Ilici

Members
  • Content count

    3500
  • Joined

  • Last visited

Community Reputation

862 Good

About Ilici

  • Rank
    Contributor
  1. Anybody left from the 2003 crowd?

    I remember getting banned once, end of 2003, for a post in one of the random threads in the lounge taunting Dave. I felt like such a bad boy.
  2. That's pretty impressive and looks like fun. What's the main challenge in coding a bot? Do you need to make separate bots for single tanks and formations? I'd change the site layout a bit if you want to attract users. I think this year's Google AI Contest site does a pretty good job of telling the user what it's about and how to get started. I'd also try to lower the game viewer loading time.. it's pretty annoying to have it load for so long. [Edit: Fixed link tag] [Edited by - InnocuousFox on November 12, 2010 10:44:17 AM]
  3. Sensation of Speed

    I'm working on a flight sim game and I can tell you, the problem with creating a good sensation of speed has been incredibly tricky. What gives you a good sensation: - detail textures help out quite a bit - they give you a lot of moving pixels under your plane. Also, the bigger you terrain texture can be, the better - ground scenary - trees or tanks or whatever you can put there will allow you to have a reference to percieve speed - faster speed near the ground, slower at high altitude - responsive controls - make the plane turn and roll quicly and you get a lot of movement of the terrain on screen - low altitude flying - fly low and make your terrain high (so you can go up/down mountains) - other planes - this will give you a reference scale when visualizing a far away mountain as well as a speed reference - how much does the plane move agains the terrain in relation to its own size
  4. Ugh.. have you actually checked the rates? The USD is at 2008 levels, which is quite low but then again gold investment isn't doing extraordinary well either.
  5. Internet Boosters

    Quote:Original post by worldextermination Thanks for the info, but wouldn't the OS be better suited to everyone's needs (indeed what they market it for) if Microsoft itself used better runtime libraries to better suit their target audience? Because although they do not say it, it seems to me that they imply that Windows is aimed towards everyone, developers, gamers, businessmen, etc as a whole. But on the contrary, they have been trying to improve their OS. It might be because Microsoft is trying new trial and error tactics, and Vista was an error. It seems to me that Microsoft is trying to steal your Megahurtz. You clearly need some bad ass Megahurtz restoration services provided by registry cleaning, mem-to-swap file forcing and time scaling software.
  6. Parabolic motion smoothing

    You could try using an exponential function for the angle (not ang velocity), something like: 1-exp(-x*5) On the 0-1 interval it increases quite fast at start and slows down quite a lot at the end. Now if you have a full flown physics engine in there and you don't want to/can't prod the angle straight away, you can get something like this with Forces/Drags. Let's say the maximum velocity is 'vmax' and the time to reach it is 'tmax'. Apply a torque that is equal to vmax*f*deltaAngle*C where f is a constant I'll get back to later. Delta Angle is the difference between the current angle and the final angle (how much you want to rotate). The apply a drag torque that is D = v*C. Where C = -log(0.01) / tmax. This expression gives a drag coefficient that will allow acceleration/braking from 0 to vmax in exactly tmax with 1% error. Thus you have: T = vmax*f*deltaAngle*C - v*C. v = v + F*dt angle = angle + v * dt; and deltaAngle = angleWant - angleCurrent The choice of f gives either a wobbly but faster rotation towards the desired angle or a slower but stable rotation towards it. The fastest stable rotation is obtained for f = C / (4*vmax) All from here.
  7. [SOLVED] Physics 2D drag problem

    Quote:Original post by NotTheCommonDose I'm starting to get the feeling that setting a maximum velocity as a result of drag isn't going to be simple. It's not simple but not too hard either: Write the equations of motion: F = a - k*v; v = integral(F) => a - k*v = dv/dt; Or as a diff equation: dv/(a - kv) = k*dt Thus you get v = vmax*(1 - exp(-k*t)) You want v ~= vmax (equal to some degree) => 0.99vmax = vmax(1-exp(-k*t)) for some t you want: Solve for k to get the constant that will give you a certain maximum velocity, reached within the required time. From here
  8. Logic to keep vehicle on course.

    My solution was something like this: It ensures you hit all your waypoints.
  9. Logic to keep vehicle on course.

    First of all you need to determine what is your minimum turn radius. If you can turn at 3 degrees per second at speed S you get: R = v / w = v / (TORAD(3)) Now take points B and C. Construct the vector from B to C. Take the perpendicular to it (cross product with Up vector) and multiply this vector by the Radius. Now you have the center point of a circle that is tangent to BC of your minimum turning radius. You now need to find the point where you have to enter this turn - the point where a line from A is tangent to this circle (center X, radius R). You can find a good method (I'm using it myself) Here. Now you have 2 states: moving and turning. You move in a straight line to the turn entry point. When you reach it you start turning at the maximum turning rate. You'll get to B on a direction pointing straight to C. Instead of using C as the next point, apply the algorithm for finding the turn entry to points C and D. Then calculate the initial tangent using this point so you get a turn that leaves you pointing straight to this future turn entry point. It gets (a lot) more complicated if you have acceleration on turning speed (instead of instantaneous change).
  10. This is quite a tough one. I'm working on a flight simulation and aircraft are modeled as a rigid solid moving under drag. I'm making the AI system so I need to plan a path for the airplane - thus i need to spin it around so it points in the right direction. Rotation is determined the usual way, by integrating torque twice. Torque is equal to the force applied using the flight stick to which a linear drag is applied. Thus T = max_rotational_velocity_konst * StickInput[-1..1] - w * C C is the drag constant and is calculated from the 'max_rotational_velocity_konst' constant and the 'time_to_reach_max_vel' constant - so that the designers can model a plane that achieves a certain angular velocity in a certain time (if the stick is kept at maximum during that time). Next, the angular velocity is equal to the integral of Torque. w = (T / (inv_mass * inertia)) * dt; And finally, the angle of rotation is equal to alpha = w * dt Unfortunately, several stick movements need to be input each frame (so that behaviors can be tweened - ground avoidance for one should increase in priority as the plane gets closer to the ground) and I can't change this. Thus my AI will input stick commands to make the plane roll to the right direction by measuring the angle between its current rotation and the desired rotation. I take the CrossProduct of the two 'Up' vectors so I get the sinus of the angle. I use this value as the StickInput value - thus it will achieve 0 input when its in the right angle. Thus: T = max_rot_speed * sin(alpha) - w * C w = integral(T) alpha = integral(w) What I want is to find the angla as a function of time in order to determine the time it will take for the aircraft to achieve a certain rotation. I've already found a way to do it for constant force from here, but I don't know how to use it for a varying force. Any ideas?
  11. Quote:Original post by Oberon_Command Quote:Original post by MikeTacular Quote:Original post by Samith Oh god I just wasted so much electricity reading this thread! Why!?!?I hope you're happy knowing that you're the reason the ice caps are melting, douche-fag. I have my computer in a refrigerator to cancel out the heat-generation, but because of you I've had to put that refrigerator in another refrigerator in order to try to cancel out your heat-generation. This makes it really, really inconvenient to write this reply. You'll be happy to know that refridgerators actually generate heat (by moving the heat from inside to outside of them), so every time you add a refridgerator on top of another one you're contributing even more to the heat death of the universe. [grin] But then you put everything into another, bigger, refrigerator and it's all cool again.
  12. Ex done stole my wii!

    A! Time for the classic head-butting into the tits eh?
  13. Sony's E3 motion tracking

    Quote:Original post by swiftcoder The obvious starting point is to use the provided blob-tracking, run some sort of histogram test to detect the correct colour, and some sort of heuristic to check the blob is (roughly) circular. I have a feeling though, that there may be a more efficient approach possible, given the pre-existing knowledge of the targets. (keep in mind that while I have experimented with computer vision, I am pretty much a novice in this area) You could also look at the Circular Hough Transform - it identifies circles in images provided you give it an approximate radius.
  14. My PC repair business flyer

    I find it kind of dull. I don't think the 'professional' look is best - i'd rather go for something more funny-geeky. First of all that picture shows a perfectly OK PC - won't really catch my eye if i'm having computer trouble. I'd go for a more explicit picture, something that shows a broken computer and someone who's pissed off about it - your customers would relate to that a bit more. This would be a good example: You make a big list of the problems you fix - but there's nothing really standing out in the picture - the tile definitely isn't - make that bigger, change the background and the font - go for something bolder and cleaner (something sans-serif maybe?) The font on the left is really dull and the format of the column makes it look uninteresting. Either make both columns the same width or ditch the 2 column format and place the lists one on top of the other. I'd make them much smaller and place them at the bottom in order to have the picture and title blown up in the middle. The ribbon is an ok idea, but the one you picked doesn't fit at all - everything has clean lines and flat colors, the ribbon looks aloof. Pick a cleaner one, make it bigger, rotate it a bit and slap it in one corner. Change the font on your brand name and make some kind of a logo (even a simple one will do). Your current font looks like something off a Roman tombstone, get something that makes people think of computers. Just my $.02
  15. GDNet at the Movies requests

    Quote:Original post by Oluseyi As for Angels & Demons, I wasn't planning on seeing that (didn't see The DaVinci Code, don't like Dan Brown). Maybe you can persuade me otherwise? Well, I found it to be severely lacking in entertainment value. The action manages to get your interest for 10 minute intervals before you have a WTF moment about the script. Sometimes it gets embarrassingly bad, the whole science-religion talk was rather silly and too in-your-face and they had that spooky choir music running for about half of the movie. Plus, though I'm not religious at all, if I were, especially catholic, I'd be rather pissed off about the idiocy spewed out about the church - at the end I was surprised they didn't anoint Tom Hanks as freaking Pope. All in all, a 3 out of 10.