Milestone has developed quite a bit of racing games with its biggest being the MotoGP series and the Monster Energy Supercross series. The company has been working with the engineering company Orobix for the past two years in developing a new neural network for artificial intelligence to implement into its racing games. The technology is called Artificial Neural Network Agent, or A.N.N.A., and it will be incorporated first for MotoGP 19 releasing in June. The basis behind this AI is that developers will not be able to dictate preset behaviors in games, but rather set goals and provide tools for teaching the AI to achieve those goals. This process is being called reinforcement learning and it allows one to obtain and extremely aware artificial intelligence.
The reinforcement learning process takes several months to complete. This will allow developers to simulate over 200,000 races per day. This allows the AI to go through two phases: the exploration phase and exploitation phase. During the exploration phase, the AI understands if consequences of its actions are useful or not to reach its goal. The exploitation phase will have the AI take advantage of its experience and knowledge to get the best reward available. The result allows for an extremely realistic and natural driving behavior, with driving systems, maneuvers and techniques very similar to those made by a professional rider.an extremely realistic and natural driving behavior, with driving systems, maneuvers and techniques very similar to those made by a professional rider. Lap times will closely resemble the real-life counterparts.
“We’re facing one of the biggest innovations in Milestone’s history” said Luisa Bixio, CEO, Milestone. “The amazing work done by our engineers, in collaboration with Orobix, it’s a great example of two independent companies innovating in an industry that is often controlled by big corporations. It’s a demonstration of creativity, passion and capacity to innovate”. You can watch a video below covering some of what A.N.N.A is capable of.