A Framework to Identify Task-Phase and Attentional-Condition for Supporting Complicated Dual-Arm Operations

Abstract

The state identification framework we propose supports complex construction machinery operations using a dual arm. Such support requires compatibility with different types of support and commonality among various operator skill levels. Our framework is organized into (i) real-time task phase identification defined using joint load applied based on environment constraints and (ii) time-series attentional condition identification defined as an internal work-state condition classified by the operational support necessity level and dependent on the vectorial or time-series data selected by the identified task phase. Experiments are conducted using the instrumented hydraulic dual arm system for transport and removal tasks, including complex dual-arm operations. Results show that the number of erroneous contacts, internal force applied, and mental workload decreased without any increase in time, confirming that operational support based on our framework greatly improves individual operator work performance.