Space Robot Emulator

My work in this project focused on the design and implementation of the electric/ electronic subsystem, as well as the software development of the second version of the space simulator's robot. This planar simulator has been placed in the Control Systems Lab of the Mechanical Engineering Department of NTUA, and it consists of robots hovering in a distance of 10 [μm] over a granite table of negligible roughness. The purpose of the simulator is the experimental study of the robots’ dynamics in conditions that simulate zero gravity in 2D, through the elimination of friction. The novelty lies in the fact that the robots are of low mass, low cost, completely autonomous and carry subsystems that resemble these of a real space robot. The motivation for the development of such a system is the increasing importance of space robots in cases such as: space exploration, construction, maintenance and inspection of systems in space, approach and attachment to other bodies in orbit.

The robots of the simulator consist of 3 subsystems: a) mechanical: they are made of aluminum, and they bear two arms, with two joints each, and a reaction wheel as an alternative way of motion, b) pneumatic: they use CO2 gas in order to hover using 3 air-bearings, and for planar motion using 3 pairs of thrusters, c) electric/ electronic: they estimate their posture using two systems of sensors (optical sensors on their body and an external camera) and through automatic control act appropriately upon the environment.

Employing the technique of model-based design, the development of the new robot was carried out as a whole, since the design of algorithms, the software development and the final integration on the hardware were studied as interdependent parts. The software used was xPC Target, from Mathworks. In this environment, C code is quickly generated from simple Simulink models, and it is finally executed on the robot’s hardware in hard real time. In this manner, validation and verification of the design were continuously performed throughout the development. As a result, emphasis was put on innovation, while at the same time potential low-level problems in hardware and software were surmounted.

In order to attain computation autonomy a PC-104 system was used, including a CPU board, a 48 Digital I/O card and a power supply board. Power autonomy was achieved using Li-Po batteries. The localization of the robot was achieved through two different systems of sensors: (a) relative estimation sensors (three optical sensors), for fast estimation and (b) an absolute estimation sensor, namely an off-board camera, providing more accurate estimation. Two wireless Ethernet bridges were used a) for the remote control of the robot, and b) to receive motion data from the camera system in real time. Four PCBs were

designed and printed for power routing, thrusters’ control, collecting optical sensors' data and another one to be used as a control panel. Finally, several experiments were carried out with the robot in full format, in order to evaluate the processing algorithms that were developed, on data derived from the camera and the optical sensors systems.

Please visit our lab site for more info.