RangedIK: An Optimization-based Robot Motion Generation Method for Ranged-Goal Tasks

Exploiting Task Tolerances in Mimicry-based Telemanipulation

“…The numerical motion accuracy and smoothness matched the perceived predictability and fluency reported by the participants. Although prior work [4] has already shown that task tolerances can be exploited to generate high-quality motions, in this experiment, the robot was controlled by a user in real time, Fig. 3: Box and whisker plots of data from the user perception and performance measures for our experiment.…”

“…While semiconstrained motion planners can effectively generate motions for path following, they are not appropriate in timesensitive scenarios such as teleoperation. In this work, we employ RangedIK [4], which is a per-instant pose optimization method, to exploit task tolerances and generate robot motions in real time.…”

“…In this section, we describe how we use RangedIK [4] to exploit task tolerances to enable functional mimicry. To perform a task with tolerances, the DoFs of a robot's endeffector can be classified into three categories: (1) a DoF with zero tolerance requires the robot to accurately match a specific goal, e.g., a writing task requires accurate pen tip positions; (2) a DoF with unbounded tolerance provides the robot a range of equally valid goals, such as allowing a pen to rotate about its principal axis; and (3) a DoF with bounded tolerance gives the robot a range of acceptable goals with a preference toward a specific goal, e.g., allowing a tilting pen but preferring it to follow user's commands.…”

“…The functional mimicry paradigm allows a robot to autonomously adjust within tolerances to generate more accurate, smooth, and feasible motions. In our previous work, we presented RangedIK [4] as a real-time motion generation method that exploits flexibility afforded by task tolerances. In this paper, we apply RangedIK in a mimicrybased telemanipulation system and investigate whether task performance and user experience will be improved given autonomous robot adjustments within task tolerances.…”