Neural Network Design for Manipulator Collision Detection Based Only on the Joint Position Sensors

Abstract

SUMMARYIn this paper, a multilayer feedforward neural network (NN) is designed and trained, for human–robot collisions detection, using only the intrinsic joint position sensors of a manipulator. The topology of one NN is designed considering the coupled dynamics of the robot and trained, with and without external contacts, by Levenberg–Marquardt algorithm to detect unwanted collisions of the human operator with the manipulator and the link that is collided. The proposed approach could be applied to any industrial robot, where only the joint position signals are available. The designed NN is compared quantitatively and qualitatively with an NN, where both the intrinsic joint position and the torque sensors of the manipulator are used. The proposed method is evaluated experimentally with the KUKA LWR manipulator, which is considered as an example of the collaborative robots, using two of its joints in a planar horizontal motion. The results illustrate that the developed system is efficient and fast to detect the collisions and identify the collided link.

Principle of Neural Network and Its Main Types: Review

“…Multilayer feedforward NN was used with our research interests in robotics. For example, in [25], the MLFFNN was designed and trained for human-robot collisions detection during the collaboration between the robot and human. In this work, two NN architectures were proposed.…”

“…The second NN architecture in [25] was composed of two hidden layers. The main benefits from using two hidden layers are the two hidden layers can achieve better result than using a single layer [26], and also they give a better generalization capability than a single hidden layer in many cases [27].…”

Mechanism Configuration and Innovation Control System Design for Power Cable Line Mobile Maintenance Robot

“…The power line maintenance inspection robots can be mainly divided into three types, inspection only, maintenance only, and both inspection and maintenance; for example, Japan, Europe, and United States have studied live maintenance automation technology and developed a variety of automation devices for different kinds of operation tasks by using intelligent operation terminal tools relied on insulated bucket arm trucks. 1,2 The representative robots in this field include the optical fiber composite overhead ground mobile robot which is developed by Tokyo Electric Power Company, 3,4 the Phase series robot which is developed by Japan Kyushu Electric Power Company, [5][6][7] subsequently, the Manitoba University in Canada 8 has developed the tactile remote control manipulator, and Canada Quebec Hydropower Research Institute 9 has developed "LineScout robots" which can perform maintenance with obstacle functions. After years of development, more and more new technologies are being applied to the inspection, operation, and maintenance of HVTL through the LineScout platform.…”

“…After years of development, more and more new technologies are being applied to the inspection, operation, and maintenance of HVTL through the LineScout platform. The Korea Electric Power Research Institute 9 has developed an insulator detection robot for the 345 kV AC transmission line, the basic configuration of a bolt tightening robot has been developed by India Welol University of Technology, 11 the bolt tightening can be carried out through the double mobile arms, the working manipulator and the fixed arms have been designed as a whole, wherein one arm fixes the hexagonal bolt and the other completes the tightening operation by rotating the sleeve motor. Additionally, American Electric Power Research Association, 12 Spain, 13 Ukrainian Electric Power Research Institute and Wenicha Polytechnic University 14 have also developed robots for transmission lines live maintenance operations.…”

Operation Motion Planning and Principle Prototype Design of Four-Wheel-Driven Mobile Robot for High-Voltage Double-Split Transmission Lines

“…System Simulation Analysis [23,24] and shown in Figure 8(a). rough the integration of the electrical system, mechanical system, and software system, the principle mobile platform can be been formed, as shown in Figures 8(b) and 8(c).…”