State-of-the-art methods to maximise the performance of human-robot interaction
It is important for human users and operators to smoothly and effortlessly interact with many robots that function in the real world. Studying and developing methods for such human-robot interaction that involves complex robotic systems and multi-robot teams in a large environment becomes impractical when experiments are conducted only with real robots. A major challenge going forward is achieving maximum synergy between teams of robots and humans. The EU-funded TRAVERSE (Towards very large scale human-robot synergy) project addressed this open question by investigating and modelling human user and operator behaviour, perception and cognition through interaction with large-scale multi-robot systems. Project partners developed cooperative methods for robot team perception functionalities that are scalable to a very large number of robots. They implemented and verified a scalable multi-robot cooperative perception technique. It includes an optimisation-based estimator that runs in real time, robot-self localisation, teammate localisation and target tracking. Researchers modelled and studied human and system factors that affect the efficiency of joint duties when human operators interact with large-scale robot teams in performing a collaborative task. The joint task involved a mission to search for survivors during a disaster scenario. Extensive experiments with 35 human subjects led to two key findings. The collaborative task was performed much better where the robots were fully autonomous in exploring surroundings while human operators were only tasked with searching for survivors in the explored areas. When operators were additionally asked to control the robots, they performed significantly poorer if they did not possess good video gaming experience or training. A second finding showed only a slight benefit in increasing the number of robots in the team that performs the collaborative task. This is under the assumption that a single human operator is responsible for the actual survivor search and classification task. Increasing the number of robots, in a scenario where the robots together maintain certain formations, becomes detrimental to exploration as it rapidly increases the computational overhead. It was also found that for a small number of robots there is no significant difference in the collaborative task efficiency between an individually controlled multirobot team (where each robot is individually commanded) or a formation controlled team (where the group is commanded as a whole). TRAVERSE’s integrated multi-robot functionalities should maximise the collective performance of robot teams, while maintaining an intuitive, effortless and natural interaction with both human users and operators.
Keywords
Human-robot interaction, multi-robot teams, TRAVERSE, human-robot synergy, survivor search
