Human Modeling for Successful Team

Modeling Team Cognition(チーム協調モデリング)

This work introduces the modeling of team cognitive processes based on a theoretical model of “we-intention.” The proposed model is composed of three layers of mental components: individual mental components, belief in one’s partner’s mental components, and belief that one’s partner believes in one’s own components. We also modeled and categorized how these mental components are acquired, shared, and modified in actual human-human cooperative activities based on observations and post-experiment interviews. We identified five major strategies along with the cognitive processes for their own situation awareness and intention in establishing shared cooperative activities: 1) observation, 2) communication, 3) knowledge-based inference, 4) cognitive inference, and 5) examination of the consistency among their own mental components. This model can be used to trace and predict the processes for sharing team awareness and intention in human cooperative activities, which can be then used to drive a system to augment human performance in a cooperative manner.

協調活動を可能にしている人間の認知的メカニズムは未だ解明されていない。人間同士の巧みな協調を説明、理解、応用する為には協調に関するヒューマンモデルが必要となる。本研究ではwe-intentionやjoint intention, theory of mind等の理論に基づき、自分の心的状態,相手の心的状態に対する信念、相手の自分の心的状態に対する信念の3層からなるチーム認知モデルの提案を行った。プラントシミュレータを用いた協調実験や航空管制における発話分析などによってモデルの妥当性の検証を行うと同時に、1)チームワークの評価、2)協調作業に適したツール、環境設計の指針・評価,3)知的インタフェースエージェントの開発等への応用を目指している。 

Mutual Beliefs in Teamwork

Model of Team Cognition*

Team Cog. in Four-Person Team*

Team Simulation (チーム協調シミュレーション)

This study is to model and simulate human-human interaction based on a concept of mutual beliefs, aiming to describe and investigate the cognitive mechanism behind human interactions that is a crucial factor for system design and assessment. The proposed model captures four important aspects of human interactions: beliefs structure, mental states and cognitive components, cognitive and belief inference processes, and meta-cognitive manipulations. This model was implemented with a Bayesian belief network and some test simulations were carried out. Results showed that some basic qualitative characteristics of human interactions as well as the effectiveness of mutual beliefs could be well simulated. The paper concludes by discussing the possibility of the application of this model and simulation to universal access and HCI design and assessment.


Team Cog. in Collaborative Diagnosis

Simulation Results: team cog. accuracy and sharedness

Team Cog.  and shared SA in an ATC Accident*

Team Situation Awareness Measurement (チームSA計測)

Our model provides a new perspective on the notion of sharedness of cognition or mental constructs such as situation awareness, mental model, knowledge, and so on. One of the innovative points is that this model can clearly distinguish what is really shared among team members and what is only believed to be shared, and can provide theoretical formulations for that. Based on this model, we have proposed a variety of measurement methods for shared and team situation awareness and have been conducting many experimental studies on team situation awareness.  

提案したチーム認知モデルの特徴は,チームメンバー間で実際に共有されている認識(what is really shared)と共有されていると信じていること(what is only believed to be shared)を明確に分けて記述・定義できる点である.ヒューマンファクタ研究で盛んに行われているチーム認知に関する研究は,この二つの視点が明確に分けて議論されていない.理論的に定義された提案する共有(sharedness)の概念に基づいて,チーム協調で重要と考えられているチーム状況認識の測定法の開発や様々な被験者実験によるチーム状況認識の測定およびその詳細なメカニズムの解明を行っている. 

The Definition of Sharedness

Experiment on Teamwork 

Experiment on Teamwork 

Team Communication Analysis(チームコミュニケーション分析)

Team commnunication contains rich information about the quality of teamwork. On the other hand, there are various aspects to be assessed in team communication, which have been long studied in different disciplines such as communication analysis, verbal protocol analysis, conversation analysis (CA), discourse analysis, etc. , focusing on different aspects of team communication. We have been working on developing a method to analyze such different aspects of team communications in an integral and systematic manner, including phonetics, contents, function, and intention, of the utterances.  We are also working on developing a method to convert discrete and qualitative analysis results into quantitative and continuous, and aiming to integrate qualitative and quantitative analysis methods to explore good behavioral and cognitive indicators/guidelines for better team cooperation. 


Team communication analysis

Recurrence plot of communication contents

Mutual information of team communication