I nonetheless would not have a reputation for my sport. It’s nonetheless in early levels of growth and concepts, however i want to share with you all this progress. It involves a 2nd racing sport, like a platform sport, with life like physics. The ideia is to evolve this sport to get a cellular sport engaged on Android OS, with perhaps 3d graphics (regardless of the 2nd physics, however I’ve some ideias on this), and multiplayer (nonetheless no ideia on learn how to obtain this).
Listed below are the present progress of it. Now it’s operating on Java desktop, however my subsequent step is to porting it to Android.
(that is the third video of it on youtube, however from now I’ll put up right here he identical manner as on youtube).
« Final Edit: Right now at 03:14:03 PM by giovz481 »
I’ve a brand new replace of the progress of my sport. I sucessfully ported the code do LibGDX, and now my sport have 3D graphics (very fundamental fashions but). It might probably run on Linux and Home windows desktop, as properly in Android platforms!
Right here is the video with a preview of it:
The subsequent step is to scrub some issues within the code, right some bugs, and put some cooler fashions for the automobile and the terrain.
I efficiently imported to the sport a 3D mannequin of a extra life like car. I additionally added texture UV mapping to the terrain, and the terrain even have physics properties (like friction and bouncing) related for every half.
Your mission seems to be very cool. It’s totally fascinating how one can change between the 2nd and 3d visualization of the scene. I might be all in favour of updates in your sport : )
Your mission seems to be very cool. It’s totally fascinating how one can change between the 2nd and 3d visualization of the scene. I might be all in favour of updates in your sport : )
Thanks for the reply! Quickly I’ll have new updates!
I put a background terrain, that was made with the implementation of a top area utilizing a top map which determines the peak of every vertex of the generated mesh. Additionally I added the controls of the automobile, just like the accelerator and brake pedals, which may be managed constantly (not solely on or off) with using the mouse or contact within the case of a cellular system. I additionally included an analog speedometer.
New replace of my sport! Now there may be an opponent automobile, managed by an machine studying mannequin!
It’s a supervised studying algorithm, I saved the information of myself driving the automobile, with parameters like for instance the pace of the automobile, angle, and distance of contact of every raycast. These information have been recorded in runtime and saved with a view to offline coaching a supervised mannequin utilizing python, after which exported the mannequin to java utilizing sklearn-porter.
Oh, a machine studying based mostly opponent. I actually like that. Within the video you offered the inexperienced and pink traces point out the inputs for the machine studying algorithm, proper? I’m not certain whether or not the method of “sensor rays” is perfect for the given setting. Possibly utilizing an outline of the present (and upcoming) slopes plus altitude and rotation of the automobile would result in improved conduct.
Sure, these sensor rays (that detect the space and angle of the purpose of contact) are a few of the inputs of the mannequin. However I used additionally some variables just like the pace of the automobile, angle of rotation, and its variation in time, the angular pace. I appreciated your suggestion of utilizing an outline of the upcoming slopes. Really the sensors are a tentative to detect this, however your suggestion made me suppose if there may be any higher different.
I created some obstacles so as to add extra enjoyable to the sport.
There are till now 4 kinds of obstacles: a form with free rotation and translation, a form with a hard and fast level with a joint which permits rotation, a dangling bridge, and a bridge with has elements that may fall when an object hits. The obstacles are objects with parameters that may be modified simply and spawned wherever on the sport. I made a lane system the place the objects can solely work together with their respective lane (first, second or any). There’s additionally a “mud” property on the terrain, which makes troublesome to maneuver the automobile.
I made a essential advance on my sport. Now the automobile’s engine has sound, based mostly on its rpm and acceleration. Additionally I added three essential objects to the sport, the engine, transmission and clutch. The transmission gears may be managed by the participant if it desires to. The clutch is managed routinely. I used Box2D joints to place all of it collectively and I’m very impressed with the outcomes, it is extremely life like and steady, even with a number of joints in sequence.
The sound is generated utilizing two recorded samples (one for low and different for top rpm), and they’re blended in actual time, based mostly on car’s rpm. The engine additionally has some parameters like idle rpm, reducing rpm, and most torque.
I’ve added additionally a minimap on the prime of the display screen, so the participant can have a high-level view of the monitor and its place in addition to its opponent place.
On this video I present the addition of a particle system and completely different sounds of tire contact and skid with the bottom. There are completely different particles and sounds for asphalt, dust street, snow and dust. Additionally I added smoke impact for the engine when the temperature is excessive.
I carried out a simulation of gas consumption and motor temperature, that are proven inside the new two gauges added.
I additionally gave a reputation to my sport. It’s known as Ignition Duel Racing.
