A Customized Quadruped Control and Sensor Integration Framework with a Construction Worker Activity Recognition Module

Quadrupeds are the most prevalent type of robots in construction sites. These robots can navigate uneven terrain found in construction sites, making them suitable vehicles for sensors and for monitoring tasks. However, a streamlined client-side software would boost rapid deployment of application-specific models for these robots. Considering the substantial prerequisite knowledge required to adapt the robots to specific construction applications, we present a comprehensive framework to address these gaps in accessibility, simplifying the customization process. This framework provides a template that facilitates seamless communication between the robot, edge devices, sensors, and a Unity simulation for mission planning in construction. To demonstrate this framework’s flexibility, we exhibit a case study using this template in the construction domain that performs worker activity recognition and features a novel self-labeling mechanism for construction activity data. The findings highlight the potential of software tools in expanding the utility of quadrupeds across various engineering domains.