Collaborative intelligence is about helping people and machines be better together. Humans and machines both have their strengths. People bring adaptability, creativity, moral values and higher-order thinking. Machines on the other hand, deliver power, scale and precision.
Many early adopters already recognise the potential of collaborative intelligence. In medicine, robots and humans collaborate in complicated surgeries. In search and rescue, robots venture into dangerous terrain while their human supervisors verify and act from a safe distance.
Within this rapidly growing field, CINTEL is developing the science and systems for humans and technology to work better, together.
A key piece in the puzzle of improving robot-humans collaboration is increasing their awareness of where the other is and what they are doing during critical moments.
At any time during a collaborative task (otherwise known as a mission), a human must take in information about the robot’s status. They must also remain alert to the bigger picture and risks in the environment. Attention is a hot commodity during a mission. The awareness that human operators and robots have of one another waxes and wanes.
Hashini is working to unpack this understudied relationship within CINTEL’s Dynamic Situational Awareness project. Her postdoc research examines how we can help both humans and robots maintain optimal awareness. To do so, she studies our world-leading robot-human team in Data61.
The team are currently developing technologies in which AI models and farmers work collaboratively and learn from each other. Hashini will also study other teams from manufacturing, surgery, and search and rescue domains.
We welcome your biofeedback
As well as establishing a better understanding of situational awareness, the project also seeks to improve it.
One possible way to do this is using biosignals. During her PhD at Monash University, Hashini developed an AI-powered suite of sensors. These sensors can detect an episode of anxiety using biosignals like sweat, the rigidity of physical movements and heart rate.
Similarly, biosignals could be used in human-robot teams to track a human’s attention and awareness by monitoring their eye gaze.
“Imagine you’re about to do a mission with a robot fleet, you come and put on a couple of unobtrusive sensors like a smartwatch and some glasses,” Hashini said.
“While you’re doing the mission, in the background, AI is analysing whether your attention is focused on the right thing at the right time, whether you’re cognitively overloaded and whether you’re stressed.
“It’s using that information to potentially redirect your attention using audio, haptic or other sensory cues to reduce workload and stress,” she said.
When combined with information about a robot’s current and predicted future situation, this AI-enabled tap on the shoulder has the potential to help overcome human limitations and improve decision-making.
Programming the future
Hashini has just kicked off this exciting work. She is currently interviewing collaborative intelligence collaborators to find out their requirements for dynamic situational awareness. Next year, the team will test whether they’ve achieved optimal situational awareness through a series of simulated experiments.
We’re still some time away from seeing human-robot teams become part of the everyday. However the CINTEL team is designing a future in which humans and machines are harmonised and working together, to the best of both of our abilities.