Coordination of an Uncalibrated 3-D Visuo-Motor System Based on Multiple Self-Organizing Maps

Abstract

This paper presents a method to coordinate an uncalibrated visuo-motor system in a 3D space. In order to handle spaces occluded by obstacles, a 3-camera system and two related self-organizing maps (SOMs) are employed. The self-organizing maps are directly connected to the camera system and are trained to perform position control. Based on the visibility of targets given in the workspace, an appropriate map is adopted. The maps determine the joint angles of the manipulator which make the end effector reach the targets precisely, and make the manipulator take obstacle-free poses. The proposed learning method ensures that the manipulator moves smoothly and consistently in whole workspace even though we use two maps to control it. In our visuo-motor system, neither any priori knowledge about the manipulator nor the camera parameters is required. In addition, the system is robust to change in its geometry. Simulation results are presented to demonstrate the effectiveness of the proposed method and the robustness of the system.

Kinematic control of a redundant manipulator using an inverse-forward adaptive scheme with a KSOM based hint generator

“…This can be done by using either geometrical models [27,13] for the obstacle and the manipulator or by obtaining information from sensors like vision [28][29][30] or sonar [3]. Once this information is available, obstacles can be avoided during online operation by using several methods.…”

Visual servoing of redundant manipulator with Jacobian matrix estimation using self-organizing map

“…[15] demonstrates that, in case of manipulators with higher degrees of freedom, the above algorithm is capable of resolving redundancy by minimizing the variations of joint angles. Recently, Han et al [16] used KSOM to avoid obstacles with a multiple camera setup for a 4-DOF manipulator, which involves computation of the pseudo-inverse during learning phase.…”

“…In addition, the computation of the pseudo-inverse at analysis about the type of solution has not been studied. The experimental results of [16] show that the redundant manipulator is tested for end-effector collision avoidance only and would suffer from positioning inaccuracy due to the open loop mode of operation. Recently, Kumar et al [17], proposed a KSOM network with joint space sub-clustering which allows the learning of multiple solutions for each end-effector position.…”

Multiple Self-Organizing Maps for Control of a Redundant Manipulator with Multiple Cameras

“…Although stereo camera systems can provide 3D information and we have used such a system in our previous works, the system could not well deal with spaces occluded by obstacles. They introduced 3-cameras system to overcome the situation (Han et al, 2006). However, the result was limited.…”

“…The method can avoid local minima. For detail, please refer to our previous study (Han et al, 2006).…”