Deep Reinforcement Learning based pedestrian and independent vehicle safety fortification using intelligent perception

The Light Detection and Ranging (LiDAR) sensor is utilized to track each sensed obstructions at their respective locations with their relative distance, speed, and direction; such sensitive information forwards to the cloud server to predict the vehicle-hit, traffic congestion and road damages. Learn the behaviour of the state to produce an appropriate reward as the recommendation to avoid tragedy. Deep Reinforcement Learning and Q-network predict the complexity and uncertainty of the environment to generate optimal reward to states. Consequently, it activates automatic emergency braking and safe parking assistance to the vehicles. In addition, the proposed work provides safer transport for pedestrians and independent vehicles. Compared to the newer methods, the proposed system experimental results achieved 92.15% higher prediction rate accuracy. Finally, the proposed system saves many humans, animal lives from the vehicle hit, suggests drivers for rerouting to avoid unpredictable traffic, saves fuel consumption, and avoids carbon emission.

Motion estimation in vehicular environments based on Bayesian dynamic networks

The cornerstone to achieve the development of autonomous ground driving with the lowest possible risk of collision in real traffic environments is the movement estimation obstacle. Predicting trajectories of multiple obstacles in dynamic traffic scenarios is a major challenge, especially when different types of obstacles such as vehicles and pedestrians are involved. According to the issues mentioned, in this work a novel method based on Bayesian dynamic networks is proposed to infer the paths of interest objects (IO). Environmental information is obtained through stereo video, the direction vectors of multiple obstacles are computed and the trajectories with the highest probability of occurrence and the possibility of collision are highlighted. The proposed approach was evaluated using test environments considering different road layouts and multiple obstacles in real-world traffic scenarios. A comparison of the results obtained against the ground truth of the paths taken by each detected IO is performed. According to experimental results, the proposed method obtains a prediction rate of 75% for the change of direction taking into consideration the risk of collision. The importance of the proposal is that it does not obviate the risk of collision in contrast with related work.

Targeted On-Demand Screening of Pesticide Panel in Soil Runoff

Using pesticides is a common agricultural and horticultural practice to serve as a control against weeds, fungi, and insects in plant systems. The application of these chemical agents is usually by spraying them on the crop or plant. However, this methodology is not highly directional, and so only a fraction of the pesticide actually adsorbs onto the plant, and the rest seeps through into the soil base contaminating its composition and eventually leaching into groundwater sources. Electrochemical sensors which are more practical for in situ analysis used for pesticide detection in soil runoff systems are still in dearth, while the ones published in the literature are attributed with complex sensor modification/functionalization and preprocessing of samples. Hence, in this work, we present a highly intuitive electroanalytical sensor approach toward rapid (10 min), on-demand screening of commonly used pesticides—glyphosate and atrazine—in soil runoff. The proposed sensor functions based on the affinity biosensing mechanism driven via thiol cross-linker and antibody receptors that holistically behaves as a recognition immunoassay stack that is specific and sensitive to track test pesticide analytes. Then, this developed sensor is integrated further to create a pesticide-sensing ecosystem using a front-end field-deployable smart device. The method put forward in this work is compared and validated against a standard laboratory potentiostat instrument to determine efficacy, feasibility, and robustness for a point-of-use (PoU) setting yielding LoD levels of 0.001 ng/ml for atrazine and 1 ng/ml for glyphosate. Also, the ML model integration resulted in an accurate prediction rate of ≈80% in real soil samples. Therefore, a universal pesticide screening analytical device is designed, fabricated, and tested for pesticide assessment in real soil runoff samples.

Discrimination of Four Cinnamomum Species with Physico-Functional Properties and Chemometric Techniques: Application of PCA and MDA Models

Discrimination of highly valued and non-hepatotoxic Cinnamomum species (C. verum) from hepatotoxic (C. burmannii, C. loureiroi, and C. cassia) is essential for preventing food adulteration and safety problems. In this study, we developed a new method for the discrimination of four Cinnamomum species using physico-functional properties and chemometric techniques. The data were analyzed through principal component analysis (PCA) and multiclass discriminant analysis (MDA). The results showed that the cumulative variability of the first three principal components was 81.70%. The PCA score plot indicated a clear separation of the different Cinnamomum species. The training set was used to build the discriminant MDA model. The testing set was verified by this model. The prediction rate of 100% proved that the model was valid and reliable. Therefore, physico-functional properties coupled with chemometric techniques constitute a practical approach for discrimination of Cinnamomum species to prevent food fraud.

An ECG Heartbeat Classification Method Based on Deep Convolutional Neural Network

The electrocardiogram (ECG) is one of the most powerful tools used in hospitals to analyze the cardiovascular status and check health, a standard for detecting and diagnosing abnormal heart rhythms. In recent years, cardiovascular health has attracted much attention. However, traditional doctors’ consultations have disadvantages such as delayed diagnosis and high misdiagnosis rate, while cardiovascular diseases have the characteristics of early diagnosis, early treatment, and early recovery. Therefore, it is essential to reduce the misdiagnosis rate of heart disease. Our work is based on five different types of ECG arrhythmia classified according to the AAMI EC57 standard, namely, nonectopic, supraventricular ectopic, ventricular ectopic, fusion, and unknown beat. This paper proposed a high-accuracy ECG arrhythmia classification method based on convolutional neural network (CNN), which could accurately classify ECG signals. We evaluated the classification effect of this classification method on the supraventricular ectopic beat (SVEB) and ventricular ectopic beat (VEB) based on the MIT-BIH arrhythmia database. According to the results, the proposed method achieved 99.8% accuracy, 98.4% sensitivity, 99.9% specificity, and 98.5% positive prediction rate for detecting VEB. Detection of SVEB achieved 99.7% accuracy, 92.1% sensitivity, 99.9% specificity, and 96.8% positive prediction rate.

Comparing ecological momentary assessment to sensor-based approaches in predicting dietary lapse

Ecological momentary assessment (EMA; brief self-report surveys) of dietary lapse risk factors (e.g., cravings) has shown promise in predicting and preventing dietary lapse (nonadherence to a dietary prescription), which can improve weight loss interventions. Passive sensors also can measure lapse risk factors and may offer advantages over EMA (e.g., objective, automatic, semicontinuous data collection), but currently can measure only a few lapse predictors, a notable limitation. This study preliminarily compared the burden and accuracy of commercially available sensors versus established EMA in lapse prediction. N = 23 adults with overweight/obesity completed a 6-week commercial app-based weight loss program. Participants wore a Fitbit, enabled GPS tracking, completed EMA, and reported on EMA and sensor burden poststudy via a 5-point Likert scale. Sensed risk factors were physical activity and sleep (accelerometer), geolocation (GPS), and time, from which 233 features (measurable characteristics of sensor signals) were extracted. EMA measured 19 risk factors, lapse, and categorized GPS into meaningful geolocations. Two supervised binary classification models (LASSO) were created: the sensor model predicted lapse with 63% sensitivity (true prediction rate of lapse) and 60% specificity (true prediction rate of non-lapse) and EMA model with 59% sensitivity and 72% specificity. EMA model accuracy was higher, but self-reported EMA burden (M = 2.96, SD = 1.02) also was higher (M = 1.50, SD = 0.94). EMA model accuracy was superior, but EMA burden was higher than sensor burden. Findings highlight the promise of sensors in contributing to lapse prediction, and future research may use EMA, sensors, or both depending on prioritization of accuracy versus participant burden.

Troponin I Levels in Neonatal Hypoxic–Ischemic Encephalopathy Are Related to Cardiopulmonary Comorbidity and Neurodevelopmental Outcomes

Troponin I is a biomarker for cardiac injury in children. The role of troponin I in neonatal Hypoxic–Ischemic encephalopathy (HIE) may have valuable clinical implications. Troponin I levels were measured within 6 h of birth to determine their relationship to HIE stage, short-term cardiac functional outcomes, and neurodevelopmental outcomes at 1 year. Seventy-three patients were divided into two groups: mild HIE and moderate to severe HIE. Troponin I levels within 6 h of birth were obtained in 61 patients, and were significantly higher in patients with moderate to severe HIE than in patients with mild HIE (Mann–Whitney U test, U = 146, p = 0.001). A troponin I cut-off level of ≥60 pg/mL predicted moderate to severe HIE with a specificity of 81.1% and a negative prediction rate of 76.9%. A troponin I cut-off level of ≥180 pg/mL was significantly (χ2 (1, n = 61) = 33.1, p = 0.001, odds ratio 96.8) related with hypotension during first admission and significantly (χ2 (1, n = 61) = 5.3, p = 0.021, odds ratio 4.53) related with abnormal neurodevelopmental outcomes at 1 year. Early troponin I level may be a useful biomarker for predicting moderate to severe HIE, and initialization of hypothermia therapy.

Evaluation of Landslide Susceptibility of Şavşat District of Artvin Province (Turkey) Using Machine Learning Techniques

The aim of this study is to produce landslide susceptibility maps of Şavşat district of Artvin Province using machine learning (ML) models and to compare the predictive performances of the models used. Tree-based ensemble learning models, including random forest (RF), gradient boosting machines (GBM), and extreme gradient boosting (XGBoost), were used in the study. A landslide inventory map consisting of 85 landslide polygons was used in the study. The inventory map comprises 32,777 landslide pixels at 30 m resolution. Randomly selected 70% of the landslide pixels were used for training the models and the remaining 30% were used for the validation of the models. In susceptibility analysis, altitude, aspect, curvature, distance to drainage network, distance to faults, distance to roads, land cover, lithology, slope, slope length, and topographic wetness index parameters were used. The validation of the models was conducted using success and prediction rate curves. The validation results showed that the success rates for the GBM, RF, and XGBoost models were 91.6%, 98.4%, and 98.6%, respectively, whereas the prediction rate were 91.4%, 97.9%, and 98.1%, respectively. Therefore, it was concluded that landslide susceptibility map produced with XGBoost model can help decision makers in reducing landslide-associated damages in the study area.

Phenylalanylphenylalanine as a Diagnostic Biomarker for Lung Cancer and Tuberculosis.

Background: Worldwide, lung cancer has the highest mortality rate, and pulmonary tuberculosis has a high incidence in China, and both may be misdiagnosed frequently because of similar clinical presentation and atypical imaging findings. Diagnostic biomarkers to distinguish between lung cancer and other pulmonary diseases can be detected by metabolomics to avoid non-essential treatment.Methods: This cohort study employed non-targeted and targeted metabolomic analysis in participants enrolled from three independent centers. Multivariate statistics, variable importance in the projection parameter, receiver operating characteristics (ROC) curve were used to build potential key diagnostic biomarkers model of lung cancer and these were subsequently analyzed using targeted metabolomics in test set. Quantitative analysis of differences in biomarker levels was conducted, and a support vector machine (SVM) classifier was used to identify the prediction rate of diagnostic biomarker model. Results: Phenylalanylphenylalanine showed opposite trends in lung cancer and tuberculosis. The area under the curve 0.8887 (95% CI 0.8064–0.9710, p<0.001, sensitivity 85.45%, specificity 84%), 0.8149 (95% CI 0.7419–0.8878, p<0.001, the sensitivity was 73.26%, the specificity was 78.43%) and SVM results (prediction rate 77.94%) showed the feasibility of using phenylalanylphenylalanine as a diagnostic marker for the differential diagnosis of lung cancer and tuberculosis.Conclusions: Changes in the levels of phenylalanylphenylalanine facilitate differential diagnosis between lung cancer and tuberculosis, thereby potentially reducing the damage caused by misdiagnosis in the clinical setting, and enabling early treatment of lung cancer patients.Trial registration: This study is registered in the China Clinical Trial Registration Center (registration number ChiCTR2000040666, Registered 07 December 2020, http://www.chictr.org.cn/index.aspx)

Analysis of Landslide Susceptibility Using Deep Neural Network

Landslides in Korea are caused by various factors, such as topographic characteristics, geology, and climate change, and they cause significant damage to property and human life. It is necessary to analyze landslide susceptibility to identify the location of landslide occurrence precisely and respond to the risk of landslides. In this study, the probability of landslide occurrence was calculated through a landslide sensitivity analysis using a deep neural network based on eight conditioning factors and 26 landslide data. In addition, verification was performed using the ROC method. The landslide susceptibility obtained using a deep neural network showed a success rate of 70% and a prediction rate of 81.7%, indicating that the prediction rate was 11.7% higher than the success rate. In addition, a landslide susceptibility map for estimating the probability of landslide occurrence was plotted using the geometric spacing method. The chi-square test results indicated that the landslide susceptibility map obtained in this study was statistically significant. The location of landslides can be identified more accurately using the proposed method.