Pub Date : 2015-07-27DOI: 10.1109/ICAR.2015.7251526
I. Mariolis, Georgia Peleka, A. Kargakos, S. Malassiotis
Category and pose recognition of highly deformable objects is considered a challenging problem in computer vision and robotics. In this study, we investigate recognition and pose estimation of garments hanging from a single point, using a hierarchy of deep convolutional neural networks. The adopted framework contains two layers. The deep convolutional network of the first layer is used for classifying the garment to one of the predefined categories, whereas in the second layer a category specific deep convolutional network performs pose estimation. The method has been evaluated using both synthetic and real datasets of depth images and an actual robotic platform. Experiments demonstrate that the task at hand may be performed with sufficient accuracy, to allow application in several practical scenarios.
{"title":"Pose and category recognition of highly deformable objects using deep learning","authors":"I. Mariolis, Georgia Peleka, A. Kargakos, S. Malassiotis","doi":"10.1109/ICAR.2015.7251526","DOIUrl":"https://doi.org/10.1109/ICAR.2015.7251526","url":null,"abstract":"Category and pose recognition of highly deformable objects is considered a challenging problem in computer vision and robotics. In this study, we investigate recognition and pose estimation of garments hanging from a single point, using a hierarchy of deep convolutional neural networks. The adopted framework contains two layers. The deep convolutional network of the first layer is used for classifying the garment to one of the predefined categories, whereas in the second layer a category specific deep convolutional network performs pose estimation. The method has been evaluated using both synthetic and real datasets of depth images and an actual robotic platform. Experiments demonstrate that the task at hand may be performed with sufficient accuracy, to allow application in several practical scenarios.","PeriodicalId":432004,"journal":{"name":"2015 International Conference on Advanced Robotics (ICAR)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2015-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132753625","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 : 2015-07-27DOI: 10.1109/ICAR.2015.7251434
É. Dorilêo, N. Zemiti, P. Poignet
Thin and long (semi-rigid) needles are well known to bend during percutaneous insertions because of needle-tissue interactions. Robotized needle insertions have been proposed to improve the efficacy of Interventional Radiology (IR) procedures such as radiofrequency ablation (RFA) of kidney tumors. However, the success of treatments and diagnosis depends on accurate prediction of needle deflection. This work aims to demonstrate the feasibility of merging needle-tissue properties, tip asymmetry and needle tip position updates to assist needle placement. In this paper we proposed a needle-tissue interaction model that matches the observations of transversal and axial resultant forces acting in the system. Analysis of a slope parameter between needle and tissue provides online and offline needle deflections predictions. Online updates of the needle tip position allow adaptive corrections of the slope parameter. Moreover, promising results were observed while evaluating the model's performance under uncertainties conditions such as tissue deformation, tissue inhomogeneity, needle-tissue friction, topological changes of the tissue and other modeling approximations. The system is evaluated by experiments in soft (homogeneous) PVC and multilayer tissue phantoms. Experiment results of needle placement into soft tissues presented average error of 1.04 mm. Meanwhile, online corrections decreased the error of offline predictions of 25%. The system shows an encouraging ability to predict semi-rigid needle deflection during interactions with elastic medium.
{"title":"Needle deflection prediction using adaptive slope model","authors":"É. Dorilêo, N. Zemiti, P. Poignet","doi":"10.1109/ICAR.2015.7251434","DOIUrl":"https://doi.org/10.1109/ICAR.2015.7251434","url":null,"abstract":"Thin and long (semi-rigid) needles are well known to bend during percutaneous insertions because of needle-tissue interactions. Robotized needle insertions have been proposed to improve the efficacy of Interventional Radiology (IR) procedures such as radiofrequency ablation (RFA) of kidney tumors. However, the success of treatments and diagnosis depends on accurate prediction of needle deflection. This work aims to demonstrate the feasibility of merging needle-tissue properties, tip asymmetry and needle tip position updates to assist needle placement. In this paper we proposed a needle-tissue interaction model that matches the observations of transversal and axial resultant forces acting in the system. Analysis of a slope parameter between needle and tissue provides online and offline needle deflections predictions. Online updates of the needle tip position allow adaptive corrections of the slope parameter. Moreover, promising results were observed while evaluating the model's performance under uncertainties conditions such as tissue deformation, tissue inhomogeneity, needle-tissue friction, topological changes of the tissue and other modeling approximations. The system is evaluated by experiments in soft (homogeneous) PVC and multilayer tissue phantoms. Experiment results of needle placement into soft tissues presented average error of 1.04 mm. Meanwhile, online corrections decreased the error of offline predictions of 25%. The system shows an encouraging ability to predict semi-rigid needle deflection during interactions with elastic medium.","PeriodicalId":432004,"journal":{"name":"2015 International Conference on Advanced Robotics (ICAR)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2015-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117170960","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 : 2015-07-27DOI: 10.1109/ICAR.2015.7251478
H. Hamzacebi, Ö. Morgül
One of the most significant outcomes of bio-inspired robotics research studies is that simple spring-mass models can accurately represent legged locomotion with various sizes and morphologies and hence the legged robots. Specifically, the Spring-Loaded Inverted Pendulum (SLIP) model became a well-known tool among the biologists and robotics researchers due to its simplicity and explanatory nature. Nevertheless, SLIP model has non-integrable system dynamics, which prevents derivation of exact analytical solutions to its dynamics despite its simple nature. In this paper, we propose a torque-enhanced active SLIP (TA-SLIP) model to support partial feedback linearization on nonlinear dynamics of the SLIP model. A linear and rotary actuator is used in TA-SLIP model to inject or remove energy from the system both to support analytic solution of the system dynamics and to control the locomotion. We also investigate the stability of the TA-SLIP model and show that the proposed model increases the region of stability with respect to original SLIP model.
{"title":"Enlarging the region of stability using the torque-enhanced active SLIP model","authors":"H. Hamzacebi, Ö. Morgül","doi":"10.1109/ICAR.2015.7251478","DOIUrl":"https://doi.org/10.1109/ICAR.2015.7251478","url":null,"abstract":"One of the most significant outcomes of bio-inspired robotics research studies is that simple spring-mass models can accurately represent legged locomotion with various sizes and morphologies and hence the legged robots. Specifically, the Spring-Loaded Inverted Pendulum (SLIP) model became a well-known tool among the biologists and robotics researchers due to its simplicity and explanatory nature. Nevertheless, SLIP model has non-integrable system dynamics, which prevents derivation of exact analytical solutions to its dynamics despite its simple nature. In this paper, we propose a torque-enhanced active SLIP (TA-SLIP) model to support partial feedback linearization on nonlinear dynamics of the SLIP model. A linear and rotary actuator is used in TA-SLIP model to inject or remove energy from the system both to support analytic solution of the system dynamics and to control the locomotion. We also investigate the stability of the TA-SLIP model and show that the proposed model increases the region of stability with respect to original SLIP model.","PeriodicalId":432004,"journal":{"name":"2015 International Conference on Advanced Robotics (ICAR)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2015-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128672179","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 : 2015-07-27DOI: 10.1109/ICAR.2015.7251488
A. Pires, D. Macharet, L. Chaimowicz
Cooperative localization allows groups of robots to improve their overall localization by sharing position estimates within the team. In Swarm Robotics a large number of very simple agents is used to perform different types of tasks, however, this simplicity may have a direct impact on the estimated localization. In this work, we consider the use of a single robot (leader) with improved localization capability (e.g. GPS) which will be used to enhance the position estimates of the rest of the group. By using a potential function, we are able to place the leader in the region (near the center) that best benefits its position broadcasting and also to execute a coordinated and continuous movement of the entire group by controlling only this unity. Numerous trials in a simulated environment were executed, providing statistical examination of the final results.
{"title":"Exploring heterogeneity for cooperative localization in Swarm Robotics","authors":"A. Pires, D. Macharet, L. Chaimowicz","doi":"10.1109/ICAR.2015.7251488","DOIUrl":"https://doi.org/10.1109/ICAR.2015.7251488","url":null,"abstract":"Cooperative localization allows groups of robots to improve their overall localization by sharing position estimates within the team. In Swarm Robotics a large number of very simple agents is used to perform different types of tasks, however, this simplicity may have a direct impact on the estimated localization. In this work, we consider the use of a single robot (leader) with improved localization capability (e.g. GPS) which will be used to enhance the position estimates of the rest of the group. By using a potential function, we are able to place the leader in the region (near the center) that best benefits its position broadcasting and also to execute a coordinated and continuous movement of the entire group by controlling only this unity. Numerous trials in a simulated environment were executed, providing statistical examination of the final results.","PeriodicalId":432004,"journal":{"name":"2015 International Conference on Advanced Robotics (ICAR)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2015-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126407986","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 : 2015-07-27DOI: 10.1109/ICAR.2015.7251446
Sebastian Buck, Richard Hanten, Goran Huskić, G. Rauscher, Alina Kloss, Jan Leininger, E. Ruff, F. Widmaier, A. Zell
Contrary to controlled experiments in a laboratory, robotics competitions pose a real challenge to an intelligent autonomous system. To allow a good performance, robotic systems have to be very robust to unforeseen circumstances in a highly constrained time frame. In this paper we present the design, implementation and performance of the robotic system we developed for participation in the SICK robot day 2014, with which we achieved very good results, placing second, with the same maximum score as the first. The task of this competition was to alternately fetch and deliver objects in a simplified warehouse scenario. Participating robots had to be able to carry out different tasks, such as detecting filling and delivery stations, receiving and analysing bar-code labelled objects as well as navigating in an arena among three other robots.
{"title":"Conclusions from an object-delivery robotic competition: SICK robot day 2014","authors":"Sebastian Buck, Richard Hanten, Goran Huskić, G. Rauscher, Alina Kloss, Jan Leininger, E. Ruff, F. Widmaier, A. Zell","doi":"10.1109/ICAR.2015.7251446","DOIUrl":"https://doi.org/10.1109/ICAR.2015.7251446","url":null,"abstract":"Contrary to controlled experiments in a laboratory, robotics competitions pose a real challenge to an intelligent autonomous system. To allow a good performance, robotic systems have to be very robust to unforeseen circumstances in a highly constrained time frame. In this paper we present the design, implementation and performance of the robotic system we developed for participation in the SICK robot day 2014, with which we achieved very good results, placing second, with the same maximum score as the first. The task of this competition was to alternately fetch and deliver objects in a simplified warehouse scenario. Participating robots had to be able to carry out different tasks, such as detecting filling and delivery stations, receiving and analysing bar-code labelled objects as well as navigating in an arena among three other robots.","PeriodicalId":432004,"journal":{"name":"2015 International Conference on Advanced Robotics (ICAR)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2015-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122304337","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 : 2015-07-27DOI: 10.1109/ICAR.2015.7251508
C. McGinn, M. Cullinan, G. Walsh, Cian Donavan, Kevin Kelly
In robotics, a system architecture refers to the manner in which a robot's control components (including sensors, actuators, microcontrollers and computers) are connected and how information flows between them. This paper describes the development of a two-tiered hybrid architecture which draws close inspiration from concepts in both embodied artificial intelligence as well as social intelligence theory. The design requirements for the architecture are postulated and their inclusion justified. A prototype embodiment of the proposed architecture has been practically implemented on the “Robbie” robot and a detailed discussion of its implementation is presented. The low-level proprioceptive system is demonstrated by tracking the forward kinematics of the robot over time. Additionally, the speed with which sensor readings are transferred from low-level to high-level control components is quantified to convey the scalability of the system.
{"title":"Towards an embodied system-level architecture for mobile robots","authors":"C. McGinn, M. Cullinan, G. Walsh, Cian Donavan, Kevin Kelly","doi":"10.1109/ICAR.2015.7251508","DOIUrl":"https://doi.org/10.1109/ICAR.2015.7251508","url":null,"abstract":"In robotics, a system architecture refers to the manner in which a robot's control components (including sensors, actuators, microcontrollers and computers) are connected and how information flows between them. This paper describes the development of a two-tiered hybrid architecture which draws close inspiration from concepts in both embodied artificial intelligence as well as social intelligence theory. The design requirements for the architecture are postulated and their inclusion justified. A prototype embodiment of the proposed architecture has been practically implemented on the “Robbie” robot and a detailed discussion of its implementation is presented. The low-level proprioceptive system is demonstrated by tracking the forward kinematics of the robot over time. Additionally, the speed with which sensor readings are transferred from low-level to high-level control components is quantified to convey the scalability of the system.","PeriodicalId":432004,"journal":{"name":"2015 International Conference on Advanced Robotics (ICAR)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2015-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121456519","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 : 2015-07-27DOI: 10.1109/ICAR.2015.7251523
Barry Ridge, Emre Ugur, A. Ude
Recent work in robotics, particularly in the domains of object manipulation and affordance learning, has seen the development of action-grounded features, that is, object features that are defined dynamically with respect to manipulation actions. Rather than using pose-invariant features, as is often the case with object recognition, such features are grounded with respect to the manipulation of the object, for instance, by using shape features that describe the surface of an object relative to the push contact point and direction. In this paper we provide an experimental comparison between action-grounded features and non-grounded features in an object affordance classification setting. Using an experimental platform that gathers 3-D data from the Kinect RGB-D sensor, as well as push action trajectories from an electromagnetic tracking system, we provide experimental results that demonstrate the effectiveness of this action-grounded approach across a range of state-of-the-art classifiers.
{"title":"Comparison of action-grounded and non-action-grounded 3-D shape features for object affordance classification","authors":"Barry Ridge, Emre Ugur, A. Ude","doi":"10.1109/ICAR.2015.7251523","DOIUrl":"https://doi.org/10.1109/ICAR.2015.7251523","url":null,"abstract":"Recent work in robotics, particularly in the domains of object manipulation and affordance learning, has seen the development of action-grounded features, that is, object features that are defined dynamically with respect to manipulation actions. Rather than using pose-invariant features, as is often the case with object recognition, such features are grounded with respect to the manipulation of the object, for instance, by using shape features that describe the surface of an object relative to the push contact point and direction. In this paper we provide an experimental comparison between action-grounded features and non-grounded features in an object affordance classification setting. Using an experimental platform that gathers 3-D data from the Kinect RGB-D sensor, as well as push action trajectories from an electromagnetic tracking system, we provide experimental results that demonstrate the effectiveness of this action-grounded approach across a range of state-of-the-art classifiers.","PeriodicalId":432004,"journal":{"name":"2015 International Conference on Advanced Robotics (ICAR)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2015-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116550373","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 : 2015-07-27DOI: 10.1109/ICAR.2015.7251522
Dani Martínez, Eduard Clotet, M. Tresanchez, Javier Moreno, Juan Manuel Jiménez-Soto, R. Magrans, S. Marco, J. Palacín
This paper presents the preliminary characterization results of a custom wind tunnel for designed for performing experiments on locating a volatile gas source with a mobile robot. Such experiments require a previous characterization of the wind tunnel as well as the definition of the configurable agents which are present during the experiments. This paper presents the experimental data gathered from the real environments. This paper shows the behavior of the evolution and diffusion of the gas depending on the gas injection rate, the mobile robot position, and the wind force. The mobile robot is equipped with a LIDAR for self localization, with a photo ionization detector (PID) for gas measurement, and with an anemometer for wind measurement. This paper shows the results obtained in static and dynamic experiments.
{"title":"First characterization results obtained in a wind tunnel designed for indoor gas source detection","authors":"Dani Martínez, Eduard Clotet, M. Tresanchez, Javier Moreno, Juan Manuel Jiménez-Soto, R. Magrans, S. Marco, J. Palacín","doi":"10.1109/ICAR.2015.7251522","DOIUrl":"https://doi.org/10.1109/ICAR.2015.7251522","url":null,"abstract":"This paper presents the preliminary characterization results of a custom wind tunnel for designed for performing experiments on locating a volatile gas source with a mobile robot. Such experiments require a previous characterization of the wind tunnel as well as the definition of the configurable agents which are present during the experiments. This paper presents the experimental data gathered from the real environments. This paper shows the behavior of the evolution and diffusion of the gas depending on the gas injection rate, the mobile robot position, and the wind force. The mobile robot is equipped with a LIDAR for self localization, with a photo ionization detector (PID) for gas measurement, and with an anemometer for wind measurement. This paper shows the results obtained in static and dynamic experiments.","PeriodicalId":432004,"journal":{"name":"2015 International Conference on Advanced Robotics (ICAR)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2015-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127663664","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 : 2015-07-27DOI: 10.1109/ICAR.2015.7251483
B. Kaleci, Cagri Mete Senler, H. Dutagaci, O. Parlaktuna
In this paper, a probabilistic approach is proposed for semantic classification in indoor environments using laser range data. Robot locations in indoor environments are categorized into three broad classes as room, corridor, and door. K-means and Learning Vector Quantization (LVQ) methods are used to classify robot positions. Circular shifting is applied to render laser range data independent of robot pose. K-means or LVQ algorithms are used to determine data clusters and their centers. In K-means method, the cluster centers are modelled with the proposed probabilistic approach to consider the semantic class of robot location. On the other hand, LVQ method inherently provides semantic classes of the cluster centers. In order to improve the rate of classification success, Markov model is integrated into the proposed approach. Experiments are conducted to demonstrate the effectiveness of the proposed approach. The results indicate that K-means method successfully classifies rooms and corridors, but door classification success rate is not satisfactory. LVQ method improves door classification rate without decreasing the classification rate of corridor and room. Lastly, effectiveness of the Markov model is discussed.
{"title":"A probabilistic approach for semantic classification using laser range data in indoor environments","authors":"B. Kaleci, Cagri Mete Senler, H. Dutagaci, O. Parlaktuna","doi":"10.1109/ICAR.2015.7251483","DOIUrl":"https://doi.org/10.1109/ICAR.2015.7251483","url":null,"abstract":"In this paper, a probabilistic approach is proposed for semantic classification in indoor environments using laser range data. Robot locations in indoor environments are categorized into three broad classes as room, corridor, and door. K-means and Learning Vector Quantization (LVQ) methods are used to classify robot positions. Circular shifting is applied to render laser range data independent of robot pose. K-means or LVQ algorithms are used to determine data clusters and their centers. In K-means method, the cluster centers are modelled with the proposed probabilistic approach to consider the semantic class of robot location. On the other hand, LVQ method inherently provides semantic classes of the cluster centers. In order to improve the rate of classification success, Markov model is integrated into the proposed approach. Experiments are conducted to demonstrate the effectiveness of the proposed approach. The results indicate that K-means method successfully classifies rooms and corridors, but door classification success rate is not satisfactory. LVQ method improves door classification rate without decreasing the classification rate of corridor and room. Lastly, effectiveness of the Markov model is discussed.","PeriodicalId":432004,"journal":{"name":"2015 International Conference on Advanced Robotics (ICAR)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2015-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130354646","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 : 2015-07-27DOI: 10.1109/ICAR.2015.7251492
Xingyun He, Lei Shi
With the rapid advancements in robotic technologies, the increasing diversity of hardware and software in robot systems has a significant impact on the design of robot control systems. The structure of robots becomes more and more sophisticated with the growing of the number of receptors and effectors. Integrating receptors and effectors to agents in multi-agent robotic system in order to complete a set of tasks is an important problem demanding efficient solution in the robot control system design. We present a bin-packing algorithm for task allocation and a graph nodes consolidation approach for resource allocation. Our bin-packing algorithm can allocate the tasks to each agent to meet the constraints of the computational ability of agents and the execution time of tasks, while guaranteeing all tasks can be completed. The graph nodes consolidation algorithm allocates all the resources to agents while minimizing the number of connections between agents, leading to a communication-efficient system structure. The proposed algorithms have polynomial time complexity compared with constrained guess-check and brutal force methods for solving complex multi-agent resource allocation problems.
{"title":"Automatic aid for robot control system design","authors":"Xingyun He, Lei Shi","doi":"10.1109/ICAR.2015.7251492","DOIUrl":"https://doi.org/10.1109/ICAR.2015.7251492","url":null,"abstract":"With the rapid advancements in robotic technologies, the increasing diversity of hardware and software in robot systems has a significant impact on the design of robot control systems. The structure of robots becomes more and more sophisticated with the growing of the number of receptors and effectors. Integrating receptors and effectors to agents in multi-agent robotic system in order to complete a set of tasks is an important problem demanding efficient solution in the robot control system design. We present a bin-packing algorithm for task allocation and a graph nodes consolidation approach for resource allocation. Our bin-packing algorithm can allocate the tasks to each agent to meet the constraints of the computational ability of agents and the execution time of tasks, while guaranteeing all tasks can be completed. The graph nodes consolidation algorithm allocates all the resources to agents while minimizing the number of connections between agents, leading to a communication-efficient system structure. The proposed algorithms have polynomial time complexity compared with constrained guess-check and brutal force methods for solving complex multi-agent resource allocation problems.","PeriodicalId":432004,"journal":{"name":"2015 International Conference on Advanced Robotics (ICAR)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2015-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115580744","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：