Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[PytonPago]

What I meant is I cannot call the position and velocity of a ground vehicle from an aircraft script, so I cannot track it. So far only aircraft to aircraft. I haven't experimented with ground vehicles tracking the position and velocity of another ground vehicle, but I expect it would work. Supposedly Bullet collisions interact with JSBSim objects, but not vice versa.

I think both objects are now as separate things in OT ... so, interactions are quite the problem yet. Hitting the target could be taken from the projectile - just finding out, if its position along the way will be in some radius from the target (at-least for now). For more complex hit-simulation, many things must be implemented and targets having some special meshes (for fuel-tanks, engine, etc.) where bullets should be traversing some of theyr faces and that to be taken as hits.

[Uriah]

Quite right, a hit or miss can be calculated. I believe JSBSim can output the flight data into CSV format, I'll run a test and see how it works with the Jupiter 246 rocket.

Gathering all the data for each FDM is very time consuming. I spent six hours and feel confident I have everything I need for the BGM-109B, C, H and G varients, that is submarine (B), ship (C), air (H) and ground (G) launched Tomahawks respectively. I have been studying guidance and control extensively, and the Tomahawk really does it all, a rocket, an airplane, and a guided bomb all in one.

Here is what I am working on for the Tomahawk, but much of it applies to all missile guidance systems.

1) PRE-LAUNCH PHASE
Target coordinates (fixed ground target defined in .js script variables LAT/LONG/AGL) are used to calculate a possible launch solution and configures/initiates the guidance program for control and navigation to the target. The Tomahawk is used primarily against slow moving or fixed targets, such as ships or bunkers... so I do not need to predict the future position of high velocity targets. This makes things less complex for now, and we will assume the targets are fixed. We will also assume the seeker is GPS/INS based, I cannot simulate DSMAC.

2) LAUNCH PHASE
The seeker, sensor and launch guidance systems are initiated and the Solid Rocket Motor is fired upon command initiation.

3) BOOST PHASE
During the boost phase, the guidance rolls to azimuth, holds yaw, and performs the pitch turn to horizontal flight. The solid booster provides 6000 lbf (pound force, or 26 kN) thrust for 12 seconds reaching maximum dynamic pressure.

4) TRANSITION TO CRUISE PHASE
Above a specified velocity the engine fairing is jettisoned and the wings, fins and intake are deployed for aerodynamic stability and oxidizer flow. Once the solid booster burns out it is jettisoned and the turbofan engine is started. The guidance makes heading and course corrections and holds roll and pitch.

5) CRUISE (MID-COURSE) PHASE
Depending on the mission profile, during the cruise phase the missile assumes the desired altitude and navigates to the target area. A high altitude flight path will have the greatest range, while a low altitude terrain follow mode would be more stealthy on approach beneath enemy radar, mere meters above the ground. This will be configured prior to launch in the script file .js variables to select the mission profile. In terrain follow mode, the missile will maintain an altitude (AGL) by calculating the average AGL per given amount of time (smooth ascent/descent, maybe Kalman filtered control output).

6) TERMINAL PHASE
Once within the target area the guidance makes final course adjustments and the seeker guides it onto target.

7) TERMINAL MANEUVER PHASE
A pop-up maneuver can be performed prior to impact, in order to gain altitude and dive vertically providing the maximum kinetic energy. Guidance controls all channels for final precision approach and impact.

Each of these guidance phases will have triggers in JSBSim guidance/executive to perform the maneuvers and switch between phases automatically. (It already does that of course)

Regards,
Uriah

[bomber]

Well consider that a plane before it takes off is technically a ground vehicle....

The jsbsim flight models I do already have damage hooks integrated into them....

I agree that for the collision coding the 3d modellers would be required to create a simplified series of not displayed mesh hit boxes that represent the external structure of the vehicle...  8 per wing, 3 fuselage.. Don't go overboard with these as online play would require a lot of data.. Once the shooting object has determined that a mesh has been hit, communicate this and allow the damaged vehicle to determine if within this area there are other systems such as fuel tanks, undercarriage,  engine etc and the probability of being hit.

Tracking individual bullets from a vehicle firing 6000 a minute plus return fire is gonna be hard enough without trying to determine if one of them hit a fuel valve or not.

[Uriah]

Not that technically any aircraft on the ground is not a ground vehicle, I'm specifically talking about the different between Bullet and JSBSim objects, and them currently being mutually exclusive. I can target an aircraft sitting on a run way, but not the tank next to it. Like I said above, to simplify things for testing missile guidance, I will just define the target position in the script and not target an object, I'm still having difficulty with doing that consistently. I think getting the guidance laws written and working is more important for now, and gameplay can wait.

You were able to implement a system for damage is JSBSim?

I am really interested in three types of damage; heat, dynamic pressure (for orbital reentry and angle of attack) and of course collision/impact. That would be awesome to be able to integrate damage to individual meshes/collision meshes for those types of damage.

I believe the physics computational demand between multiplayer and solo-sandbox will be different by many orders of magnitude, I do not see any problems for solo-mode, but multiplayer may present performance roadblocks.

Regards,
Uriah

[bomber]

Most combat flight sims create a flight model first and then put damage as an add on.... This produces a very awkward feel to the whole damage experience....  Some use 0, 33, 66 & 100% damage graduation with the whole wing falling off at 100%.

For a planes flight model there are 3 factors
1) thrust
2) lift
3) drag

On modern planes there's also

4) systems

But losing an instrument won't make the plane fall out the sky....although a hydraulics failure might. So it's a grey area depending upon what system is damaged but as this normally effects 1 2 or 3 to bring the plane down systems cannot be considered a primary flight requirement.

When creating the flight model in jsbsim I added a graduated damage effect into the 22 lift/drag generating parts. The thrust is another issue but when I've finished with jsbsim piston code I'll have that nailed.

Online play will create issues but to my mind working them out now is a whole lot better than having to rewrite.

[Uriah]

Very nice, I would like to scope that code when you finish up. It would be very useful in all areas!

[bomber]

The Ask 13 glider comes with it... Each point mass area also has a damaged value, initially at 1, reducing this value reduces the lift, increases a certain type of drag for the area concerned.

[Uriah]

Thank you!

[PytonPago]

Well, most modern flyers have now around 3 computing units (whyte a reserve one), each doing its stuff, but capable to take over any that goes bad. So from electronic point of view, you have to aether hit all, cut off the power-supply (aether engine-generated and batteries) or scramble all sensors and control servos they're connected (or at least those doing fly-by-wire for fighters, witch, at least for the modern flankers, are more needed as people would think). Meaning, leaving damage on the aerodynamic state of the plane ( physical damage to it ) is ok. Any on-board systems can be just derived from them when some would want. I just hope something like that will be possible for ground-units. (it would go, but not sure if all stuff should be just in the single .js file - could make a lot of problems)

[Uriah]

Right, battle proven aircraft have multiple redundant systems in separate locations so if one is destroyed you can still fly. Just like DCS, or any other good combat flight sim, simulating component damage and failure modes is possible if there is already damage in JSBSim it should work in OT. Collisions is still the big road block.

That reminds me, I am also going to simulate electronic countermeasures , electronic warfare, against the bomb or missile seeker, such as jamming and signal interference.

JSBSim uses Sensors to simulate real world data, instead of perfect inertial measurements calculated from the equations of motion, it adds noise to the output signal. I could use this to simulate sensor Signal to Noise Ratio jamming, of any type, radar, IR, RF, GPS, etc... to effectively confuse the bomb/missile guidance system.

[Uriah]

Well... it kind of works.

[PytonPago]

FLYES THAT THING ! 

[M7]

Could you make a quick video?

[Uriah]

Yes most definitely, but not quite yet!

I am having difficulty running the F107 turbofan at the moment, and because of that I only have 12 seconds of thrust from the Solid Rocket Booster, and I am not able to fly to the target coordinates. Not to mention it is flying blind not knowing the target altitude (I just need to add this to the targeting function), or ground slope/ ground cover. Also, I am having problems with the terrain follow mode. Since I use current AGL to set autopilot altitude, the autopilot attempts to fly a flight path which is an exact profile of the terrain contour along the ground track below. As you can guess, it tries to climb over a steep mountain at too high an angle of attack and stalls, instead of making the pitch correction sooner, and at less of an angle. The smaller the AGL hold distance, the worse this problem becomes. I need to filter the AGL input so it is smooth (Kalman filter would be great), and take the AGL measurement from a point some distance (not far) ahead of the aircraft so it knows when there is an elevation gain before it reaches that point. In terrain follow mode the Tomahawk should have a range of over 600 miles, and greater than 1500 miles for high altitude level flight.

http://jsbsim.sourceforge.net/AutomaticFlightInJSBSim.pdf

Once I have something significant to show I will definitely record a video hitting targets at the White Sands Missile Test Range I have been working on!

-Uriah

[Uriah]