Pub Date : 2019-08-01DOI: 10.1109/COASE.2019.8843126
P. Rajendran, Shantanu Thakar, Satyandra K. Gupta
We present a bi-directional tree-search framework for point-to-point path planning for manipulators. By design, it integrates human assistance seamlessly. Our framework consists of six modules: tree selection, focus selection, node selection, target selection, extend selection and connection type selection. Each module consists of a set of interchangeable strategies. By exploiting interaction among these strategies and selecting appropriate strategies based on the contextual cues from the search state, our method computes high quality solutions in a variety of complex scenarios with a low failure rate. We compare our approach with popular methods in a set of very hard scenarios. Without human assistance, our approach reduces the failure rate drastically. With human assistance, our approach has a zero failure rate as well as high solution quality.
{"title":"User-Guided Path Planning for Redundant Manipulators in Highly Constrained Work Environments","authors":"P. Rajendran, Shantanu Thakar, Satyandra K. Gupta","doi":"10.1109/COASE.2019.8843126","DOIUrl":"https://doi.org/10.1109/COASE.2019.8843126","url":null,"abstract":"We present a bi-directional tree-search framework for point-to-point path planning for manipulators. By design, it integrates human assistance seamlessly. Our framework consists of six modules: tree selection, focus selection, node selection, target selection, extend selection and connection type selection. Each module consists of a set of interchangeable strategies. By exploiting interaction among these strategies and selecting appropriate strategies based on the contextual cues from the search state, our method computes high quality solutions in a variety of complex scenarios with a low failure rate. We compare our approach with popular methods in a set of very hard scenarios. Without human assistance, our approach reduces the failure rate drastically. With human assistance, our approach has a zero failure rate as well as high solution quality.","PeriodicalId":6695,"journal":{"name":"2019 IEEE 15th International Conference on Automation Science and Engineering (CASE)","volume":"27 1","pages":"1212-1217"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89759979","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-08-01DOI: 10.1109/COASE.2019.8843237
Samuel Doctolero, C. Macnab
Although many hybrid force-position controllers appear in the literature, the problem of touching and leaving a surface rarely gets addressed - many leave this as a practical matter for the engineers. If the force control results in inappropriate signals in free space then the designer must try to switch controllers at the surface, a solution that can introduce unwanted vibrations; note that stability problems can easily result with such a design in light of imperfect knowledge/measurement of where the surface actually lies and the reality of (possibly unmodelled) joint elasticity. In this work we propose an adaptive backstepping approach that guarantees Lyapunov stability when in contact with the surface and in free-space i.e. without switching, for both non-redundant and redundant manipulators. We develop the controls for a flexible-joint robot in order to demonstrate the guarantee of stability and the ability to avoid excessive vibrations even in the case of elasticity. The proposed controls use neural networks to estimate nonlinear terms and unmodelled dynamics. Simulations show the proposed method significantly outperforms a proportional-derivative hybrid force-position control.
{"title":"CMAC-Adaptive Force-Position Control of a Flexible-Joint Robot","authors":"Samuel Doctolero, C. Macnab","doi":"10.1109/COASE.2019.8843237","DOIUrl":"https://doi.org/10.1109/COASE.2019.8843237","url":null,"abstract":"Although many hybrid force-position controllers appear in the literature, the problem of touching and leaving a surface rarely gets addressed - many leave this as a practical matter for the engineers. If the force control results in inappropriate signals in free space then the designer must try to switch controllers at the surface, a solution that can introduce unwanted vibrations; note that stability problems can easily result with such a design in light of imperfect knowledge/measurement of where the surface actually lies and the reality of (possibly unmodelled) joint elasticity. In this work we propose an adaptive backstepping approach that guarantees Lyapunov stability when in contact with the surface and in free-space i.e. without switching, for both non-redundant and redundant manipulators. We develop the controls for a flexible-joint robot in order to demonstrate the guarantee of stability and the ability to avoid excessive vibrations even in the case of elasticity. The proposed controls use neural networks to estimate nonlinear terms and unmodelled dynamics. Simulations show the proposed method significantly outperforms a proportional-derivative hybrid force-position control.","PeriodicalId":6695,"journal":{"name":"2019 IEEE 15th International Conference on Automation Science and Engineering (CASE)","volume":"104 1","pages":"794-799"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87748080","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-08-01DOI: 10.1109/COASE.2019.8842860
H. Phalen, P. Vagdargi, Michael Pozin, S. Chakravarty, G. Chirikjian, I. Iordachita, R. Taylor
The treatment of malaria is a global health challenge that stands to benefit from the widespread introduction of a vaccine for the disease. A method has been developed to create a live organism vaccine using the sporozoites (SPZ) of the parasite Plasmodium falciparum (Pf), which are concentrated in the salivary glands of infected mosquitoes. Current manual dissection methods to obtain these PfSPZ are not optimally efficient for large-scale vaccine production. We demonstrate the automation of a key step in this production process, the picking and placing of mosquitoes from a staging apparatus into a dissection assembly. This unit test of a robotic mosquito pick-and-place system is performed using a custom-designed micro-gripper attached to a four degree of freedom (4-DOF) robot under the guidance of a computer vision system. Mosquitoes are autonomously grasped from a mesh platform and pulled to a pair of notched dissection blades to remove the head of the mosquito, allowing access to the salivary glands. Placement into these blades is adapted based on output from computer vision to accommodate for the unique anatomy and orientation of each grasped mosquito. In this pilot test of the system on 50 mosquitoes, we demonstrate a 100% grasping accuracy and a 90% accuracy in placing the mosquito with its neck within the blade notches such that the head can be removed. This is a promising result for this difficult and non-standard pick-and-place task. An analysis of the failure cases provides insights for improvements to be implemented as this robotic pick-and-place system is integrated into a larger automated mosquito dissection system under development.
{"title":"Mosquito Pick-and-Place: Automating a Key Step in PfSPZ-based Malaria Vaccine Production","authors":"H. Phalen, P. Vagdargi, Michael Pozin, S. Chakravarty, G. Chirikjian, I. Iordachita, R. Taylor","doi":"10.1109/COASE.2019.8842860","DOIUrl":"https://doi.org/10.1109/COASE.2019.8842860","url":null,"abstract":"The treatment of malaria is a global health challenge that stands to benefit from the widespread introduction of a vaccine for the disease. A method has been developed to create a live organism vaccine using the sporozoites (SPZ) of the parasite Plasmodium falciparum (Pf), which are concentrated in the salivary glands of infected mosquitoes. Current manual dissection methods to obtain these PfSPZ are not optimally efficient for large-scale vaccine production. We demonstrate the automation of a key step in this production process, the picking and placing of mosquitoes from a staging apparatus into a dissection assembly. This unit test of a robotic mosquito pick-and-place system is performed using a custom-designed micro-gripper attached to a four degree of freedom (4-DOF) robot under the guidance of a computer vision system. Mosquitoes are autonomously grasped from a mesh platform and pulled to a pair of notched dissection blades to remove the head of the mosquito, allowing access to the salivary glands. Placement into these blades is adapted based on output from computer vision to accommodate for the unique anatomy and orientation of each grasped mosquito. In this pilot test of the system on 50 mosquitoes, we demonstrate a 100% grasping accuracy and a 90% accuracy in placing the mosquito with its neck within the blade notches such that the head can be removed. This is a promising result for this difficult and non-standard pick-and-place task. An analysis of the failure cases provides insights for improvements to be implemented as this robotic pick-and-place system is integrated into a larger automated mosquito dissection system under development.","PeriodicalId":6695,"journal":{"name":"2019 IEEE 15th International Conference on Automation Science and Engineering (CASE)","volume":"121 1","pages":"12-17"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89137527","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-08-01DOI: 10.1109/COASE.2019.8843125
Jingze Dai, H. Yoshiuchi, Tomohiro Matsuda
Service robot is an emerging robotic application. To be approved as an assistant or replacement for human employee, the robots should be able to solve team-level task effectively and thereby achieve high productivity. With regarded to the point, work assignment mechanism for multiple robots is essential. This paper originally gives a novel definition of effectiveness of deploying service robots from the viewpoint of multi-robot management, and then proposes a pack of multi-robot work assignment methods to improve the effectiveness by making service robots more accessible to the people who want to or are willing to use service robots. Simulation results demonstrated an improvement in the effectiveness of service deployment especially for application scenarios where the possibility of robot users is geographically uneven or changes over time.
{"title":"Multi-Robot Work Assignment Methods for Effectiveness Improvement of Deploying Service Robots","authors":"Jingze Dai, H. Yoshiuchi, Tomohiro Matsuda","doi":"10.1109/COASE.2019.8843125","DOIUrl":"https://doi.org/10.1109/COASE.2019.8843125","url":null,"abstract":"Service robot is an emerging robotic application. To be approved as an assistant or replacement for human employee, the robots should be able to solve team-level task effectively and thereby achieve high productivity. With regarded to the point, work assignment mechanism for multiple robots is essential. This paper originally gives a novel definition of effectiveness of deploying service robots from the viewpoint of multi-robot management, and then proposes a pack of multi-robot work assignment methods to improve the effectiveness by making service robots more accessible to the people who want to or are willing to use service robots. Simulation results demonstrated an improvement in the effectiveness of service deployment especially for application scenarios where the possibility of robot users is geographically uneven or changes over time.","PeriodicalId":6695,"journal":{"name":"2019 IEEE 15th International Conference on Automation Science and Engineering (CASE)","volume":"34 8 1","pages":"1224-1229"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80086353","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-08-01DOI: 10.1109/COASE.2019.8842876
Frederik Hagelskjær, T. Savarimuthu, N. Krüger, A. Buch
This paper presents a method for high precision visual pose estimation along with a simple setup procedure. Robotics for industrial solutions is a rapidly growing field and these robots require very precise position information to perform manipulations. This is usually accomplished using e.g. fixtures or feeders, both expensive hardware solutions. To enable fast changes in production, more flexible solutions are required, one possibility being visual pose estimation. Although many current pose estimation algorithms show increased performance in terms of recognition rates on public datasets, they do not focus on actual applications, neither in setup complexity or high accuracy during object localization. In contrast, our method focuses on solving a number of specific pose estimation problems in a seamless manner with a simple setup procedure. Our method relies on a number of workcell constraints and employs a novel method for automatically finding stable object poses. In addition, we use an active rendering method for refining the estimated object poses, giving a very fine localization, suitable for robotic manipulation. Experiments with current state-of-the-art 2D algorithms and our method show an average improvement from 9 mm to 0.95 mm uncertainty. The method was also used by the winning team at the 2018 World Robot Summit Assembly Challenge.
{"title":"Using spatial constraints for fast set-up of precise pose estimation in an industrial setting","authors":"Frederik Hagelskjær, T. Savarimuthu, N. Krüger, A. Buch","doi":"10.1109/COASE.2019.8842876","DOIUrl":"https://doi.org/10.1109/COASE.2019.8842876","url":null,"abstract":"This paper presents a method for high precision visual pose estimation along with a simple setup procedure. Robotics for industrial solutions is a rapidly growing field and these robots require very precise position information to perform manipulations. This is usually accomplished using e.g. fixtures or feeders, both expensive hardware solutions. To enable fast changes in production, more flexible solutions are required, one possibility being visual pose estimation. Although many current pose estimation algorithms show increased performance in terms of recognition rates on public datasets, they do not focus on actual applications, neither in setup complexity or high accuracy during object localization. In contrast, our method focuses on solving a number of specific pose estimation problems in a seamless manner with a simple setup procedure. Our method relies on a number of workcell constraints and employs a novel method for automatically finding stable object poses. In addition, we use an active rendering method for refining the estimated object poses, giving a very fine localization, suitable for robotic manipulation. Experiments with current state-of-the-art 2D algorithms and our method show an average improvement from 9 mm to 0.95 mm uncertainty. The method was also used by the winning team at the 2018 World Robot Summit Assembly Challenge.","PeriodicalId":6695,"journal":{"name":"2019 IEEE 15th International Conference on Automation Science and Engineering (CASE)","volume":"144 1","pages":"1308-1314"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76530190","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-08-01DOI: 10.1109/COASE.2019.8843119
Eric N. Sihite, Daniel J. Yang, T. Bewley
Brushed DC motors are usually driven with PWM forcing in one of two modes: drive/brake or drive/coast. That is, at the low state of the PWM forcing profile, the motor driver will either “brake” the motor with its own back EMF, or allow the motor to “coast” (i.e., spin freely). Drive/brake motor drivers, which are by far the most common, may be represented by a Multilevel Four-Quadrant DC Chopper model, while drive/coast motor drivers may be represented by two independent Bipolar Two-Quadrant DC Chopper models. Conveniently, when averaged over the PWM duty cycle, drive/brake motor drivers are accurately modeled as linear systems over their entire operational range. On the other hand, drive/coast motor drivers, when averaged over the PWM duty cycle, exhibit significant nonlinear behaviors that are dependent on factors such as inductance, PWM frequency, and rotor speed. Though there are some existing partial derivations of drive/coast motor driver models, no comprehensive, experimentally-validated modeling approaches appropriate for feedback control applications over the full dynamic range of the motor could be readily found in the literature. In this paper, we derive a practical nonlinear model of a drive/coast motor driver, validate this model using a motor dynamometer, and demonstrate a real-time implementation of this model on a Mobile Inverted Pendulum (MIP) robot.
{"title":"Derivation of a new drive/coast motor driver model for real-time brushed DC motor control, and validation on a MIP robot","authors":"Eric N. Sihite, Daniel J. Yang, T. Bewley","doi":"10.1109/COASE.2019.8843119","DOIUrl":"https://doi.org/10.1109/COASE.2019.8843119","url":null,"abstract":"Brushed DC motors are usually driven with PWM forcing in one of two modes: drive/brake or drive/coast. That is, at the low state of the PWM forcing profile, the motor driver will either “brake” the motor with its own back EMF, or allow the motor to “coast” (i.e., spin freely). Drive/brake motor drivers, which are by far the most common, may be represented by a Multilevel Four-Quadrant DC Chopper model, while drive/coast motor drivers may be represented by two independent Bipolar Two-Quadrant DC Chopper models. Conveniently, when averaged over the PWM duty cycle, drive/brake motor drivers are accurately modeled as linear systems over their entire operational range. On the other hand, drive/coast motor drivers, when averaged over the PWM duty cycle, exhibit significant nonlinear behaviors that are dependent on factors such as inductance, PWM frequency, and rotor speed. Though there are some existing partial derivations of drive/coast motor driver models, no comprehensive, experimentally-validated modeling approaches appropriate for feedback control applications over the full dynamic range of the motor could be readily found in the literature. In this paper, we derive a practical nonlinear model of a drive/coast motor driver, validate this model using a motor dynamometer, and demonstrate a real-time implementation of this model on a Mobile Inverted Pendulum (MIP) robot.","PeriodicalId":6695,"journal":{"name":"2019 IEEE 15th International Conference on Automation Science and Engineering (CASE)","volume":"51 1","pages":"1099-1105"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81204113","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper proposed an enhanced method on the basis of product of exponentials (POE) formula for open chain robotic kinematic calibration. The advantages of the method mainly lay in two aspects: Above all, both the joint offsets and the link length errors can be distinctly identified using the decoupling matrix B to constrain redundant degrees of freedom of POE-based model. Moreover, the absolute positioning accuracy can be further improved by identifying the errors of the reduction ratio and the coupling ratio, namely the transmission errors. Through this model, the following parameters can be identified: 1) the joint offsets, 2) the link lengths, 3) the transformation from measurement coordinate system to base, 4) the transformation from flange coordinate system to tool, and 5) the errors of joint reduction ratio and coupling ratio. The experimental results on a 6R type and a SCARA type manipulators revealed that there were significant improvements in the positioning accuracy after the calibration process: the 6R manipulator and the SCARA manipulator were enhanced from 2.175 mm to 0.291 mm, from 7.078 mm to 0.203 mm, respectively.
{"title":"An Enhanced POE-Based Method with Identified Transmission Errors for Serial Robotic Kinematic Calibration","authors":"Chentao Mao, Zhang-wei Chen, Hongfei Zu, Xiang Zhang","doi":"10.1109/COASE.2019.8843105","DOIUrl":"https://doi.org/10.1109/COASE.2019.8843105","url":null,"abstract":"This paper proposed an enhanced method on the basis of product of exponentials (POE) formula for open chain robotic kinematic calibration. The advantages of the method mainly lay in two aspects: Above all, both the joint offsets and the link length errors can be distinctly identified using the decoupling matrix B to constrain redundant degrees of freedom of POE-based model. Moreover, the absolute positioning accuracy can be further improved by identifying the errors of the reduction ratio and the coupling ratio, namely the transmission errors. Through this model, the following parameters can be identified: 1) the joint offsets, 2) the link lengths, 3) the transformation from measurement coordinate system to base, 4) the transformation from flange coordinate system to tool, and 5) the errors of joint reduction ratio and coupling ratio. The experimental results on a 6R type and a SCARA type manipulators revealed that there were significant improvements in the positioning accuracy after the calibration process: the 6R manipulator and the SCARA manipulator were enhanced from 2.175 mm to 0.291 mm, from 7.078 mm to 0.203 mm, respectively.","PeriodicalId":6695,"journal":{"name":"2019 IEEE 15th International Conference on Automation Science and Engineering (CASE)","volume":"62 1","pages":"1568-1573"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83930679","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-08-01DOI: 10.1109/COASE.2019.8843150
Lingxuan Liu, Hongyu L. He, Leyuan Shi
This paper addresses a flow shop scheduling problem with a sum-of-process-times based learning effect. The objective is to find schedules that can minimize the maximum completion time. For constructing a solution framework, we propose a new random-sampling-based solution procedure called Bounds-based Nested Partition (BBNP). In order to enhance the effectiveness of BBNP, we develop a composite bound for guidance. Two heuristic algorithms are conducted with worst-case analysis as benchmarks. Numerical results show that the BBNP algorithm outperforms benchmark algorithms.
{"title":"A new solution approach for flow shop scheduling with an exponential time-dependent learning effect","authors":"Lingxuan Liu, Hongyu L. He, Leyuan Shi","doi":"10.1109/COASE.2019.8843150","DOIUrl":"https://doi.org/10.1109/COASE.2019.8843150","url":null,"abstract":"This paper addresses a flow shop scheduling problem with a sum-of-process-times based learning effect. The objective is to find schedules that can minimize the maximum completion time. For constructing a solution framework, we propose a new random-sampling-based solution procedure called Bounds-based Nested Partition (BBNP). In order to enhance the effectiveness of BBNP, we develop a composite bound for guidance. Two heuristic algorithms are conducted with worst-case analysis as benchmarks. Numerical results show that the BBNP algorithm outperforms benchmark algorithms.","PeriodicalId":6695,"journal":{"name":"2019 IEEE 15th International Conference on Automation Science and Engineering (CASE)","volume":"76 1","pages":"468-473"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73571017","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-08-01DOI: 10.1109/COASE.2019.8843020
Lorenz Halt, Fengjunjie Pan, Philipp Tenbrock, A. Pott, T. Seel
In industrial robotics, controller parameters for force control must be adjusted to the specific robot that performs a task and they must be re-adjusted when the same task is to be performed by another robot. We address this challenge by proposing a transferable force controller for contact establishment between robot and surface. The controller is implemented based on task frame formalism. The proposed controller is based on prescribed performance control (PPC) and does not rely on a dynamic model of the environment. Due to the inherent robustness of PPC, it can be used to ensure similar performance for the same task across different robots and environments. The proposed controller is validated experimentally in a simple contact establishment task performed by three different robots (Universal Robots UR5, Franka Emika Panda, Denso Wave VS087) and three different board materials providing different stiffness (steel, aluminum, PVC). The PPC is found to yield an up to two orders of magnitude smaller variance of closed-loop settling time across all robots and materials than a conventional impedance controller.
{"title":"A Transferable Force Controller based on Prescribed Performance for Contact Establishment in Robotic Assembly Tasks*","authors":"Lorenz Halt, Fengjunjie Pan, Philipp Tenbrock, A. Pott, T. Seel","doi":"10.1109/COASE.2019.8843020","DOIUrl":"https://doi.org/10.1109/COASE.2019.8843020","url":null,"abstract":"In industrial robotics, controller parameters for force control must be adjusted to the specific robot that performs a task and they must be re-adjusted when the same task is to be performed by another robot. We address this challenge by proposing a transferable force controller for contact establishment between robot and surface. The controller is implemented based on task frame formalism. The proposed controller is based on prescribed performance control (PPC) and does not rely on a dynamic model of the environment. Due to the inherent robustness of PPC, it can be used to ensure similar performance for the same task across different robots and environments. The proposed controller is validated experimentally in a simple contact establishment task performed by three different robots (Universal Robots UR5, Franka Emika Panda, Denso Wave VS087) and three different board materials providing different stiffness (steel, aluminum, PVC). The PPC is found to yield an up to two orders of magnitude smaller variance of closed-loop settling time across all robots and materials than a conventional impedance controller.","PeriodicalId":6695,"journal":{"name":"2019 IEEE 15th International Conference on Automation Science and Engineering (CASE)","volume":"34 1","pages":"830-835"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88824407","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-08-01DOI: 10.1109/COASE.2019.8842973
Yu-Ming Chang, C. G. Li, Yi-Feng Hong
Objects embedded in the environment, such as switches, control buttons, sockets, et al., are devices that need frequent operations. To devise manipulators to operate such devices automatically, we propose a visual-position control scheme that directly converts the visual coordinate detections to motor commands. We train ConvNets with rigid 3D coordinate information, which is obtained from a single basis image of the target object. Our proposed training data preparation frameworks automatically generate and organize the required structure of the training images for the network. The ConvNet’s superior image recognition capability results in high success rate in object detection and high precision in coordinate estimation. In our static experiments, in-range plane coordinate detection achieves an average success rate of 91% from various view-point directions; the depth coordinate detection achieves an average success rate of 86% based on an extended success range. In our dynamic experiments, a low-precision manipulator was used to press a down elevator call button and achieved an overall success rate of 98%. A high-precision manipulator was used for an object localization task and achieved a precision of ± 0.3 mm using a low-resolution camera.
{"title":"Real-Time Object Coordinate Detection and Manipulator Control Using Rigidly Trained Convolutional Neural Networks","authors":"Yu-Ming Chang, C. G. Li, Yi-Feng Hong","doi":"10.1109/COASE.2019.8842973","DOIUrl":"https://doi.org/10.1109/COASE.2019.8842973","url":null,"abstract":"Objects embedded in the environment, such as switches, control buttons, sockets, et al., are devices that need frequent operations. To devise manipulators to operate such devices automatically, we propose a visual-position control scheme that directly converts the visual coordinate detections to motor commands. We train ConvNets with rigid 3D coordinate information, which is obtained from a single basis image of the target object. Our proposed training data preparation frameworks automatically generate and organize the required structure of the training images for the network. The ConvNet’s superior image recognition capability results in high success rate in object detection and high precision in coordinate estimation. In our static experiments, in-range plane coordinate detection achieves an average success rate of 91% from various view-point directions; the depth coordinate detection achieves an average success rate of 86% based on an extended success range. In our dynamic experiments, a low-precision manipulator was used to press a down elevator call button and achieved an overall success rate of 98%. A high-precision manipulator was used for an object localization task and achieved a precision of ± 0.3 mm using a low-resolution camera.","PeriodicalId":6695,"journal":{"name":"2019 IEEE 15th International Conference on Automation Science and Engineering (CASE)","volume":"76 1","pages":"1347-1352"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90562840","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：