Collision detection, localization &amp; classification for industrial robots with joint torque sensors

Popov, Dmitry; Klimchik, Alexandr; Mavridis, Nikolaos

doi:10.1109/roman.2017.8172400

Cited by 46 publications

(28 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…A comparison of classical dynamic modeling (DM) with simple artificial neural network (ANN) modeling (and thresholding) is presented in [ 34 ]. There, the authors show that the respective threshold values applied to the torque error (in the range 1–2.5 Nm) lead to quite good performance when the torque calculated by the KUKA environment is used as a baseline.…”

Section: Problem Formulationmentioning

confidence: 99%

A Simple Neural Network for Collision Detection of Collaborative Robots

Czubenko

Kowalczuk

2021

Sensors

View full text Add to dashboard Cite

Due to the epidemic threat, more and more companies decide to automate their production lines. Given the lack of adequate security or space, in most cases, such companies cannot use classic production robots. The solution to this problem is the use of collaborative robots (cobots). However, the required equipment (force sensors) or alternative methods of detecting a threat to humans are usually quite expensive. The article presents the practical aspect of collision detection with the use of a simple neural architecture. A virtual force and torque sensor, implemented as a neural network, may be useful in a team of collaborative robots. Four different approaches are compared in this article: auto-regressive (AR), recurrent neural network (RNN), convolutional long short-term memory (CNN-LSTM) and mixed convolutional LSTM network (MC-LSTM). These architectures are analyzed at different levels of input regression (motor current, position, speed, control velocity). This sensor was tested on the original CURA6 robot prototype (Cooperative Universal Robotic Assistant 6) by Intema. The test results indicate that the MC-LSTM architecture is the most effective with the regression level set at 12 samples (at 24 Hz). The mean absolute prediction error obtained by the MC-LSTM architecture was approximately 22 Nm. The conducted external test (72 different signals with collisions) shows that the presented architecture can be used as a collision detector. The MC-LSTM collision detection f1 score with the optimal threshold was 0.85. A well-developed virtual sensor based on such a network can be used to detect various types of collisions of cobot or other mobile or stationary systems operating on the basis of human-machine interaction.

show abstract

Section: Problem Formulationmentioning

confidence: 99%

A Simple Neural Network for Collision Detection of Collaborative Robots

Czubenko

Kowalczuk

2021

Sensors

View full text Add to dashboard Cite

show abstract

“…There are many methods and techniques for robot collision checking or detection in the literature. One effective way of achieving collision avoidance is to utilize hardware, including impedance actuators [22] [23], force and torque sensors [24] [25], depth sensor and camera [26] [27]. There are two main classes of algorithms for collision detection, namely Discrete Collision Detection (DCD) and Continuous Collision Detection (CCD) [28].…”

Section: E Collision Checkingmentioning

confidence: 99%

A framework for multi-robot coverage analysis of large and complex structures

Dai

Hassan

Sun³

et al. 2021

J Intell Manuf

View full text Add to dashboard Cite

Coverage analysis is essential for many coverage tasks (e.g., robotic grit-blasting, painting and surface cleaning) performed by Autonomous Industrial Robots (AIRs). Coverage analysis enables (1) the performance evaluation (e.g., coverage rate and operation efficiency) of AIRs for a coverage task, and (2) the configuration design of a multi-AIR system (e.g., decision on the number of AIRs to be used). Multi-AIR coverage analysis of large and complex structures involves addressing various problems. Thus, a framework is presented in this paper that incorporates various modules (e.g., AIR reachability, AIR base placement, collision avoidance, and area partitioning and allocation) for appropriately addressing the associated problems. The modules within the framework provide the flexibility of utilizing different methods and algorithms, depending on the requirements of the target application. The framework is tested and validated by extensive analyses of 10 different scenarios with up-to 10 AIRs.

show abstract

“…Pozwala on na wyliczenie momentu obrotowego dla każdego silnika (w każdym stopniu swobody), co jest proporcjonalne do poboru prądu pomierzonego na silnikach. W przypadku kolizji różnica między wyznaczonym poborem prądu a pomierzonym jest znacząca, dzięki temu wirtualny czujnik może wykryć kolizję [12,14]. Sensor taki jest stosowany w cobotach Universal Robots.…”

Section: Systemy Bezpieczeństwa Cobotówunclassified

Safety System for an Industrial Cooperating Robot Based on Depth Cameras

Masłowski¹,

Czubenko²

2019

PAR

View full text Add to dashboard Cite

Collision detection, localization & classification for industrial robots with joint torque sensors

Cited by 46 publications

References 20 publications

A Simple Neural Network for Collision Detection of Collaborative Robots

A Simple Neural Network for Collision Detection of Collaborative Robots

A framework for multi-robot coverage analysis of large and complex structures

Safety System for an Industrial Cooperating Robot Based on Depth Cameras

Contact Info

Product

Resources

About