As far as complex contact-based manufacturing tasks are concerned, humans outperform machines. Indeed, conventionally controlled robotic manipulators are limited to basic applications in close to ideal circumstances. However, tedious work in hazardous environments, make some tasks unsuitable for humans. Therefore, the interest in expanding the application-areas of robots arose. This paper employs a bottom-up approach to develop robust and adaptive learning algorithms for trajectory tracking: position and torque control in the presence of uncertainties and switching constraints. The robotic manipulators mimicking the human behavior based on bio-inspired algorithms, take advantage of their know-how. Simulations and experiments validate the concept-performance.