Operational Evaluation of Uncontrolled Intersections using Artificial Intelligence Techniques

The objective of this paper is to develop and assess motorized level of service models using advanced artificial intelligence techniques like functional networks, multi-gene genetic programming and multi-variate auto-regressive spline for uncontrolled intersections under mixed traffic conditions. Thirteen intersections from India are chosen and geometric, traffic and roadside environmental data are collected using high-definition cameras. An innovative Influence for Gap Acceptance (INA) method for critical gap and follow-up-time measurement is also developed. About seven thousand effective uncontrolled intersection motorized driver responses are collected based on user satisfaction scores (1=excellent to 6=awfully bad). Eight variables with high significance effects on perceived scores from Spearman’s correlation technique are modelled. The proposed functional network model showed better efficiency with volume to capacity ratio, percentage of on-street parking and service delay showing supreme effects in level of service predictions. The imperative outcomes of this research may help transport planners and traffic engineers to quantify operational evaluation of uncontrolled intersections and take crucial decisions towards their improvement.

YOLO with adaptive frame control for real-time object detection applications

You only look once (YOLO) is being used as the most popular object detection software in many intelligent video applications due to its ease of use and high object detection precision. In addition, in recent years, various intelligent vision systems based on high-performance embedded systems are being developed. Nevertheless, the YOLO still requires high-end hardware for successful real-time object detection. In this paper, we first discuss real-time object detection service of the YOLO on AI embedded systems with resource constraints. In particular, we point out the problems related to real-time processing in YOLO object detection associated with network cameras, and then propose a novel YOLO architecture with adaptive frame control (AFC) that can efficiently cope with these problems. Through various experiments, we show that the proposed AFC can maintain the high precision and convenience of YOLO, and provide real-time object detection service by minimizing total service delay, which remains a limitation of the pure YOLO.

Computational Offloading in Mobile Edge with Comprehensive and Energy Efficient Cost Function: A Deep Learning Approach

In mobile edge computing (MEC), partial computational offloading can be intelligently investigated to reduce the energy consumption and service delay of user equipment (UE) by dividing a single task into different components. Some of the components execute locally on the UE while the remaining are offloaded to a mobile edge server (MES). In this paper, we investigate the partial offloading technique in MEC using a supervised deep learning approach. The proposed technique, comprehensive and energy efficient deep learning-based offloading technique (CEDOT), intelligently selects the partial offloading policy and also the size of each component of a task to reduce the service delay and energy consumption of UEs. We use deep learning to find, simultaneously, the best partitioning of a single task with the best offloading policy. The deep neural network (DNN) is trained through a comprehensive dataset, generated from our mathematical model, which reduces the time delay and energy consumption of the overall process. Due to the complexity and computation of the mathematical model in the algorithm being high, due to trained DNN the complexity and computation are minimized in the proposed work. We propose a comprehensive cost function, which depends on various delays, energy consumption, radio resources, and computation resources. Furthermore, the cost function also depends on energy consumption and delay due to the task-division-process in partial offloading. None of the literature work considers the partitioning along with the computational offloading policy, and hence, the time and energy consumption due to task-division-process are ignored in the cost function. The proposed work considers all the important parameters in the cost function and generates a comprehensive training dataset with high computation and complexity. Once we get the training dataset, then the complexity is minimized through trained DNN which gives faster decision making with low energy consumptions. Simulation results demonstrate the superior performance of the proposed technique with high accuracy of the DNN in deciding offloading policy and partitioning of a task with minimum delay and energy consumption for UE. More than 70% accuracy of the trained DNN is achieved through a comprehensive training dataset. The simulation results also show the constant accuracy of the DNN when the UEs are moving which means the decision making of the offloading policy and partitioning are not affected by the mobility of UEs.

Internet of Things-Based Energy Efficiency Optimization Model in Fog Smart Cities

In recent years, the IoT) Internet of Things (IoT) allows devices to connect to the Internet that has become a promising research area mainly due to the constant emerging of the dynamic improvement of technologies and their associated challenges. In an approach to solve these challenges, fog computing came to play since it closely manages IoT connectivity. Fog-Enabled Smart Cities (IoT-ESC) portrays equitable energy consumption of a 7% reduction from 18.2% renewable energy contribution, which extends resource computation as a great advantage. The initialization of IoT-Enabled Smart Grids including (FESC) like fog nodes in fog computing, reduced workload in Terminal Nodes services (TNs) that are the sensors and actuators of the Internet of Things (IoT) set up. This paper proposes an integrated energy-efficiency model computation about the response time and delays service minimization delay in FESC. The FESC gives an impression of an auspicious computing model for location, time, and delay-sensitive applications supporting vertically -isolated, service delay, sensitive solicitations by providing abundant, ascendable, and scattered figuring stowage and system associativity. We first reviewed the persisting challenges in the proposed state-of-the models and based on them. We introduce a new model to address mainly energy efficiency about response time and the service delays in IoT-ESC. The iFogsim simulated results demonstrated that the proposed model minimized service delay and reduced energy consumption during computation. We employed IoT-ESC to decide autonomously or semi-autonomously whether the computation is to be made on Fog nodes or its transfer to the cloud.

Application of Cluster Analysis to Define Level of Service Criteria of U-turns at Median Openings

Defining Level of Service (LOS) criteria of U-turns is important for proper planning, design of transportation projects and also allocating resources. The present study attempts to establish a framework to define LOS criteria of U-turns keeping in mind the peculiar behavior of drivers and heterogeneity in urban Indian context. The U-turns at uncontrolled (no traffic sign, no signal, no traffic personnel) median openings are very risky. Upon arrival at the median opening, the U-turning vehicle looks for a suitable gap in the approaching traffic stream before initiating the merging process. While waiting for a suitable gap the U-turning vehicle experiences service delay. This service delay has been studied to quantify the delay ranges for different LOS categories. In this study, service delay data were collected from 7 different sections and microscopic analysis procedure was adopted to extract data from the recorded video. Subsequently, clustering technique has been utilized to defining delay ranges of different level of service categories. Four clustering methods, namely; K-mean, K-medoid, Affinity Propagation (AP), and Fuzzy C-means (FCM) are used. Four validation parameters are applied to determine most suitable clustering algorithm for the study and to determine the optimal number of cluster. AP was found to be the most suitable clustering method and 6 was found to be the optimal number and accordingly the collected delay data were clustered into 6 categories using AP. The delay range is found to be less than 4 s for LOS A is greater than 35 s for LOS F.

Time to Sputum Smear Conversion and Its Predictors Among Smear-Positive Pulmonary TB Patients in East Gojjam Zone, Northwest Ethiopia: A Longitudinal Study.

Background: Sputum smear conversion is a key indicator of treatment response and reduced infectivity among smear-positive pulmonary tuberculosis (SPPTB). This study aimed at estimating sputum smear conversion time and identifying factors hindering sputum smear conversion among SPPTB patients in East Gojjam Zone, Northwest Ethiopia. Methods: A total of 282 SPPTB patients were followed for 22 weeks through weekly sputum smear evaluation. Due to the absence of sputum culture and rapid diagnostic services, sputum smear conversion evaluation was conducted microscopically using acid-fast-bacilli staining technique of sediments from a 5% sodium hypochlorite concentration technique. Data on socio-demographic, clinical profile, and personal behavior variables were collected using a pretested interviewer-administered questionnaire. Various descriptive statistics including mean, median with interquartile range (IQR), proportions, and cross-tabulations were computed. Factors affecting sputum smear conversion were identified by multivariable logistic regression analysis. The statistical significance of variables was determined at a p-value < 0.05. Results: Over half, 166(59%) of SPPTB cases were males and 147(52%) were rural dwellers. The mean age of respondents was 35±5SD years. About 88(31.2%) of SPPTB patients had comorbidities, 102(36.2%) faced stigma, and 54(19%) smoked a cigarette. The median sputum smear conversion time at the end of the intensive phase was 35 days (IQR: 21 -56 days). The majority, 85% (95%CI: 76% -93%) and 95% (95% CI: 85% -99%) of SPPTB patients underwent sputum smear conversion at the end of 2nd and 5th months of treatment, respectively. Poor knowledge on TB, being HIV positive, higher smear grading, having diabetes mellitus, undernutrition, cigarette smoking, facing societal stigma, and TB service delay were positively associated with the length of sputum smear conversion. Conclusion: Although the treatment success rate was 95%, the median sputum smear conversion time was higher compared to the TB program expectations and some former studies. Factors of sputum smear conversion were related to nutritional status, smear grading, comorbidity status, knowledge on TB, personal behavior, stigma, and TB service delay. Improving the health literacy of the community by revising the existing community awareness strategies is essential to enhance treatment adherence and lower infectiousness after treatment initiation.

Predictive Migration Performance in Vehicular Edge Computing Environments

Advanced learning algorithms for autonomous driving require lots of processing and storage power, which puts a strain on vehicles’ computing resources. Using a combination of 5G network connectivity with ultra-high bandwidth and low latency together with extra computing power located at the edge of the network can help extend the capabilities of vehicular networks. However, due to the high mobility, it is essential that the offloaded services are migrated so that they are always in close proximity to the requester. Using proactive migration techniques ensures minimum latency for high service quality. However, predicting the next edge server to migrate comes with an error that can have deteriorating effects on the latency. In this paper, we examine the influence of mobility prediction errors on edge service migration performances in terms of latency penalty using a large-scale urban vehicular simulation. Our results show that the average service delay increases almost linearly with the migration prediction error, with 20% error yielding almost double service latency.