The gait pattern of prosthesis wearers differs from that of able-bodied humans. For example, higher energy consumption, asymmetry and increased loading are observed. For a targeted improvement of existing prostheses, the dynamic interplay between a human and the prosthesis needs to be well comprehended. An appropriate method to analyze the dynamics is Real-Time Hybrid Substructuring (RTHS), where the prosthesis is tested on a test bench and the amputee is co-simulated in real-time. In this thesis, the foundation is laid for testing prosthetic feet using RTHS. For this purpose, control schemes are first compared that allow safe and accurate testing of systems with contact. Specifically, Iterative Learning Control as a feedforward controller, in combination with Normalized Passivity Control, is proposed and experimentally investigated. A key research question in RTHS is the fidelity assessment of the experiments. This thesis proposes a novel approach called Fidelity Assessment based on Convergence and Extrapolation (FACE). The presentation and analysis of the method is done using virtual RTHS tests, as well as experimental RTHS tests. Additionally, an RTHS test was conducted, where the human is modeled using the Virtual Pivot Point model and the prosthesis tested on the test bench. In this test, one gait cycle is completed and the results reveal that the dynamic interplay between an amputee and a prosthesis can be emulated. By this work, not only the feasibility and potential to test prostheses are demonstrated, but also methods are presented that represent an advance for various RTHS applications in the field of actuator control and fidelity assessment.weiterlesen