Smart & Autonomous Built Environments
Understanding Sensory Information Requirements for Teleoperation in Construction
The construction industry has recently experienced an increased deployment of automation and robotics on- and off-site. Despite technological developments, however, the use of fully autonomous machines has been restricted to a limited number of applications, with remotely-operated and, more recently, teleoperated machines representing the majority of robots found in construction sites.
For teleoperated robots, specifically, although important results have been borrowed from more technologically-advanced industries such as aerospace and military, in most cases, the human-machine interfaces are rather primitive, with control systems relying solely on levers that are used to control the equipment's joints and end-effectors, and feedback systems based on 2D images on flat screens and audio. Given the problems that arise from decoupling the worker from the physical environment where the construction task takes place in the teleoperation context, it is key to understand what the sensory information requirements are for different classes of construction tasks at various levels of task complexity to design human-machine interfaces that reduce the operator's cognitive loads while increasing safety, health, and potentially productivity levels. In this project, we aim at understanding the operator's sensory information requirements for a group of construction tasks at varying levels of task complexity in order to propose teleoperation workstations that optimize the operator's performance during teleoperation.
Status: On-going
Eustress vs. Distress: Automated Stress Detection in Offices
Stress is recognized by the World Health Organization as the “epidemic of the 21st century”. While there are many reasons for stress, job pressure is the main cause of people’s stress. However, depending on how the workers perceive job stressors, their stress experience can be positive (eustress) or negative (distress).
Distress is what most people refer to when they feel “stressed out”. It usually results in people feeling overwhelmed when the cause of stress is not within their control. On the other hand, eustress motivates individuals to reach their goals, face challenges and achieve success and fulfillment. Differentiating between a positive and negative appraisal of stress is important, as eustress may be one of the most powerful resources to prevent or reduce distress at work and as such lead to a productive and energizing work environment. While extensively studied, stress detection research widely considers all stress-related data as distress. Therefore, the aim of this project attempts to address this significant gap in real-world understanding of distress versus eustress, by creating automated eustress vs distress detection framework using machine learning techniques. The framework will examine facial expressions, physiological signals, and human-computer interactions as potential predictors of eustress vs distress during office work.
Status: On-going
Reskilling & Upskilling for Human-Robot Interactions on Construction Sites
The construction industry is increasingly adapting robots for automating various construction tasks. Unlike other industries, where multiple robots complete various tasks independently, worker–robot collaboration is necessary for construction activities due to the specific project and site requirements, such as the need for multiple parallel or sequential activities, exposure to outdoor conditions, and dangerous working conditions.
his project aims to contribute to the fundamental understanding of workforce training impacts on workers’ safety behavior, operational skills, trust-in-automation, robot operation self-efficacy, situational awareness, and mental workload. We focus on understanding the impact of VR-based training as a whole package and how its components moderate the development of human-related factors in Human-Robot Interaction (HRI). Through this effort, compared with real-life situations, we will systematically collect data on the affordances and hindrances of existing automation and common problems that arise at the individual- and construction site levels. In addition, the environment developed as part of this project will serve as a platform for the research community to explore important research questions related to human-machine collaboration and its impact on construction work processes.
Status: On-going
Smart IoT Desk
Office workers spend most of their working time at their desks, where they are engaged in sedentary behavior, and usually subjected to poor indoor thermal and lighting conditions. Prolonged sitting, and long term exposure to homogeneous temperatures have been linked with reduced metabolism, increased obesity, and increased risk of cardiovascular disease and type two diabetes.
Furthermore, uncomfortable thermal environment and improper lighting conditions have a negative influence on worker productivity and well being. We envision smart desks in the near future to provide a solution to adverse health and productivity impacts that current office work entails. The smart desk uses a wide range of sensors to monitor the environment around the user, as well as the user behavior. The desk uses reinforcement learning to learn and adapt to their thermal, lighting and posture preferences. The desk will control the local thermal and lighting environment around the user based on user preferences while trying to promote health and productivity based on the best practices identified in literature. The goal is to engage users in a bi-directional interaction where the desk promotes exposure to wider range of thermal conditions, and healthier use of sit-stand regimen to improve their productivity and reduce adverse health impacts from the indoor environment.
Status: Completed
Worker-Robot Collaboration on Construction Sites
The construction industry is increasingly adapting robots for automating various construction tasks. Unlike other industries, where multiple robots complete various tasks independently, worker–robot collaboration is necessary in construction activities due to specific project and site requirements, such as the need for multiple parallel or sequential activities, exposure to outdoor conditions and dangerous working conditions .
In this project, we aim to contribute to the fundamental understanding of trust-in-automation, specifically understanding how construction workers develop trust-in-automation. We focus on understanding the role of individual, as well as task differences, and how they moderate the development of trust-in-automation. Through this effort, compared with real-life situations, we will systematically collect data on the affordances and hindrances of existing automation and common problems that arise at the individual- and construction site-level. In addition, the environment, developed as part of this project, will serve as a platform for the research community to explore important research questions as they relate to human-machine collaboration, as well as human-machine collaboration’s impact on construction work processes.
Status: Completed