[en] In this article, we address the problem of controlling robots with arbitrarily-switched constraints and unknown dynamics. Switching between different constraints of a robot would result in a switched nonlinear system that does not inherit the behavior of its individual subsystems. In order to guarantee stable performance of robots with arbitrarily switched constraints and unknown dynamics, we propose a Robust Adaptive Fuzzy Control (RAFC) strategy that can guarantee global stable performance under such challenging conditions. The suggested control strategy relies on the synergy of the Sliding Mode Control (SMC) that adds robustness against possible dynamics parameters drift, finding a Common Lyapunov Function (CLF) that guarantees stability under arbitrary constraints switching, and Direct Adaptive Fuzzy System (DAFS) that relaxes the need for knowing the precise robot dynamics. Experiments are performed on a KUKA Lightweight Robot (LWR) doing camshaft caps assembly of an automotive powertrain. The given robotic assembly process falls in the category of switched constrained robots and the efficiency of the suggested RAFC strategy in controlling such a robotic task will be shown.