An implemented theory of mind to improve human-robot shared plans execution

Abstract

When a robot has to execute a shared plan with a human, a number of unexpected situations and contingencies can happen due, essentially, to human initiative. For instance, a temporary absence or inattention of the human can entail a partial, and potentially not sufficient, knowledge about the current situation. To ensure a successful and fluent execution of the shared plan the robot might need to detect such situations and be able to provide the information to its human partner about what he missed without being annoying or intrusive. To do so, we have developed a framework which allows a robot to estimate the other agents mental states not only about the environment but also about the state of goals, plans and actions and to take them into account when executing human-robot shared plans.

Functional advantages of an adaptive Theory of Mind for robotics: a review of current architectures

“…Similarly, a ToM model is integrated in the framework illustrated in Devin et al [12] which permits the estimation of both the state of the environment and human partners' internal states, in particular goals and plans, which are considered by the robot for successful human-robot shared plans performance. More specifically, this architecture responses to unexpected situations and ambiguity which may happen during human-robot interaction given humans' initiative.…”

“…Beliefs are generally included amongst humans' mental states [8] and their understanding and tracking is considered an essential requirement for successful human-robot interactions [12], especially during collaborative tasks. The realization that other people represent the world from their perspective, and thus may have similar or different beliefs, is a distinguishing feature of the ToM (or mentalistic) account [16].…”

“…Several papers integrated the ability of visual perspective taking in their architecture [7,12,5] to infer the mental states (thus beliefs) of agents. In fact, by enabling robots to put themselves in the agents' shoes and infer their sensorial access, a better recognition of mental states and increased performances in belief recognition tasks were observed.…”

“…However, as we have seen in the previous paragraph, this architecture is limited to the understanding of agents' intentions. Finally, [7,12] provide some advancements related to the preparation of a robot to social interactions. More specifically, in [7] robots take into account the human's potential unwillingness to be assisted from the robot, whereas in [12] robots only execute their actions when they are considered as needed and possible, thus rendering the humanrobot interactions more fluid and effortless.…”

Transferring Adaptive Theory of Mind to Social Robots: Insights from Developmental Psychology to Robotics

“…In fact, determining whether a human partner holds true beliefs about a situation is an essential requirement for successful human-robot interactions [20], especially during collaborative tasks. Previous studies introduced in robots the ability to assume the spatial perspective of the agent they were interacting with [19,21], which is a fundamental characteristic of human mentalizing. By enabling robots to put themselves in the agents' shoes and infer their sensorial access, they showed a better recognition of mental states and increased performances in belief recognition tasks.…”