This paper presents an experimental robotic setup dedicated to human arm endpoint impedance measurement in a cyclic task with physical interactions. Understanding human endpoint impedance adaptation during a cyclic task in interaction with the environment can provide insights for the design of innovative robot controllers for collaborative robotics both in manufacturing and service contexts. The ball bouncing benchmark task, classical in human movement science, is selected here for its simplicity, yet representativeness of handeye synchronization and intermittent physical interactions. For a better control over the environmental conditions, the task is performed in a simulated environment, where the paddle is actuated by the human participant by interacting with an admittance-controlled collaborative robotic arm. First experimental feasibility and identification results are presented.