> > To put it a bit provoking (no personal offense intended!) for the sake > > of getting the status right, this looks to me that we're never going to > > have this feature of proper tire/ground reaction simulation in > > FlightGear simply because the involved folks prefer to play around with > > smaller, 'approximative' workarounds forever instead of biting the > > bullet just once. Right ? > > > > Cheers, > > Martin. > > > Not quite. I'll write more this evening. Off to work, now. > > JB
OK. The answer to your question (which is really more like an assertion disguised as a question ;^) follows below. There are many kinds of simulations, that all address particular needs. Are we talking about a training simulator, or an engineering simulator, or an entertainment simulator, or ... ? In my day job I work on a sophisticated launch vehicle simulator. For our specific uses, we are not overly concerned about the exact dynamics of the vehicle during its flight, but we are concerned that the launch vehicle arrives at a particular point in time and space with a particular state, so we can test the abort systems. The level of fidelity for the various subsystem models should be driven by the purpose or intended use of the simulator. In the case of ground reactions for JSBSim, from the start we noticed that it is difficult to model ground reactions when at rest or at small velocity. The wheel slip angle, for instance, varies wildly in JSBSim when no filtering or other compensation is used: http://jsbsim.sourceforge.net/NoFiltering.png If the output is filtered at the frame rate, the slip angle doesn't oscillate quite so fast, but it's still bad: http://jsbsim.sourceforge.net/FilteringOn.png If the forces are ramped down to zero below some threshold, the dynamics improve greatly: http://jsbsim.sourceforge.net/FilterAndRamp.png Finally, adjusting the longitudinal and lateral ramp thresholds can improve response even more. Note the scale on the left side of the plot: http://jsbsim.sourceforge.net/FilterAndRamp2.png So, I have to ask you: what is "proper tire/ground reaction"? With a few lines of code I've reduced the ground reactions artifacts for the f-16 to very small levels - and I really haven't optimized them, yet. They could maybe be improved even more. So what is the cost/benefit to adding more code to do gear modeling more precisely? How much would the user really notice the difference? If we model the ground reactions to such high fidelity, should we then also model the aero characteristics using the Navier Stokes equations? There are always tradeoffs in simulation. In the case of JSBSim, I am hoping to model ground reactions so that the end user is satisfied with the response. There's not a simple answer to this question. There is definitely a need for more documentation on our gear modeling, and perhaps some software tools for people to use when creating the landing gear portion of a JSBSim configuration file. Jon ------------------------------------------------------------------------------ Create and Deploy Rich Internet Apps outside the browser with Adobe(R)AIR(TM) software. With Adobe AIR, Ajax developers can use existing skills and code to build responsive, highly engaging applications that combine the power of local resources and data with the reach of the web. Download the Adobe AIR SDK and Ajax docs to start building applications today-http://p.sf.net/sfu/adobe-com _______________________________________________ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
