Learning AI to use APEX
Learning my AI to compete on racing circuit
Upgraded to ultrealstic suspension physical model + correct geometry calculation.
Now it will be able to drive real rally. Its driven by AI driver, but works perfect with my Logitech G29 as well with real feel Force Feedback. I also increased physics rate to 200hz.
For 8 years of Racing Engine physics development, each car has approximately 300 real engineering parameters to setup.
Features:
Aerodynamic
Tire : Pacejka + Brush model
Realistic Engine Setup,
Force Feedback,
up to 1000hz asynchronous physics rate with hight fps frame rendering rate.
Road corridor racing navigation system + Tactical Navigation system & AI
Data ready for Hydraulic Racing seats.
Destructible cars
Next steps:
- realistic car sound.
- engine drivetrain upgrade.
- AI improovements
- Rally enviroment
When will you have a test ready to play?
Very soon, i have finished realistic drivetrain which works with realistic engine, differentials setup rwd awd fwd, drift, race.
You can see that the telemetry has expanded a lot for final tests and polishing.
The most important part now will be the sound to achieve audio visual feeling also.
Looking forward to it
For people who thinks that making physically correct vehicle physics is just one day with AI todays, WRONG WAY ! I was trying to be the best friend of AI for a few months, Hat off, it real, i was totally executing Ai about what i do, shared samples. Many times AI do lets compare a code, and AI say no! T_Net ■■■■ always, . Not true, T_net is not math solve , thats the information for all of you trying to make copy of engine like that. T_net came from 1000000hz telemetry systems, not possible to solve this
in game. This took 8 years passively, 4 times totally reworked from real math