This thesis presents a new method of simulating the sensation of driving. There are many different driving simulators currently on the market. All of which use the force of gravity to simulate the sensation of driving, which limits their capabilities (especially in regard to simulating long-term forces). Our method works on the principle of generating forces to the head of the user, using actuators.
The first part gives a quick overview of the history and development of motion simulators. There have been numerous developments throughout the last decades, but the major issues persist. The new method of simulating the sensation of driving faces its own challenges.
The new method of simulating the sensation of driving generates the same forces on the head of the user as he would experience driving an actual vehicle. The simulator, as the thesis, is composed of several units that work together as a whole. The first unit outlines the possible actuators for generating force. The choice of the actuator has a major impact on the choice of the open- or closed-loop controller. The simulation shows that, in the case we choose a DC motor, best results are obtained if controlling the force by measuring it directly. For this purpose we will have a look at the method for measuring the forces generated between the guides and the user. To generate any kind of force we need a power supply. The chosen DC motor requires a high power supply, which has an immense impact on the cost. Therefore we have decided to use two computer power supply units, which together produce a voltage of 24V, to control our actuator. The last and most important part is the software which consists of two programs.
The main program is entwined with the simulation game; its purpose is to send information from the simulation to the controller Netduino within the shortest time possible. The program on Netduino receives the data and, by means of measuring the force between the user and the guide, controls it appropriately so that the error between the measured and the obtained force is minimized.
The main program is the first bottleneck of the time complexity, it has to ensure data transfer every few milliseconds. This can be achieved by using the principle of memory-mapped files. The simulation game gets free access to an allotted space on the RAM. This gives us the advantage of having the file we are accessing handy, the size of which is many times smaller than the whole memory of the simulation game.
The second bottleneck is the program on Netduino, which is responsible for measuring all the forces acting between the user and guides, and for controlling the loop. Because of the noise of the analogue-digital converter, the accuracy of the measurement correlates to its speed. The speed of the loop control does not pose a problem, therefore we focused our attention on improving the measuring speed. Because of the noise, we had to use the average of hundred measurements. At the beginning of each loop we replaced the oldest result with the new one. With this algorithm we were able to achieve the desired speed as well as accuracy.
The comparison of our driving simulator to the classical type, contained in the last part of the thesis, shows that the former has many advantages, not the least of which are no maximum user weight and the capability of generating long term forces of up to 6G. In the course of the development we constructed a prototype to help us thoroughly explore the idea. Test users provided us with a lot of positive and constructive criticism to aid us on the path of further refinement and development of the product.
|
