Manipulator Collision Detection and Collided Link Identification Based on Neural Networks

Sharkawy, Abdel-Nasser; Koustoumpardis, Panagiotis Ν.; Aspragathos, Nikos Α.

doi:10.1007/978-3-030-00232-9_1

Cited by 26 publications

(51 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is found that the execution time of the trained PTSNN (3.0 μs) presented in ref. [17] is slightly lower than the one of the trained PSNN (4.3 µs). Therefore, the collision detection time obtained from the trained PSNN is slightly longer.…”

Section: Collision Detectionmentioning

confidence: 61%

“…In our previous papers, 16,17 a multilayer feedforward NN was presented for collisions detection applied in 1-DOF and 2-DOF manipulators, respectively. In the latter, 17 the dynamics coupling between the joints of the manipulator was considered and showed that one NN, that couples the inputs from two joints, can be trained for detecting the collisions and identifying the collided link. In both papers, the detection efficiency of the trained NNs was high and the generalization ability was presented too.…”

Section: Related Workmentioning

confidence: 99%

“…After trying many different weights' initializations and number of hidden neurons, it is found that, in case of using the PTSNN as presented in ref. [17], the best one is with 120 hidden neurons and 1000 iterations, which gives the lowest MSE (0.018244) and the appropriate collision thresholds (τ th1 = 1.5529 Nm and τ th2 = 0.8006 Nm). The training MSE of the PTSNN is shown in Fig.…”

Section: Nn Trainingmentioning

confidence: 99%

“…The trained PTSNN presented in ref. [17] is based on both the intrinsic joint positions and the torques sensors of the manipulator. Therefore, its application is restricted to the robots having joint torque sensors, or where the estimations of the joint torques are available.…”

Section: Comparison Between the Two Nn Architecturesmentioning

confidence: 99%

See 3 more Smart Citations

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

2019

Self Cite

View full text Add to dashboard Cite

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.

show abstract

Section: Collision Detectionmentioning

confidence: 61%

Section: Related Workmentioning

confidence: 99%

Section: Nn Trainingmentioning

confidence: 99%

Section: Comparison Between the Two Nn Architecturesmentioning

confidence: 99%

See 2 more Smart Citations

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

2019

Self Cite

View full text Add to dashboard Cite

show abstract

“…Modern collaborative and also some lightweight robots with multiple sensors are an exciting field where this interaction opens new possibilities in production, and many interesting research has been performed on them. [1][2][3][4][5][6][7][8] However, it is clear that not all robots in production should or can be collaborative. Furthermore, already installed, conventional industrial robots will be predominant work force for years to come.…”

Section: Introductionmentioning

confidence: 99%

Collision Detection on Industrial Robots in Repetitive Tasks Using Modified Dynamic Time Warping

Gordić¹,

Jovanović²

2019

Robotica

View full text Add to dashboard Cite

SUMMARYThis paper presents a non-model-based collision detection algorithm for robots without external sensors and with closed control architecture. A reference signal of repetitive motion is recorded from the robot operation. To detect collisions, the reference is compared with measurements from the robot. One of the key contributions is a novel approach to optimal matching of compared signals, which is ensured by the newly developed modified Dynamic Time Warping (mDTW) method presented in this paper. One of the main improvements of the mDTW is that it enables comparing a signal with the most similar section of the other signal. Partial matching also enables online application of time warping principles and reduces the time and computation resources needed to perform matching. In addition to mDTW, two complementary decision rules are developed to identify collisions. The first rule, based on the absolute difference between compared matched samples, uses statistically determined thresholds to perform rapid detection of unambiguous collisions. The second rule is based on Eigen values of the covariance matrix of matched samples, and it employs its higher sensitivity to detect collisions with lower intensity. Results from experimental validation of the proposed collision algorithm on two industrial robots are shown.

show abstract

Assessing Deep Learning Models for Human-Robot Collaboration Collision Detection in Industrial Environments

Silva

Barbosa

Ledebour

et al. 2020

Intelligent Systems

View full text Add to dashboard Cite

Manipulator Collision Detection and Collided Link Identification Based on Neural Networks

Cited by 26 publications

References 9 publications

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

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

Collision Detection on Industrial Robots in Repetitive Tasks Using Modified Dynamic Time Warping

Assessing Deep Learning Models for Human-Robot Collaboration Collision Detection in Industrial Environments

Contact Info

Product

Resources

About