Personal tools
You are here: Home Publications Modeling the cognitive state of urban search and rescue robot operators in real time
About NIFTi
NIFTi is about human-robot cooperation. About teams of robots and humans doing tasks together, interacting together to try and reach a shared goal. NIFTi looks at how the robot could bear the human in mind. Literally. When determining what to do or say next in human-robot interaction; when, and how. NIFTi puts the human factor into cognitive robots, and human-robot team interaction in particular.   Each year, NIFTi evaluates its systems together with several USAR organizations. Rescue personnel teams up with NIFTi robots to carry out realistic missions, in real-life training areas. 
Impressum

This site uses Google Analytics to record statistics on site visits - see Legal information.

 

T. R Colin (2013)

Modeling the cognitive state of urban search and rescue robot operators in real time

Master's Thesis, Utrecht University.

Urban Search And Rescue (USAR) robots collaborate with humans in the wake of disasters such as earthquakes or terrorist attacks. These robots are being augmented with artificial intelligence to take over some tasks from their human operators, who are affected by extreme stress and workload. But they still require human supervision and assistance. This thesis investigates means to improve human-robot cooperation in USAR missions. Keeping track of the cognitive state of the operator would allow USAR robots to adapt to variations in his workload. A cognitive task load model is presented for this purpose: using a rule-based system, the model detects the recent tasks of the robot operator, based on system events such as interactions between the operator and the robot. The task history can be used to determine the value of three metrics of cognitive task load. Finally, these values are used as input for a naive Bayes classifier which outputs the most likely cognitive state of the operator. In future applications, knowledge of the cognitive state of an operator could be used to adapt the robot's interface and autonomy. To test the model, eight participants drove a shape-shifting USAR robot, accumulating over 16 hours of driving time in 485 USAR missions with varying objectives and difficulty. Over 30,000 events were recorded for 351 of these missions, and used as input for the model. Accuracy results were insufficient for real-world use (around 70\% for mental effort and 62\% for performance), with important variations between participants. However these results demonstrate that such a model can contribute, in a completely non-invasive manner, to estimating an operator's mental workload. The results were also insightful with regards to the factors affecting USAR robot operators' cognitive state, performance, and their preferred adaptation in terms of robot autonomy and interface. The experiment suggests that the model would benefit from taking into account individual participant differences, additional task-relevant data, experience and mental fatigue. A sketch of an improved model is given based on these results, opening promising new perspectives for the improvement of human-robot interaction (HRI) in USAR, and in other fields in which robots and humans cooperate.
Document Actions