The pressing issue of electronic waste (e-waste) presents a significant environmental and economic challenge. Automating e-waste disassembly is crucial for resource recovery, but achieving stable robotic contact is difficult due to highly variable material stiffnesses. To address this, this paper proposes a robust control framework with three primary contributions. First, an accurate dynamic model is derived via frequency-domain system identification, enabling analytical gain optimization within theoretical stability limits. Second, we present a rigorous comparative analysis of Admittance-based Direct Force Control (ADFC) and Hybrid Force/Velocity Control (HFVC), demonstrating a quantified 59% reduction in impact forces. Third, we introduce a binary Gain Scheduling strategy driven by Adaptive Stiffness Estimation, which identifies effective stiffness in real-time ( < 10 ms) to adjust gains according to soft or hard material properties. Validated through simulations and real-world experiments on a STÄUBLI TX2-140 industrial robot, this approach effectively bridges the sim-to-real gap, eliminating the model mismatch typical of fixed-gain controllers without requiring continuous re-tuning, ensuring reliable contact across diverse industrial materials. • Accurate manipulator model derived via frequency identification for gain optimization • Adaptive Stiffness Estimation algorithm identifies material properties in < 10 ms. • Binary gain scheduling at 110 kN/m stiffness threshold bridges the sim-to-real gap. • Compared direct force vs. hybrid force/velocity control for contact effectiveness. • Contact impact forces reduced by 59% compared to classic robotic controllers.
Granados et al. (Thu,) studied this question.