A hierarchical machine learning framework for the identification of automated construction

A robust monitoring system is essential for ensuring safety and reliability in automated construction. Activity recognition one of the critical tasks monitoring. Existing studies this area have not fully exploited potential enhancing performance machine learning algorithms using domain knowledge, especially problem formulation. This paper presents a hierarchical framework improving accuracy identification Automated Construction System (ACS) operations. The proposed arranges operations to be identified form hierarchy uses multiple classifiers that are organized hierarchically separating operation classes. It tested on laboratory prototype an ACS, which follows top-down construction method. ACS consists set lightweight portable machinery designed automate structural frame low-rise buildings . Accelerometers were deployed at locations structure. acceleration data collected while operating equipment used identify through techniques. compared with conventional approach involves flat list classes separated. was comparable top level. However, outperformed when fine levels identified. versatility noise tolerance also reported. Results demonstrate robust, it feasible precisely. Although validated full-scale effects strong ambient disturbances actual sites been evaluated. study will support development assist main operator ensure safe high-level details purpose can utilised project assessment progress application outlined.