Team task allocation support by using the human body as a sensor. Real-time estimated stress of team members as input for decision support for military commanders
Steen, C.E. van der
MetadataShow full item record
The quality of decision making by military commanders can be enhanced by providing extra information. Modern military equipment can be expanded by devices that monitor the condition of soldiers. Because the condition of a soldier influences his behavior and thereby his performance, this can be considered valuable input for the commander in making strategic decisions. The purpose of this study is to investigate whether the quality of decision making of a military commander improves when he can monitor the condition of the members of his team. An experiment is carried out with 18 participants. Per session, five team members play a virtual first person shooter game where they have to work together to eliminate enemies. By using interbeat intervals, a stress model estimates the stress value of each team member real-time, as an indication of the condition of the team member. One commander has to allocate the team members to one of three tasks. To provide the team leader with the information, a decision support system (DSS) has been designed and built. This DSS presents the information visually and can give advice to the team leader in making decisions. The team leader can get three types of support: a) no support, b) stress information support, c) stress information and decision advice support. A fourth condition is added to check the quality of decision making by the model individually (without human interference). The purpose of the experiment can be divided in two areas: 1) to test the relationship between the stress value of team members and their performance within the experiment, and 2) to test the effect of different types of support on the quality of decision making of the team leader. Firstly, we expect that the stress value is a negative predictor for performance. Secondly, we expect that a team leader makes better decisions when he can consider the advice from the DSS, as opposed to making the decision based on solely verbal communication with the team members, and only stress information support. The analyses show that no clear relationship is found between the stress value and the performance of the team members, although trends indicate that the severeness of stress plays a role in this relationship. Though the DSS model bases the advice on this assumption, the model individually yielded better quality of decisions than the human team leader with both stress information and decision advice support. The subjective support effectivity results show that the team leaders prefer the decision advice support. All other relationships between support effectivity and conditions were not significant. An important finding is that the objective and the subjective results of support effectivity did not match. Also, both expectations could not be confirmed, which indicates discrepancies between the theory and the results of this experiment. This can be due to the lack of stressfulness of the team member task and the lack of knowledge of human team leaders to deal with the stress information.