Shared autonomy for assisted mobile robot teleoperation by recognizing operator intention as contextual task

For robot applications in unknown or hazardous environments, such as search and rescue, it is difficult and stressful for human beings to merely simply teleoperate a mobile robot without assistance. Consequently, much research work focuses on the means to facilitate an efficient shared autonomy between human and robot in the field of robotics. In our previous work, a novel shared autonomy system has been proposed, which models the user intention as the contextual task the user is executing, and provides motion assistance based on the intention recognition. We described and instantiated a certain task by defining task feature based on the contextual information. We employed a recursive Bayesian filter to maintain belief of task being executed, which is implemented with a Gaussian Mixture Regression model. In this paper, the proposed approach is applied to the problem of operating a holonomic mobile robot in the context of two task types: doorway crossing and object inspection. We detail the algorithm used by the system for processing the environment information, which assists the calculation of the task features for corresponding task types. Finally, the system is evaluated within multiple scenarios in an indoor environment and shows a good performance.