NVIDIA suggests using machine learning for haptic feedback in games
NVIDIA is considering using artificial intelligence to improve haptic feedback technology in video games. This was reported by PC Gamer magazine, which discovered the corresponding patent of the company. In the document, NVIDIA suggests taking a new approach to implementing haptic feedback through machine learning.
The patent draws attention to the fact that video game developers now program in advance the moments of triggering feedback effects or associate them with certain audio frequencies within the audio recordings of the game. Instead, NVIDIA proposes to train artificial intelligence to detect certain “features” within the game in real-time and correlate them with different types of haptic feedback.
The patent provides little or no detail on how this would work. The processing tactile control algorithms will likely be assigned to the central processor or GPU or a combination of hardware and software.
The original AI-based tactile feedback algorithm, whatever form it takes, will require prior training to enable it to master some of the game’s features, such as video images, objects, and audio within the game. After that, he will adjust on the fly without prior knowledge of a particular video game.
“Feedback effects can be predefined according to a specific feature. Such a feature could be, for example, a shot being fired in a game. Upon recognizing it, the tactile control algorithm will give a certain feedback effect to the game controller, thus coordinating the player’s recoil effect with the recoil experienced by the game character in a video game, ”the patent says.
