Highlights –
- REMP algorithm will enrich the collaboration between humans and robots.
- The research team wanted to investigate the extent to which gagging the capability of robots can help humans assign suitable tasks for seamless collaboration.
Led by Prof. Song Chun Zu, a team of researchers at the University of California, Los Angeles (UCLA) Center for Vision, Cognition, Learning, and Autonomy (VCLA) has designed a way that can help align a human user’s assessment of what robots can achieve with their true and advanced capabilities. Published in IEEE Robotics and Automation Letters, the paper says that the approach is based on a new algorithm that enhances the physical cost and expressiveness of a robot motion at the same time to understand how well human observers will estimate its reachable workspace.
Xiaofeng Xao, one of the leading researchers on the team to carry out the research results, said, “In human society, people have different roles based on their expertise and capabilities.” He added, “Such capability-aware role assignments allow humans to collaborate with each other more efficiently. We believe that when humans are working with robots, it is equally important for them to understand the robot’s capability, as a failure to do so can affect their trust in and acceptance of robots.”
California university’s research team and Gao believe that it’s important for humans to evaluate the robot’s capacity and capabilities accurately. Because if a robot’s abilities are underestimated, the user will not be able to use it efficiently. On the other hand, if a robot’s abilities are overestimated, users will be disappointed or may use it only use it in critical situations.
The core purpose of the research study was to assist humans in understanding the reachable workspace of robots through a capability calibration process involving multiple motion demonstrations. Additionally, the research team wanted to investigate the extent to which gagging the capability of robots can help humans assign suitable tasks for seamless collaboration.
Gao and his research team tested their REMP algorithm in a series of experiments involving human participants. The researchers compared its performance with two baseline methods that use pure functional motions and randomly traversing trajectories as demonstrations. The results of these tests were remarkable because they found that their algorithm could significantly improve the user’s estimation of the robot’s capabilities. Furthermore, it could also enrich the collaboration between humans and robots.
DARPA Explainable Artificial Intelligence (XAI) funded the research project through grants. They envision that this algorithm can help strengthen the collaboration skills of both present and newly developed robotic systems. The research team conducted their experiments online, hence they could test the algorithm’s performance on a 2D plane. They plan to develop their method in their next studies to ensure that it is applicable in 3D environments as well.
Experts’ view
“We propose reachability-expressive motion planning (REMP), an algorithm that generates expressive motion demonstrations to calibrate the perceived robot reachability via trajectory optimization,” Gao explained. “One unique feature of REMP is that it models how the human belief of the robot’s reachable workspace changes after each trajectory. As a result, it can improve the human’s reachability understanding of a robot quite efficiently, as only a small number of demonstrations are necessary to achieve decent calibration.”
“We are excited to see that when using our method, users perceive the robot more positively, as the robot is considered to be more reliable, more predictable and easier to understand,” Gao said. “These results highlight the necessity of building intelligent machines that are aware of people they work with and help us envision a better future where humans and AIs can work together.”
“As reaching is one of the most basic tasks in human-robot interaction, we believe understanding reachability greatly helps users understand robot capacities in different tasks,” Gao said. “We view our work as a successful first step towards a more general capability calibration setting. We are now also interested in using a variety of other modalities (e.g., speech, gesture) as means of communicating capabilities.”