The system uses a multi-computer architecture to optimize the allocation of resources.

All modules, except the 3D model and video display, connect with another module (Java server) for communicating with the web server (host server). This procedure prevents the use of multiple socket connections by concentrating all the traffic in one channel.

The communication between the several modules inside the HTML page is done using the class AppletContext. To use this class all the modules must be inside the same frame. The exception is the 3D model which uses its own socket connection, and thus does not need the communication module. The display is broken into 3 frames.

The 3D textured model is coded in VRML 2.0 and has been optimized for visualization with Comsmo Player 2.0 as a plug-in for the brower, and is contained in frame 3.

Frame 2 uses both the GPDD and Internet video server.

The images acquired from the robot's video camera are transmitted through a wireless video connection to a third computer. These images are encoded and pushed to the Internet using Microsoft Netshow with ReadVideo.

Commands betweent the host computer and mobile robots computer are transmitted via a wireless RS232 connection. The commands are intercepted by the communications server running on the robot's computer and redirected to the proper real-time task.

Background tasks both in the robot and the host computer monitor the status of the communications between the robot and the host.