehicles are usually similar. However, the redundancy in the vehicles and the background is not fully exploited in the current video coding schemes. To address this problem, we propose a library-based video coding scheme. Specifically, for each static monitoring camera, we can collect video in a period, and build a library of vehicles and background from that video. When encoding the following video of that camera, we utilize the pre-built library by searching similar vehicles and background from the library and using the retrieved vehicles and background to help encode. Accordingly, the decoder also refers to the same library to reconstruct the video. We design efficient algorithms for building the library, searching in the library, as well as utilizing the library for encoding/decoding, to fulfill the proposed scheme. Our scheme is implemented upon the state-of-the-art video coding system–High Efficiency Video Coding (HEVC), and is tested on our own collected traffic surveillance videos. Experimental results show that, compared to the HEVC anchor, our proposed scheme achieves as high as 47.0%, 37.1%, and 34.8% BD-rate reduction, under random-access, low-delay B, and low-delay P settings, respectively. Our scheme also achieves higher compression efficiency than the existing background-based methods for surveillance video coding.

SWATH-ID: An instrument method which combines identification and quantification in a single analysis.

Data-independent acquisition (DIA) approaches, such as SWATH® -MS, are showing great potential to reliably quantify significant numbers of peptides and proteins in an unbiased manner. These developments have enhanced interest in developing a single DIA method that integrates qualitative and quantitative analysis, eliminating the need of a prebuilt library of peptide spectra, which are created through data-dependent acquisition methods or from public repositories. Here, we introduce a new DIA approach, referred to as "SWATH-ID," which was developed to allow peptide identification as well as quantitation. The SWATH-ID method is composed of small Q1 windows, achieving better selectivity and thus significantly improving high-confidence peptide extractions from data files. Furthermore, the SWATH-ID approach transmits precursor ions without fragmentation as well as their fragments within the same SWATH acquisition period. This provides a single scan that includes all precursor ions within the isolation window as well as a record of all of their fragment ions, substantially negating the need for a survey scan. In this way all precursors present in a small Q1 window are associated with their fragment ions, improving the identification specificity and providing a more comprehensive and in-depth view of protein and peptide species in complex samples.

Automatic ontology generation from patents using a pre-built library, WordNet and a class-based n-gram model

An ontology is defined as a structured, hierarchical way for describing domain knowledge. Research work regarding ontological engineering has yielded fruitful results, but these methods share a common drawback: they require significant manual work to generate an ontology, which limits the usefulness of these approaches in practice. In this paper, we propose a computational model that combines data mining, Natural Language Processing (NLP), WordNet and a novel class-based n-gram model for automatic ontology discovery and recognition from existing patent documents. A pre-built ontology library was constructed by gathering knowledge from engineering textbooks and dictionaries. Then a data set of engineering patent claims was split into training (80%) and validation (20%) subsets. The pre-built library and WordNet were used to generate class labels for constructing class-based n-gram models in a training process. The holdout validation showed that the average accuracy was 87.26% for all validation samples.