Rational design of hybrid sensor arrays combined synergistically with machine learning for rapid response to a hazardous gas leak environment in chemical plants.

Combinations of semiconductor metal oxide (SMO) sensors, electrochemical (EC) sensors, and photoionization detection (PID) sensors were used to discriminate chemical hazards on the basis of machine learning. Sensing data inputs were exploited in the form of either numerical or image data formats, and the classification of chemical hazards with high accuracy was achieved in both cases. Even a small amount of gas sensing or purging data (input for ∼30 s) input can be exploited in machine-learning-based gas discrimination. SMO sensors exhibit high performance even in a single-sensor mode, presumably because of the intrinsic cross-sensitivity of metal oxides, which is otherwise considered a major disadvantage of SMO sensors. EC sensors were enhanced through synergistic integration of sensor combinations with machine learning. For precision detection of multiple target analytes, a minimum number of sensors can be proposed for gas detection/discrimination by combining sensors with dissimilar operating principles. The Type I hybrid sensor combines one SMO sensor, one EC sensor, and one PID sensor and is used to identify NH3 gas mixed with sulfur compounds in simulations of NH3 gas leak accidents in chemical plants. The portable remote sensing module made with a Type I hybrid sensor and LTE module can identify mixed NH3 gas with a detection time of 60 s, demonstrating the potential of the proposed system to quickly respond to hazardous gas leak accidents and prevent additional damage to the environment.

Synergistic Integration of Machine Learning with Microstructure/Composition-Designed SnO2 and WO3 Breath Sensors.

A high-performance semiconductor metal oxide gas sensing strategy is proposed for efficient sensor-based disease prediction by integrating a machine learning methodology with complementary sensor arrays composed of SnO2- and WO3-based sensors. The six sensors, including SnO2- and WO3-based sensors and neural network algorithms, were used to measure gas mixtures. The six constituent sensors were subjected to acetone and hydrogen environments to monitor the effect of diet and/or irritable bowel syndrome (IBS) under the interference of ethanol. The SnO2- and WO3-based sensors suffer from poor discrimination ability if sensors (a single sensor or multiple sensors) within the same group (SnO2- or WO3-based) are separately applied, even when deep learning is applied to enhance the sensing operation. However, hybrid integration is proven to be effective in discerning acetone from hydrogen even in a two-sensor configuration through the synergistic contribution of supervised learning, i.e., neural network approaches involving deep neural networks (DNNs) and convolutional neural networks (CNNs). DNN-based numeric data and CNN-based image data can be exploited for discriminating acetone and hydrogen, with the aim of predicting the status of an exercise-driven diet and IBS. The ramifications of the proposed hybrid sensor combinations and machine learning for the high-performance breath sensor domain are discussed.

Road Speed Prediction Scheme by Analyzing Road Environment Data.

Road speed is an important indicator of traffic congestion. Therefore, the occurrence of traffic congestion can be reduced by predicting road speed because predicted road speed can be provided to users to distribute traffic. Traffic congestion prediction techniques can provide alternative routes to users in advance to help them avoid traffic jams. In this paper, we propose a machine-learning-based road speed prediction scheme using road environment data analysis. The proposed scheme uses not only the speed data of the target road, but also the speed data of neighboring roads that can affect the speed of the target road. Furthermore, the proposed scheme can accurately predict both the average road speed and rapidly changing road speeds. The proposed scheme uses historical average speed data from the target road organized by the day of the week and hour to reflect the average traffic flow on the road. Additionally, the proposed scheme analyzes speed changes in sections where the road speed changes rapidly to reflect traffic flows. Road speeds may change rapidly as a result of unexpected events such as accidents, disasters, and construction work. The proposed scheme predicts final road speeds by applying historical road speeds and events as weights for road speed prediction. It also considers weather conditions. The proposed scheme uses long short-term memory (LSTM), which is suitable for sequential data learning, as a machine learning algorithm for speed prediction. The proposed scheme can predict road speeds in 30 min by using weather data and speed data from the target and neighboring roads as input data. We demonstrate the capabilities of the proposed scheme through various performance evaluations.

Unified camera tamper detection based on edge and object information.

In this paper, a novel camera tamper detection algorithm is proposed to detect three types of tamper attacks: covered, moved and defocused. The edge disappearance rate is defined in order to measure the amount of edge pixels that disappear in the current frame from the background frame while excluding edges in the foreground. Tamper attacks are detected if the difference between the edge disappearance rate and its temporal average is larger than an adaptive threshold reflecting the environmental conditions of the cameras. The performance of the proposed algorithm is evaluated for short video sequences with three types of tamper attacks and for 24-h video sequences without tamper attacks; the algorithm is shown to achieve acceptable levels of detection and false alarm rates for all types of tamper attacks in real environments.

Complete mitochondrial genome of a troglobite millipede Antrokoreana gracilipes (Diplopoda, Juliformia, Julida), and juliformian phylogeny.

The complete mitochondrial genome of a troglobite millipede Antrokoreana gracilipes (Verhoeff, 1938) (Dipolopoda, Juliformia, Julida) was sequenced and characterized. The genome (14,747 bp) contains 37 genes (2 ribosomal RNA genes, 22 transfer RNA genes and 13 protein-encoding genes) and two large non-coding regions (225 bp and 31 bp), as previously reported for two diplopods, Narceus annularus (order Spirobolida) and Thyropygus sp. (order Spirostreptida). The A + T content of the genome is 62.1% and four tRNAs (tRNA(Ser(AGN)), tRNA(Cys), tRNA(Ile) and tRNA(Met)) have unusual and unstable secondary structures. Whereas Narceus and Thyropygus have identical gene arrangements, the tRNA(Thr) and tRNA(Trp) of Antrokoreana differ from them in their orientations and/or positions. This suggests that the Spirobolida and Spirostreptida are more closely related to each other than to the Dipolopoda. Three scenarios are proposed to account for the unique gene arrangement of Antrokoreana. The data also imply that the Duplication and Nonrandom Loss (DNL) model is applicable to the order Julida. Bayesian inference (BI) and maximum likelihood (ML) analyses using amino acid sequences deduced from the 12 mitochondrial protein-encoding genes (excluding ATP8) support the view that the three juliformian members are monophyletic (BI 100%; ML 100%), that Thyropygus (Spirostreptida) and Narceus (Spirobolida) are clustered together (BI 100%; ML 83%), and that Antrokoreana (Julida) is a sister of the two. However, due to conflict with previous reports using cladistic approaches based on morphological characteristics, further studies are needed to confirm the close relationship between Spirostreptida and Spirobolida.