Task frame estimation during model-based teleoperation for satellite servicing

Abstract

Model-based teleoperation is suitable for systems with large communication delay because the operator interacts with a model of the task while the remote robot uses sensorbased control to replicate that interaction on the physical system. When the model geometry is accurately known, it is only necessary to register it to the remote physical system. If registration errors can be detected during the task, it is possible to update the local task frame (as in model mediated teleoperation) or the remote task frame. This paper proposes the latter approach for the case of telerobotic satellite servicing where the remote (on-orbit) robot cuts the tape that secures the patch of insulation covering the satellite access panel. This task can be modeled as sliding a tool along a planar surface (or one that is locally planar). The remote task frame is used by a hybrid position/force controller to maintain contact with the planar surface. Registration errors, however, can affect the orientation of the cutting tool and cause cutting failures. Therefore, the registration is updated during the task by using position measurements under the effect of hybrid control. The contributions of this paper are in the application of this method to teleoperation, where it must handle user-specified motions, and in the experimental verification during cutting, where compliance of the environment must be estimated and taken into account.

Haptic Shared-Control Methods for Robotic Cutting under Nonholonomic Constraints

“…The authors propose a robotic control architecture that helps a human operator in handling long objects by imposing virtual nonholonomic constraints. More recently, Li and Kazanzides [21] proposed a shared-control architecture for cutting in satellite servicing scenarios under time delay. The task consisted of cutting a straight line in multi-layer insulation (MLI) blankets (a thermal insulation patch used to cover satellites).…”

“…While the approaches of [22] and [21] are promising, they mainly focus on treating the time delay in the system rather than the cutting task itself. In fact, they all consider rather simple cutting tasks (straight lines).…”

Teleoperation and Visualization Interfaces for Remote Intervention in Space

“…First, we developed a method to enable an operator to more accurately align the virtual plane by dynamically texturing the virtual plane with the purpose of adding visual distortion until the virtual plane is accurately aligned with the real plane, Leonard (2015) . Then, Li and Kazanzides (2016) reported the development of a method to update the orientation of the virtual plane based on sensor feedback during the task, as described in Section 2.5 .…”

“…In Section 2.4.1 , we observed that in the directions of position/velocity control, the measured force could be used to monitor the task performance. Here, we consider that in directions of force control, the measured position can be used to update the task model, Li and Kazanzides (2016) .…”

“…The goal of the technique proposed by Li and Kazanzides (2016) is to allow the remote robot to automatically align the cutter to the plane, during the cutting task, by estimating the two DOF rotation between z c and n p . The problem is illustrated in Figure 7B .…”