Robot - Pattern computer

Intellectual Computer Technology for the Control of the Mobile Robot Purposeful Actions

The proposed technology is intended for the intellectual control system of the indoor class autonomous mobile robot (MR) and for the man-machine interface with the operator-supervisor organization. Mathematical modeling and universal representation structures of the different MR physical modules operational processes (odometry navigation system, location sensors, video cameras) are based on a 3D geometric approach. Appropriate software of the modeling and control environment for robot is developed in the form of Windows applications. The software environment provides an intellectual interface for the operator-supervisor and is based on the client-server technology.

Autonomous modeling mode – for control and simulation purely in the virtual environment;
Remote control mode – for controlling real robot via radio channel;
Supervisor control mode – for MR autonomous functioning within the scope of the existing library typical tasks. This mode uses the purposeful actions self-programming algorithms and integrates informational streams of real and model data concerning the robot state.

Mobile indoor robot (Laptop on the wheals)
Notebook
Web-camera
Sensory CAN-network
Operator-supervisor PC equipped with an audio system
Windows 2000/XP operating system
Adapters for setting up a wireless communication channel (Wi-Fi 802.11) between stationary and on-board computers
Modeling and control system software with intellectual operator interface (Client)
On-board control system software (Server)
Preliminary processing and interpretation of the range finder data software which is also intended for the measurement results insonation

Visually estimate robot position in the remote control session using the most informative perspectives;
State tasks for the robot;
Check the fundamental feasibility of these tasks implementation in the modeling environment;
Route trajectories to the target points;
Opt among possible approaches to the target objects;
Generate program trajectories;
Accumulate both model and real data and knowledge about the robots and their operational environment in the database;
Estimate visually the algorithms quality of the location sensor data model representation;
Communicate both with the model and the real robot by means of the high-level abstraction naturally perceivable images, namely virtual spatial scenes and natural language commands and messages.

A sensor network prototype which is based on the standard of CAN (Controller Area Network) ISO 11898 and a station based on C8051F005 microprocessor for the preliminary processing of the signal obtained from Sharp GP2Y0A02YK range finder in real time are developed. Insonation of the information about the distance to objects in the range of 25-150 cm is implemented.