Biological inspired templates can help to control complex robotic movements. In this thesis a control strategy, enabling hopping motions of a segmented robotic leg, is developed. The control bases on the spring loaded inverse pendulum (SLIP) model, which can describe the courses of displacement of the center of mass and the ground reaction force during human or animal hopping motions. To use this template as a calculation model for desired control values a method called virtual model control (VMC) is used. VMC implements virtual components in real structures to design a desired behavior. Existing actuators of the real system are controlled in a manner to mimic the effects, the virtual components would have on the system. The virtual component used in this work is a spring with certain properties. Like in the role model, the SLIP template, the spring is virtually attached between hip and foot of the robotic leg. The knee is the only actuated part of the structure. Through the control of the knee torque the effects of the virtual spring are mimicked. The used test-bed necessitates it to adjust the developed control laws for the compensation of losses. Different methods for the calculation of a variable virtual spring stiffness have been developed, resulting in stable hopping motions of the robotic leg in the used test-bed. The resulting control strategy does not need a feedback loop of the controlled parameter and is therefore a kind of a feed-forward approach. The possibility of an overlaid force-feedback control has been examined and the limits of this method have been estimated.

Realization and optimization of robotic hopping motions using bio-inspired virtual model control