Slashdot Mirror
F1 Simulators Revealed
An anonymous reader writes "Racecar Engineering has posted an exclusive look inside the simulator of a leading grand prix team. Particularly interesting is that the Formula 1 team uses software based on the free simulator Racer (with source code available) albeit with a custom vehicle model and hardware interface via CAN-bus. The article highlights the importance that mainstream racing sims (rFactor, iRacing) have in simulation at the pinnacle of the worlds most advanced sport."
Along similar lines, reader PatPending writes "Engineers at the Max Planck Institute for Biological Cybernetics in Germany (surely the greatest of all institutes) have turned a massive robot arm into a Ferrari F1 simulator, discovering a new strain of awesome in the process. The contraption, known as the CyberMotion Simulator, consists of an industrial robotic arm fitted with a racing seat, a force feedback steering wheel and a 3D simulation of the Monza Formula 1 track beamed from a projector on to a curved display."
Most of the industry is using rF Pro. Despite the constant negativity around the tyre/aero maps with the baseline version of rFactor, I've always been of the opinion that rF is the most scalable sim around. Nice to see RaceCarEngineering get a plug, also. Those guys do good work.
The CyberMotion Simulator isn't the first to use an industrial robot as the motion platform for a game. When I visited Legoland (Billund) in 2004, they had several robots set up as a thrillride, with the robot going through a user-programmable motion pattern.
I'm not sure that is true. At the incident at 00:51, the arm is moving to the left of picture, and then suddenly starts moving to the right. It is the acceleration that counts, not the position or speed. The sudden acceleration from moving left to moving right appears to happen right on the moment the driver turns the wheel; the fact that it takes the arm some time to move to the right of picture is irrelevant.
On the other side the german economy has the biggest growth in europe with over 2%, looks we do get something right.
In another iteration of the company I work for (we've had a few mergers), one of the divisions ran a centrifuge for human factors research on pilots. Then someone had the bright idea to turn the technology, and software into an amusement park ride.
It had an enclosed gondola with six axis movement and a display inside to show the environment that was being simulated. The arm spun at a constant rate, and with the gondola at a certain angle it could trick the inner ear to think you were sitting still while you were turning. And then by changing the the angle of the gondola in relation to the centrifugal force vector it could give the sensation of roll, pitch and yaw.
Although the tech was cool, and some parks showed some interest. It never went anywhere because they couldn't figure out a way to get the throughput that the park operators where looking for.
Sounds like that system would be the best of both methods mentioned in the summary.
I want to shoot the messenger!
Forget the simulator part! Give me an office chair strapped to the robot arm...Imagine delivering those TPS reports without getting up!
David Coultard was about to race at Monaco, but he had never raced there before, so he fired up the Microprose F1 Grand Prix to get used to the course, and won it too!
I'm not. Not at all.
The high reliance on simulators is not necessarily because it is in any way better than physical testing. The FIA now severely limit the amount of physical testing that can be done.
It's now regular for a team to receive updated parts mid week straight from the factory and the first real-world testing is the Friday practice session, the day before qualifying. This Friday is effectively the only testing day, since the car you complete your time in during qualifying is literally put in a bag and only opened shortly before the race. This is also why drivers who for whatever reason have no chance of gaining anything from finishing a race do so anyway; they use it as free testing time.
What brings a package like Racer to a level usable by F1 teams is not so much the software itself (even though the openness of it helps), but the data that it is being fed by the team and their suppliers (e.g. performance and feedback data from the car, professional track scans, etc.).
Since the casual user does NOT have access to these data set, all they're left with is the "empty sheet of paper", on which they can paint their own fantasies, but, just because they're using the same "paper" as a race team, this does not mean that the outcome (the simulation feel) will be the same as that of an F1 team.
So - unless you have some realistic data to plug into it (and to test in real-life feedback loops), don't be under any illusion that it's any better than any other racing sim.
Commercially available automatic gearing degrades performance with 25% (for an average driver, much more for a good one). There is no reason a professional drivers would volunteer to waste that much performance just so they can rest one foot while driving. The automatics have only reentered professional racing due to new more complex engine and gear types, that are harder for a human but easier for a computer to control.