Department / Course Research Organization of Science and Technology
Title / Position Senior Researcher
言語種別 英語
研究概要 Development of a streaming multiplayer motion game with audience-participation mechanisms for promotion of well-being

My current research is to develop an audience participation game and make it available on a streaming platform. It is planned to be a fighting game, in which two on-screen characters fight each other. The game characters are mainly controlled by AIs, but human players (audiences) can tell the AIs to perform actions they want by performing body motions in front of a camera to send commands. Beside such commands for requesting AI actions, there will also be other commands for impacting gameplay via chatting messages, like those for buffing and debuffing characters (e.g., dropping some items to enhance or to damage characters). Battles will be streamed online, welcoming anyone to join. The system is designed to promote three dimensions of well-being of audiences: physical, mental, and social; “physical” is by promoting body movement, while “mental” and “social” are by promoting fun collaborative gameplay. It is also noted that real-time motion detection using only a general webcam is a very challenging task in the field of image processing.

Existing Problems:
1.Need for eliminating the need for motion controller: Motion games are generally developed specifically to be used with a certain motion controller. Nevertheless, with a programming library called OpenPose, it is possible to detect motions using a low-cost webcam, hence eliminating the need for motion controller.
2.Need for Self-Playing AI: It is expected that the characters in the proposed game will be mainly controlled by AI. This is to respond to the diversification in the time zones and preferences. We embrace the technology of procedural play generation (PPG) to build an AI to play the game. Such AI capable of generating gameplay by itself is called Self-Playing (SP) AI. We plan to build AI for generating SP contents that optimize mental and social well-being of their audiences.
3.Need to evaluate effects of audience-participation games on health promotion: In previous research studies, there had been no full-scale attempt to promote health by SP contents on any live streaming delivery platforms. In the proposed game, human players can partially control AI characters using body movement, hence promoting their physical well-being (the AI executes skills based on motions performed by human players). Mental and social well-being are promoted through SP contents generation algorithms.

Research Purposes, for solving the above problems:
1.To develop a multiplayer audience-participation game that can be played by using body movement.
2.To develop an AI method for generating SP contents (battle gameplay).
3.To develop mechanisms for promoting physical, mental, and social well-being.

Motion games via streaming: We will apply image processing technology for detecting motion using any webcam.
Content generation using AI for game live streaming: Studies on audience-participation games are a new trend of research, but generation and distribution of contents is usually done by human players. We aim to automatically generate gameplay using artificial intelligence technologies.
Game live streaming for promoting well-being: We will be the first group to explore the potential of game live streaming for health promotion.

There is no existing study targeting health promotion through audience-participation games and no existing audience-participation fighting game.

Expected impact
Success in automatic generation of contents for live game streaming will impact mass of people, as according to a recent global market survey, the population of gameplay audiences will be 860 million in 2022. In addition, the proposed game can be further used as a platform for research on AI for generating SP contents, as well as on features for audience participation for promoting well-being.