Design of a general interface of a driving simulator cockpit

Abstract: This thesis has covered the design, implementation and evaluation of a new hardware interface for an automobile simulator cockpit - the front end of a SAAB 9-3. Two circuit boards were produced, one for analog and one for digital signals. The boards were manufactured at ITN’s PCB lab at Campus Norrköping. Both boards were equipped with terminal blocks, enabling easy connection of input signals. The analog signals were of such configuration that voltage division was to be used. I therefore had 32 potentiometers of different resistance intervals soldered to the analog board. These potentiometers, along with the ones in the cockpit (mounted by the manufacturer), provided fixed analog output levels. The digital board did not have any signal processing but for every signal a LED was soldered to the board. This was done to make trouble-shooting easier. As an interface to the simulator itself, two I/O boards and two PCI-cards with both analog and digital inputs were used. The I/O boards were situated close to the analog and digital boards to which they were connected via planar cable. Two SCSI cables then transmitted the information to the PCI-cards, 5 m away. Installation of the boards went smoothly. The interface then underwent and passed a test to everyone’s satisfaction. Furthermore, Controller Area Network (CAN) has been examined in order to determine its implementability in simulator applications. First, the protocol and specification of CAN was examined. The next step was to investigate its general implementation in automobiles. This proved to be a difficult task since there is no general way of implementing it which is used by every car manufacturer. I was therefore forced to make individual contact with manufacturers and it became apparent that such information was not something they were keen on sharing. These contacts were still of use because none of them recommended CAN to be implemented in a simulator. The reasons mentioned for this was that because of its asynchronous protocol, CAN is not suited for message transmission which must be sent instantly. When that was the case, an implementation of CAN would mean that other protocols would have to be used as well in order to transmit all signals. Because of that, my conclusion was that CAN is not a feasible solution in simulator applications.