Extended range telepresence allows a human user to intuitively teleoperate a mobile robot through arbitrarily large remote environments by natural walking. In order to give the user the possibility to navigate the robot through an arbitrarily large remote environments, while his own environment is of limited size, Motion Compression is used. The Motion Compression framework provides a nonlinear transformation between the user's path and the robot's path, which preserves path length and turning angles. There is, however, a difference in path curvature, which is minimized in order to guarantee a high degree of immersion. A major drawback of the current system is its inability to deal with non-convex time-variant environments or environments shared by multiple users. This paper presents a systematic approach to extending Motion Compression to non-convex environments. This solution will then be used to cover the multi-user case.