Toward Immersive Teleoperation with Assistive Force Cues to Guide Users with Obstacle and Target Information

Kurzfassung

With advances in recent years, we now have several different possibilities for robot operations in complex environments. Scaling from direct teleoperation to fully au- tonomous planning and execution, the user may choose their level of involvement as called for by the task at hand. Direct teleoperation requires continuous user engagement, which can result in significant physical and mental workload. On the other hand, autonomous systems are frequently unreliable in unstructured or dynamic environments. Supervised autonomy bridges the gap between the two methods by automating some tasks while keeping the human in the loop and allowing them to intervene between autonomous stages. However, there is still a lack of access for intervention and course correction during each autonomous execution. The presented research introduces an assistive teleoperation approach on top of an autonomous system, where a human teleoperator can intervene, with a haptic user interface, in the robot's motion during autonomous task execution. Our proposed approach is divided into two stages. The first stage plans the entire trajectory using a motion planner. The second stage generates assistive force cues for the planned trajectory toward the target position while avoiding known obstacles. The robot is in full control of the task execution until the operator introduces conflicting inputs to the haptic device. This inherently allows the system to function autonomously till the operator notices an error and intervenes by providing haptic inputs.