Personal tools
You are here: Home News NIFTi Joint Exercises 2011: Studying human-robot teaming
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. 

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


NIFTi Joint Exercises 2011: Studying human-robot teaming

Early July, NIFTi met at the Training Center of the Fire Department of Dortmund (FDDO) for the NIFTi Joint Exercises 2011. The goal: Study human-robot teaming. Given a team of humans (researchers, end users), a microcopter, and the new NIFTi robot - how could and should they interact to jointly explore a disaster site, like a multi-story apartment building on fire?

NIFTi is about how humans and robots can work together, in a single team. This moves well beyond the idea of a single human pilot tele-operating a robot through the rubble. Instead, humans and robots form a team, possibly distributed over several locations, aiding each other in building up a (distributed) situation awareness through jointly exploring a disaster site. If the robots have sufficient autonomy to perform tasks under supervision rather than through tele-operation, and if the communication and dynamics of coordination work well within the team, then this would mean a big step forward in deploying robots under realistic circumstances. 

Of course, those are several Big Ifs. Both autonomy and human-robot interaction have been identified as two big issues in developing robots for assisting in USAR missions. We have not solved those yet, but, truth be told, there is also still a lot to be investigated in what it means for a robot to be autonomous in a human-robot team context, or how humans and robots are supposed to interact within a team. 

We organized the NIFTi Joint Exercises 2011 (NJEx 2011) with the explicit goal to get more insight in the latter issue: Study the dynamics of interaction in a human-robot team, with a particular focus on how humans interact "about" robots while coordinating their efforts in exploring a disaster. 

NJEx 2011 took place early July, at the Training Center of the Fire Department of Dortmund (FDDO). We build teams around researchers and firemen, with each team getting a microcopter and a brand-new NIFTi UGV at their disposal. FDDO provided radio handsets for communication, and arranged for us to have the huge mobile command center available during NJEx -- the Einsatzleitungswagen/ELW 3. 


Inside the ELW3 we built up a command post, with several displays for visualizing information coming from the microcopter and the UGV, and controlling the UGV. For each team, people playing "mission director", "UGV pilot," and "mission specialist" would be at this remote location, whereas other team members for safeguarding the robots (the "safety directors") and piloting the UAV would be in-field. 

All the communication going on between these team members went over the radio handsets, and was recorded (digitally) at servers in the ELW3. 

We started the week with lectures on how FDDO organizes its response teams, and how people within and across these teams communicate. After which we naturally extensively tried out how to do that in reality, using the radios ... 

Tuesday and Wednesday we used to practice with the new NIFTi platforms, using Euro-Pallets to build up a variety of mazes up to NIST Orange/Red complexity. Each team also had to perform "competitively" using the full team setup (remote UGV ops & mission director, local UAV team) to traverse a particularly complex maze. 

 Maze at NJEx 2011


Wednesday afternoon and all of Thursday we then went for the real mission: Exploring a multi-story building "on fire." The UGV would drive in through a door on the ground floor, explore the (smoked up) corridor and rooms there, then drive outside again, take the staircase outside the building, and move to another door to enter the building on the first floor. The UAV could also enter here, and could be used to explore further rooms, which typically also were smoked up. (We ended up not going via the winding staircase inside the building, as this turned out to be too risky for remote tele-operation: Situation awareness was too limited to deal with scaling slopes that were occasionally up to 70 degrees...)

 Red building

And this is "of course" where all the training earlier in the week, together with the firemen from FDDO and Vigili del Fuoco (to tell use how to really do it), started paying off. More and more, people moved away from saying "move your left flipper up!" to communicating about situation awareness, requests for observations from the UAV, talking about what tactics to follow to explore the next floor. 

 Robot exploring smoky red building

Team-work, in other words. 

Over the next period we will be looking more closely at the recorded data, in conjunction with the video-coverage of each mission, to analyze what was going on in terms of team-dynamics. These first insights will then feed into further developments of the NIFTi "GUI" (which is rapidly transforming now into a Houston-style operations center ...), and the interaction between spoken dialogue, planning, and visualization to adapt team/robot operations to the current task- and team context. 

For more photos of NJEx 2011, see our FaceBook photo album: [ WWW


 ELW3 teaming setup



Document Actions