Researchers are using Machine Learning to make VR sword fights more realistic.
Researchers at the University of Bath, England, partnered with Cambridge-based game studio Ninja Theory (Hellblade VR, DmC: Devil May Cry, Eslaved: West to Odyssey) to deliver more realistic, data-driven sword fighting.
“With screen-based video games, a player presses ‘attack’ and their character displays a sequence of animations, but in a VR game, the player input is much harder to process.” said Dr. Christof Lutteroth, Senior Lecturer in Computer Science at the University of Bath.
These repetitive actions can easily break the immersion of players, so researchers started working on a new system that uses machine learning to deliver more realistic sword fights against less predictable opponents. The programmable system, called Touché by the team, is supported using a collection of pre-recorded sword fighting animations captured using motion capture technology.
Using a form of artificial intelligence called "machine learning", non-playable characters can actually learn from these recorded moves and increase their skill levels. Opponents can also be trained to respond more realistically to the attacks and positioning of the human player, resulting in more challenging, more complex engagements. According to researchers, developers can control each enemy "skill and aggression" and scale their experience to meet the needs of any player, regardless of their abilities.
Machine learning is an artificial intelligence (AI) application that provides systems with the ability to learn and develop automatically from experience without being explicitly programmed. Machine learning focuses on the development of computer programs that can access data and use learning for themselves.
The learning process starts with observations or data, such as examples, direct experience or instructions, to look for patterns in the data and make better decisions in the future, based on the examples we provide. The primary goal is to enable computers to learn automatically without human intervention or assistance, and adjust actions accordingly.
“Touche increases the realism of a sword fight by generating responsive animations against attacks and eliminating non-reactive behavior from characters. Using our model, a game character can anticipate all possible fight situations and react to them, resulting in a more enjoyable and immersive game experience.” added Dr. Lutteroth.
While games like Blade & Sorcery are approaching delivering realistic VR sword fights, there is still a sense of automation in the movement of enemy characters. After just a few hours in the game, even those who are not familiar with VR will find themselves relatively easy dancing around enemies. Hopefully, by including machine learning in the development process, we will soon see a new standard for VR sword fighting supported by the human movement. If this type of model were to have an increased range of movements, maybe one day people could use it to learn real world fighting and martial arts. It could be useful for military combat training in the future also.