Video, as a key driver in the global explosion of digital information, can create tremendous benefits for human society. Governments and enterprises are deploying innumerable cameras for a variety of applications, e.g., law enforcement, emergency management, traffic control, and security surveillance, all facilitated by video analytics (VA). This trend is spurred by the rapid advancement of deep learning (DL), which enables more precise models for object classification, detection, and tracking. Meanwhile, with the proliferation of Internet-connected devices, massive amounts of data are generated daily, overwhelming the cloud. Edge computing, an emerging paradigm that moves workloads and services from the network core to the network edge, has been widely recognized as a promising solution. The resulting new intersection, edge video analytics (EVA), begins to attract widespread attention. Nevertheless, only a few loosely-related surveys exist on this topic. The basic concepts of EVA (e.g., definition, architectures) were not fully elucidated due to the rapid development of this domain. To fill these gaps, we provide a comprehensive survey of the recent efforts on EVA. In this paper, we first review the fundamentals of edge computing, followed by an overview of VA. EVA systems and their enabling techniques are discussed next. In addition, we introduce prevalent frameworks and datasets to aid future researchers in the development of EVA systems. Finally, we discuss existing challenges and foresee future research directions. We believe this survey will help readers comprehend the relationship between VA and edge computing, and spark new ideas on EVA.

Data fusion can be defined as the process of combining data or information for estimating the state of an entity. Data fusion is a multidisciplinary field that has several benefits, such as enhancing the confidence, improving reliability, and reducing ambiguity of measurements for estimating the state of entities in engineering systems. It can also enhance completeness of fused data that may be required for estimating the state of engineering systems. Data fusion has been applied to different fields, such as robotics, automation, and intelligent systems. This paper reviews some examples of recent applications of data fusion in civil engineering and presents some of the potential benefits of using data fusion in civil engineering.

Freeway work zone with lane closure has a direct negative impact on travel time, safety, and environmental sustainability. The capacity drop at the onset of the congestion can also further reduce the discharging rate at the work zone area and worsen traffic conditions. Existing studies have developed various variable speed limit (VSL) control methods to mitigate the congestion; however, a simple yet robust VSL control strategy that considers the nonlinearity induced by the capacity drop is still lacking. To address the above-mentioned issue, this study proposes a VSL strategy using a nonlinear traffic flow model and a discrete-time sliding mode control for freeway work zone. The developed traffic flow model incorporates the nonlinearity caused by the capacity drop at the work zone using the cell transmission model. The sliding mode controller is designed to drive the traffic state, which is acquired from the built traffic flow model, to the desired equilibrium state with different convergence rates. Under speed limit constraints, the VSL scheme is generated to regulate the traffic and mitigate the congestion. The proposed system is implemented and evaluated using the traffic microscopic simulator SUMO. The results indicate that the proposed VSL control can consistently improve the traffic mobility, safety, and environmental sustainability under the noisy traffic demand and different control scenarios. Compared with the uncontrolled scenario, the developed system shows improvement by approximately reducing 17% of the average travel time, 90% of the safety risk, and 6% of NOx, CO2 emissions, and fuel consumption.

Over the past decade, several smart and automated systems have been developed to address the issue of struck-by hazards in construction—that is, workers on foot struck by equipment or equipment struck by equipment. False alarms (false positives and false negatives) are common in such systems, but methods for limiting struck-by hazards have not yet been thoroughly studied or tested for real-world implementations. This study presents two novel four-dimensional (4D) [time and three-dimensional (3D) space] models, a time-sphere model and a time-cuboid model, that are effective in reducing the rate of false alarms. In each developed 4D model, (1) entities’ state information, including 3D position, orientation (roll, pitch, and yaw), and velocity, is acquired and analyzed over time; and (2) the hazardous area around equipment or workers is represented by a sphere or a cuboid with the warning distance adjusted and updated according to the entities’ collected state information; and (3) unsafe-proximity qu...

Having an approach that can analyze and identify safety risks in the dynamic and hazardous construction environment is one of the key steps to the success of health and safety plans. Struck-by-equipment hazard is one of the leading causes of construction injuries and fatalities. Therefore, the primary objective of this paper is to investigate and provide an effective method of safety risk analysis on struck-by-equipment hazard. A spatiotemporal network-based model is developed in this paper which performs dynamic risk analysis on the struck-by-equipment hazard at both entity and network levels. The developed model performs safety risk analysis in a real-time and proactive manner by considering the spatiotemporal interactions among all construction entities (equipment and workers on foot) across the jobsite. Three risk factors of struck-by-equipment hazards were selected, including proximity, blind spots, and velocity. The interactions of the selected factors between entities are quantified and used to generate the network. Three safety leading indicators—degree centrality, eigenvector centrality, and relative risk score—are adopted to represent the risk levels of individual entities and jobsites. The implementation and evaluation of the network-based model with the three leading indicators are conducted and illustrated based on four simulated jobsites. Accordingly, the derived practical applications for risk analysis and hazard identification and prevention are described. The work presented in this paper enables the quantification and analysis of struck-by-equipment risk for both individual entities and jobsites from multiple perspectives. Further insight into the temporal aspects of risk and the safety performance of entities, as well as jobsites, is gained to proactively identify and prevent hazards.