Search for: All records

Recent advancements in wearable physiological sensing and artificial intelligence have made some remarkable progress in workers’ health monitoring in construction sites. However, the scalable application is still challenging. One of the major complications for deployment has been the distribution shift observed in the physiological data obtained through sensors. This study develops a deep adversarial domain adaptation framework to adapt to out-of-distribution data(ODD) in the wearable physiological device based on photoplethysmography (PPG). The domain adaptation framework is developed and validated with reference to the heart rate predictor based on PPG. A heart rate predictor module comprising feature generating encoder and predictor isinitially trained with data from a given training domain. An unsupervised adversarial domain adaptation method is then implemented for the test domain. In the domain adaptation process, the encoder network is adapted to generate domain invariant features for the test domain using discriminator-based adversarial optimization. The results demonstrate that this approach can effectively accomplish domain adaptation, as evidenced by a 27.68% reduction in heart rate prediction error for the test domain. The proposed framework offers potential for scaled adaptation in the jobsite by addressing the ODD problem. 

Exoskeletons, also known as wearable robots, are being studied as a potential solution to reduce the risk of work-related musculoskeletal disorders (WMSDs) in construction. The exoskeletons can help enhance workers’ postures and provide lift support, reducing the muscular demands on workers while executing construction tasks. Despite the potential of exoskeletons inreducing the risk of WMSDs, there is a lack of understanding about the potential effects ofexoskeletons on workers’ psychological states. This lack of knowledge raises concerns thatexoskeletons may lead to psychological risks, such as cognitive overload, among workers. Tobridge this gap, this study aims to assess the impact of back-support exoskeletons (BSE) onworkers’ cognitive load during material lifting tasks. To accomplish this, a physiologically basedcognitive load assessment framework was developed. This framework used wearable biosensorsto capture the physiological signals of workers and applied Autoencoder and Ensemble Learningtechniques to train a machine learning classifier based on the signals to estimate cognitive loadlevels of workers while wearing the exoskeleton. Results showed that using BSE increasedworkers’ cognitive load by 33% compared to not using it during material handling tasks. Thefindings can aid in the design and implementation of exoskeletons in the construction industry. 

Work-related musculoskeletal disorders (WMSDs) are a leading cause of injury for workers who are performing physically demanding and repetitive construction tasks. With recent advances in robotics, wearable robots are introduced into the construction industry to mitigate the risk of WMSDs by correcting the workers’ postures and reducing the load exerted on their body joints. While wearable robots promise to reduce the muscular and physical demands on workers to perform tasks, there is a lack of understanding of the impact of wearable robots on worker ergonomics. This lack of understanding may lead to new ergonomic injuries for worker swearing exoskeletons. To bridge this gap, this study aims to assess the workers’ ergonomic risk when using a wearable robot (back-support exoskeleton) in one of the most common construction tasks, material handling. In this research, a vision-based pose estimation algorithm was developed to estimate the pose of the worker while wearing a back-support exoskeleton. As per the estimated pose, joint angles between connected body parts were calculated. Then, the worker’s ergonomic risk was assessed from the calculated angles based on the Rapid Entire Body Assessment (REBA) method. Results showed that using the back-support exoskeleton reduced workers’ ergonomic risk by 31.7% by correcting awkward postures of the trunk and knee during material handling tasks, compared to not using the back-support exoskeleton. The results are expected to facilitate the implementation of wearable robots in the construction industry. 

Recent advances in robotics have enabled robots to collaborate with workers in shared, fenceless workplaces in construction and civil engineering, which can improve productivity and address labor shortages. However, this collaboration may lead to collisions between workers and robots. Targeting safe collaboration, this study proposes an intention‐aware motion planning method for robots to avoid collisions. This method involves two novel deep networks that allow robots to anticipate the motions of workers based on inferences about workers' motion intentions. Then, a probabilistic collision‐checking mechanism is developed that enables robots to estimate the collision probability with the motions of workers and generate collision‐free adjustments. The results verify that the method enables robots to predict workers' intended motions 1 s in advance and generate adjustments with a collision probability of less than 5.0% during collaborative masonry tasks. This study facilitates the safe implementation of collaborative robots in construction and civil engineering. 

The information privacy of the Internet users has become a major societal concern. The rapid growth of online services increases the risk of unauthorized access to Personally Identifiable Information (PII) of at-risk populations, who are unaware of their PII exposure. To proactively identify online at-risk populations and increase their privacy awareness, it is crucial to conduct a holistic privacy risk assessment across the internet. Current privacy risk assessment studies are limited to a single platform within either the surface web or the dark web. A comprehensive privacy risk assessment requires matching exposed PII on heterogeneous online platforms across the surface web and the dark web. However, due to the incompleteness and inaccuracy of PII records in each platform, linking the exposed PII to users is a non-trivial task. While Entity Resolution (ER) techniques can be used to facilitate this task, they often require ad-hoc, manual rule development and feature engineering. Recently, Deep Learning (DL)-based ER has outperformed manual entity matching rules by automatically extracting prominent features from incomplete or inaccurate records. In this study, we enhance the existing privacy risk assessment with a DL-based ER method, namely Multi-Context Attention (MCA), to comprehensively evaluate individuals’ PII exposure across the different online platforms in the dark web and surface web. Evaluation against benchmark ER models indicates the efficacy of MCA. Using MCA on a random sample of data breach victims in the dark web, we are able to identify 4.3% of the victims on the surface web platforms and calculate their privacy risk scores.