Abstract: Haptic-robot control based on local autonomy and a dual-proxy model is provided. The dual proxy guarantees generation of safe and consistent commands for two local controllers, which ensure the compliance and stability of the systems on both sides. A Force-Space Particle Filter enables an autonomous modeling and rendering of the task contact geometry from the robot state and sensory data. The method suppresses the instability issues caused by the transfer of power variables through a network with communication delays in conventional haptic-robot controllers. The results demonstrated the transparency and high fidelity of the method, and robustness to communication delays. While the conventional method failed for communication delays higher that 150 milliseconds, the dual proxy method maintained high performance for delays up to one and a half seconds. The local autonomy-based haptic control of robots with the dual-proxy model enables applications in areas such as medical, underwater and space robotics.