Dynamic robot manipulation using visual tracking

Abstract

Abstmct -We consider the multiple applications of vision in the solution of dynamic robot hand positioning problems such as catching, hitting, and intercepting a moving object. We propose a framework for qualitative reasoning about such problems and then present a robust vision-based control strategy that uses a Kalman filter to track image features. In our formulation, a given positioning problem is transformed into one of constrained trajectory planning on a topological surface defined by the relationship between positions of the robot and observable image features. By considering only the qualitative properties of this surface we eliminate the need for precise calibration of the various components of the system.

Noncontact impedance control for redundant manipulators

“…In the impedance control method proposed in this paper, it is necessary to measure the accelerations of the joint angles and the LED [see (7) and (12)]. For digital differentiation, the low-pass first-order differentiation filter invented by Usui and Amidror [21] were used.…”

“…In other cases, vision-based control for robot manipulators has been actively exploited in recent years [5]- [7], where the robot is controlled according to the visual information on the task space. Based on the framework of vision-based control, Castano and Hutchinson [8] proposed the concept of visual compliance.…”

Planning Under Sensing Uncertainty

“…For versions that involve minimal sensing and no prior given map, see [75,87,151,159,196]. The problem of visually tracking a moving target both with [8,70,71,102,128,129] and without [59,79,161] obstacles is closely related to pursuit-evasion. For a survey of combinatorial algorithms for computing visibility information, see [137].…”

Long-term stealth navigation in a security zone where the movement of the invader is monitored

“…The tracking based strategy purposes to minimize the difference between the state (location and velocity) of the target and the state of the robot over each control period [1]. On the other hand, the APPE based strategy [1][2][3] predicts the target motion and plans the robot motion in order to block the predicted path of the target. A robot-target rendezvous point, where the robot captures the target, is selected out of the points on the predicted path of the target, where the robot can arrive earlier than or simultaneously with the target.…”