For decades our visions of large space stations orbiting the Earth or humans colonizing other planets have included intelligent robotic assistants. They are now, at least partially, a reality; but they are not teammates yet -- current systems have limited autonomy and the most typical scenario is still human guided teleoperation.
We propose here to demonstrate the feasibility of autonomous robotic teammates in the space domain, and in particular in a relevant scenario of joint construction in earth orbit manned flights. The goal is to develop a human-robot joint task demonstrator -- using a simulator and by conducting tests in a physical implementation at the MAR Lab’s facilities -- that is able to communicate and work intelligently with astronauts on mutual tasks. Our demonstrator will be based on the most recent developments in neuro-cognitive robotics and supported by the consortium’s ground-breaking work in these areas. The work is integrated in the ESA Project proposal JAIS – Joint Action In Space.
Objectivos/Main goals and tasks:
· Detail the joint task, in particular the key performance parameters for benchmarking
· Train object recognition module for the objects involved in the joint construction task.
· Adapt the robot’s cognitive architecture for decision making in human-robot Joint Action, to the joint construction task
· Test of human-robot joint task: Teleoperation with different delays & autonomous robot
Metodologia/Methodology:
· Use of the Anthropomorphic Robot ARoS
· Conduct human-human and human-robot joint construction tasks
· Use of programming languages (C/C++)