Publications
Only selected publications from 2014 onward are listed here. See older publications for a list of selected publications from 2013 and earlier.
2024
Jaquier N, Welle M C, Gams A, Yao K, Fichera B, Billard A, Ude A, Asfour T, Kragic D
Transfer learning in robotics: An upcoming breakthrough? A review of promises and challenges Journal Article
In: The International Journal of Robotics Research, vol. 0, no. 0, pp. 02783649241273565, 2024.
@article{Gams2024IJRR,
title = {Transfer learning in robotics: An upcoming breakthrough? A review of promises and challenges},
author = {Noémie Jaquier and Michael C Welle and Andrej Gams and Kunpeng Yao and Bernardo Fichera and Aude Billard and Aleš Ude and Tamim Asfour and Danica Kragic},
doi = {10.1177/02783649241273565},
year = {2024},
date = {2024-10-01},
journal = {The International Journal of Robotics Research},
volume = {0},
number = {0},
pages = {02783649241273565},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Kuster B, Simonič M, Ude A
Vision-Language Model Based Robot Action Prediction for E-Waste Disassembly Operations Proceedings Article
In: 40th Anniversary of the IEEE International Conference on Robotics and Automation (ICRA), pp. 539-541, 2024.
@inproceedings{kuster2024,
title = {Vision-Language Model Based Robot Action Prediction for E-Waste Disassembly Operations},
author = {Boris Kuster and Mihael Simonič and Aleš Ude},
year = {2024},
date = {2024-09-26},
booktitle = {40th Anniversary of the IEEE International Conference on Robotics and Automation (ICRA)},
pages = {539-541},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Simonič M, Kuster B, Mavsar M, Šavle G, Ruiz S, Bem M, Hrovat M M, Ude A
An application study on reconfigurable robotic workcells and policy adaptation for electronic waste recycling Proceedings Article
In: IEEE International Conference on Cybernetics and Intelligent Systems (CIS) and IEEE International Conference on Robotics, Automation and Mechatronics (RAM), pp. 180-186, Hangzhou, China, 2024.
@inproceedings{simonic24,
title = {An application study on reconfigurable robotic workcells and policy adaptation for electronic waste recycling},
author = {Mihael Simonič and Boris Kuster and Matija Mavsar and Gašper Šavle and Sebastian Ruiz and Martin Bem and Matevž Majcen Hrovat and Aleš Ude},
year = {2024},
date = {2024-08-20},
booktitle = {IEEE International Conference on Cybernetics and Intelligent Systems (CIS) and IEEE International Conference on Robotics, Automation and Mechatronics (RAM)},
pages = {180-186},
address = {Hangzhou, China},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Jafari-Tabrizi A, Gruber D P, Gams A
Exploiting image quality measure for automatic trajectory generation in robot-aided visual quality inspection Journal Article
In: The International Journal of Advanced Manufacturing Technology, pp. 1–17, 2024.
@article{jafari2024,
title = {Exploiting image quality measure for automatic trajectory generation in robot-aided visual quality inspection},
author = {Atae Jafari-Tabrizi and Dieter P Gruber and Andrej Gams},
doi = {10.1007/s00170-024-13609-5},
year = {2024},
date = {2024-01-01},
urldate = {2024-01-01},
journal = {The International Journal of Advanced Manufacturing Technology},
pages = {1–17},
publisher = {Springer doi = 10.1007/s00170-024-13609-5},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Mavsar M, Ridge B, Pahič R, Morimoto J, Ude A
Simulation-Aided Handover Prediction From Video Using Recurrent Image-to-Motion Networks Journal Article
In: IEEE Transactions on Neural Networks and Learning Systems, vol. 35, no. 1, pp. 494-506, 2024.
@article{mavsar2022,
title = {Simulation-Aided Handover Prediction From Video Using Recurrent Image-to-Motion Networks},
author = {Matija Mavsar and Barry Ridge and Rok Pahič and Jun Morimoto and Aleš Ude},
doi = {10.1109/TNNLS.2022.3175720},
year = {2024},
date = {2024-01-01},
journal = {IEEE Transactions on Neural Networks and Learning Systems},
volume = {35},
number = {1},
pages = {494-506},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Simonič M, Ude A, Nemec B
Hierarchical learning of robotic contact policies Journal Article
In: Robotics and Computer-Integrated Manufacturing, vol. 86, pp. 102657, 2024.
@article{Simonic2024,
title = {Hierarchical learning of robotic contact policies},
author = {Mihael Simonič and Aleš Ude and Bojan Nemec},
url = {https://www.sciencedirect.com/science/article/pii/S0736584523001321},
doi = {https://doi.org/10.1016/j.rcim.2023.102657},
issn = {0736-5845},
year = {2024},
date = {2024-01-01},
urldate = {2024-01-01},
journal = {Robotics and Computer-Integrated Manufacturing},
volume = {86},
pages = {102657},
abstract = {The paper addresses the issue of learning tasks where a robot maintains permanent contact with the environment. We propose a new methodology based on a hierarchical learning scheme coupled with task representation through directed graphs. These graphs are constituted of nodes and branches that correspond to the states and robotic actions, respectively. The upper level of the hierarchy essentially operates as a decision-making algorithm. It leverages reinforcement learning (RL) techniques to facilitate optimal decision-making. The actions are generated by a constraint-space following (CSF) controller that autonomously identifies feasible directions for motion. The controller generates robot motion by adjusting its stiffness in the direction defined by the Frenet–Serret frame, which is aligned with the robot path. The proposed framework was experimentally verified through a series of challenging robotic tasks such as maze learning, door opening, learning to shift the manual car gear, and learning car license plate light assembly by disassembly.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The paper addresses the issue of learning tasks where a robot maintains permanent contact with the environment. We propose a new methodology based on a hierarchical learning scheme coupled with task representation through directed graphs. These graphs are constituted of nodes and branches that correspond to the states and robotic actions, respectively. The upper level of the hierarchy essentially operates as a decision-making algorithm. It leverages reinforcement learning (RL) techniques to facilitate optimal decision-making. The actions are generated by a constraint-space following (CSF) controller that autonomously identifies feasible directions for motion. The controller generates robot motion by adjusting its stiffness in the direction defined by the Frenet–Serret frame, which is aligned with the robot path. The proposed framework was experimentally verified through a series of challenging robotic tasks such as maze learning, door opening, learning to shift the manual car gear, and learning car license plate light assembly by disassembly.
Savevska K, Ude A
Analysis of Cost Functions for Reinforcement Learning of Reaching Tasks in Humanoid Robots Journal Article
In: Applied Sciences, vol. 14, no. 1, 2024.
@article{app14010039,
title = {Analysis of Cost Functions for Reinforcement Learning of Reaching Tasks in Humanoid Robots},
author = {Kristina Savevska and Aleš Ude},
url = {https://www.mdpi.com/2076-3417/14/1/39},
doi = {10.3390/app14010039},
issn = {2076-3417},
year = {2024},
date = {2024-01-01},
journal = {Applied Sciences},
volume = {14},
number = {1},
abstract = {In this paper, we present a study on transferring human motions to a humanoid robot for stable and precise task execution. We employ a whole-body motion imitation system that considers the stability of the robot to generate a stable reproduction of the demonstrated motion. However, the initially acquired motions are usually suboptimal. To successfully perform the desired tasks, the transferred motions require refinement through reinforcement learning to accommodate the differences between the human demonstrator and the humanoid robot as well as task constraints. Our experimental evaluation investigates the impact of different cost function terms on the overall task performance. The findings indicate that the selection of an optimal combination of weights included in the cost function is of great importance for learning precise reaching motions that preserve both the robot’s postural balance and the human-like shape of the demonstrated motions. We verified our methodology in a simulated environment and through tests on a real humanoid robot, TALOS.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
In this paper, we present a study on transferring human motions to a humanoid robot for stable and precise task execution. We employ a whole-body motion imitation system that considers the stability of the robot to generate a stable reproduction of the demonstrated motion. However, the initially acquired motions are usually suboptimal. To successfully perform the desired tasks, the transferred motions require refinement through reinforcement learning to accommodate the differences between the human demonstrator and the humanoid robot as well as task constraints. Our experimental evaluation investigates the impact of different cost function terms on the overall task performance. The findings indicate that the selection of an optimal combination of weights included in the cost function is of great importance for learning precise reaching motions that preserve both the robot’s postural balance and the human-like shape of the demonstrated motions. We verified our methodology in a simulated environment and through tests on a real humanoid robot, TALOS.
Mavsar M, Morimoto J, Ude A
GAN-based semi-supervised training of LSTM nets for intention recognition in cooperative tasks Journal Article
In: IEEE Robotics and Automation Letters, vol. 9, no. 1, pp. 263–270, 2024.
@article{mavsar2024gan,
title = {GAN-based semi-supervised training of LSTM nets for intention recognition in cooperative tasks},
author = {Matija Mavsar and Jun Morimoto and Aleš Ude},
doi = {10.1109/LRA.2024.1234567},
year = {2024},
date = {2024-01-01},
urldate = {2024-01-01},
journal = {IEEE Robotics and Automation Letters},
volume = {9},
number = {1},
pages = {263–270},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Lončarević Z, Reberšek S, Šela S, Skvarč J, Ude A, Gams A
Adaptive Visual Quality Inspection Based on Defect Prediction From Production Parameters Journal Article
In: IEEE Access, vol. 12, pp. 93899-93910, 2024.
@article{10587254,
title = {Adaptive Visual Quality Inspection Based on Defect Prediction From Production Parameters},
author = {Zvezdan Lončarević and Simon Reberšek and Samo Šela and Jure Skvarč and Aleš Ude and Andrej Gams},
doi = {10.1109/ACCESS.2024.3424664},
year = {2024},
date = {2024-01-01},
journal = {IEEE Access},
volume = {12},
pages = {93899-93910},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Deniša M, Ude A
Improving Ergonomics of Collaborative Robot Workcells Using Passive Reconfigurable Fixtures Journal Article
In: IEEE Access, vol. 12, pp. 124871-124888, 2024.
@article{10654788,
title = {Improving Ergonomics of Collaborative Robot Workcells Using Passive Reconfigurable Fixtures},
author = {Miha Deniša and Aleš Ude},
doi = {10.1109/ACCESS.2024.3451975},
year = {2024},
date = {2024-01-01},
journal = {IEEE Access},
volume = {12},
pages = {124871-124888},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
2023
Razmjoo A, Brecelj T, Savevska K, Ude A, Petrič T, Calinon S
Learning Joint Space Reference Manifold for Reliable Physical Assistance Proceedings Article
In: 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 10412–10417, IEEE, 2023.
@inproceedings{Razmjoo2023,
title = {Learning Joint Space Reference Manifold for Reliable Physical Assistance},
author = {Amirreza Razmjoo and Tilen Brecelj and Kristina Savevska and Aleš Ude and Tadej Petrič and Sylvain Calinon},
url = {https://ieeexplore.ieee.org/document/10342173/},
doi = {10.1109/IROS55552.2023.10342173},
issn = {21530866},
year = {2023},
date = {2023-10-01},
booktitle = {2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
pages = {10412–10417},
publisher = {IEEE},
abstract = {This paper presents a study on the use of the Talos humanoid robot for performing assistive sit-to-stand or stand-to-sit tasks. In such tasks, the human exerts a large amount of force (100-200 N) within a very short time (2-8 s), posing significant challenges in terms of human unpredictability and robot stability control. To address these challenges, we propose an approach for finding a spatial reference for the robot, which allows the robot to move according to the force exerted by the human and control its stability during the task. Specifically, we focus on the problem of finding a 1D manifold for the robot, while assuming a simple controller to guide its movement on this manifold. To achieve this, we use a functional representation to parameterize the manifold and solve an optimization problem that takes into account the robot's stability and the unpredictability of human behavior. We demonstrate the effectiveness of our approach through simulations and experiments with the Talos robot, showing robustness and adaptability.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
This paper presents a study on the use of the Talos humanoid robot for performing assistive sit-to-stand or stand-to-sit tasks. In such tasks, the human exerts a large amount of force (100-200 N) within a very short time (2-8 s), posing significant challenges in terms of human unpredictability and robot stability control. To address these challenges, we propose an approach for finding a spatial reference for the robot, which allows the robot to move according to the force exerted by the human and control its stability during the task. Specifically, we focus on the problem of finding a 1D manifold for the robot, while assuming a simple controller to guide its movement on this manifold. To achieve this, we use a functional representation to parameterize the manifold and solve an optimization problem that takes into account the robot's stability and the unpredictability of human behavior. We demonstrate the effectiveness of our approach through simulations and experiments with the Talos robot, showing robustness and adaptability.
Nemec B, Hrovat M M, Simonič M, Shetty S, Calinon S, Ude A
Robust Execution of Assembly Policies Using a Pose Invariant Task Representation Proceedings Article
In: Proc. Intl Conf. on Ubiquitous Robots (UR), pp. 779–786, 2023.
@inproceedings{Nemec23,
title = {Robust Execution of Assembly Policies Using a Pose Invariant Task Representation},
author = {Bojan Nemec and Matevž Majcen Hrovat and Mihael Simonič and Suhan Shetty and Sylvain Calinon and Aleš Ude},
doi = {10.1109/UR57808.2023.10202430},
year = {2023},
date = {2023-06-25},
urldate = {2023-06-25},
booktitle = {Proc. Intl Conf. on Ubiquitous Robots (UR)},
pages = {779–786},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Deniša M, Ude A, Simonič M, Kaarlela T, Pitkäaho T, Pieskä S, Arents J, Judvaitis J, Ozols K, Raj L, Czmerk A, Dianatfar M, Latokartano J, Schmidt P A, Mauersberger A, Singer A, Arnarson H, Shu B, Dimosthenopoulos D, Karagiannis P, Ahonen T, Valjus V, Lanz M
Technology Modules Providing Solutions for Agile Manufacturing Journal Article
In: Machines, vol. 11, no. 9, 2023.
@article{machines11090877,
title = {Technology Modules Providing Solutions for Agile Manufacturing},
author = {Miha Deniša and Aleš Ude and Mihael Simonič and Tero Kaarlela and Tomi Pitkäaho and Sakari Pieskä and Janis Arents and Janis Judvaitis and Kaspars Ozols and Levente Raj and András Czmerk and Morteza Dianatfar and Jyrki Latokartano and Patrick Alexander Schmidt and Anton Mauersberger and Adrian Singer and Halldor Arnarson and Beibei Shu and Dimosthenis Dimosthenopoulos and Panagiotis Karagiannis and Teemu-Pekka Ahonen and Veikko Valjus and Minna Lanz},
url = {https://www.mdpi.com/2075-1702/11/9/877},
doi = {10.3390/machines11090877},
issn = {2075-1702},
year = {2023},
date = {2023-01-01},
journal = {Machines},
volume = {11},
number = {9},
abstract = {In this paper, we address the most pressing challenges faced by the manufacturing sector, particularly the manufacturing of small and medium-sized enterprises (SMEs), where the transition towards high-mix low-volume production and the availability of cost-effective solutions are crucial. To overcome these challenges, this paper presents 14 innovative solutions that can be utilized to support the introduction of agile manufacturing processes in SMEs. These solutions encompass a wide range of key technologies, including reconfigurable fixtures, low-cost automation for printed circuit board (PCB) assembly, computer-vision-based control, wireless sensor networks (WSNs) simulations, predictive maintenance based on Internet of Things (IoT), virtualization for operator training, intuitive robot programming using virtual reality (VR), autonomous trajectory generation, programming by demonstration for force-based tasks, on-line task allocation in human–robot collaboration (HRC), projector-based graphical user interface (GUI) for HRC, human safety in collaborative work cells, and integration of automated ground vehicles for intralogistics. All of these solutions were designed with the purpose of increasing agility in the manufacturing sector. They are designed to enable flexible and modular manufacturing systems that are easy to integrate and use while remaining cost-effective for SMEs. As such, they have a high potential to be implemented in the manufacturing industry. They can be used as standalone modules or combined to solve a more complicated task, and contribute to enhancing the agility, efficiency, and competitiveness of manufacturing companies. With their application tested in industrially relevant environments, the proposed solutions strive to ensure practical implementation and real-world impact. While this paper presents these solutions and gives an overview of their methodologies and evaluations, it does not go into their details. It provides summaries of comprehensive and multifaceted solutions to tackle the evolving needs and demands of the manufacturing sector, empowering SMEs to thrive in a dynamic and competitive market landscape.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
In this paper, we address the most pressing challenges faced by the manufacturing sector, particularly the manufacturing of small and medium-sized enterprises (SMEs), where the transition towards high-mix low-volume production and the availability of cost-effective solutions are crucial. To overcome these challenges, this paper presents 14 innovative solutions that can be utilized to support the introduction of agile manufacturing processes in SMEs. These solutions encompass a wide range of key technologies, including reconfigurable fixtures, low-cost automation for printed circuit board (PCB) assembly, computer-vision-based control, wireless sensor networks (WSNs) simulations, predictive maintenance based on Internet of Things (IoT), virtualization for operator training, intuitive robot programming using virtual reality (VR), autonomous trajectory generation, programming by demonstration for force-based tasks, on-line task allocation in human–robot collaboration (HRC), projector-based graphical user interface (GUI) for HRC, human safety in collaborative work cells, and integration of automated ground vehicles for intralogistics. All of these solutions were designed with the purpose of increasing agility in the manufacturing sector. They are designed to enable flexible and modular manufacturing systems that are easy to integrate and use while remaining cost-effective for SMEs. As such, they have a high potential to be implemented in the manufacturing industry. They can be used as standalone modules or combined to solve a more complicated task, and contribute to enhancing the agility, efficiency, and competitiveness of manufacturing companies. With their application tested in industrially relevant environments, the proposed solutions strive to ensure practical implementation and real-world impact. While this paper presents these solutions and gives an overview of their methodologies and evaluations, it does not go into their details. It provides summaries of comprehensive and multifaceted solutions to tackle the evolving needs and demands of the manufacturing sector, empowering SMEs to thrive in a dynamic and competitive market landscape.
Kuster B, Simonič M, Ude A
Adaptive Robotic Levering for Recycling Tasks Proceedings Article
In: Petrič, Tadej; Ude, Aleš; Žlajpah, Leon (Ed.): Advances in Service and Industrial Robotics, pp. 417–425, Springer Nature Switzerland, Cham, 2023.
@inproceedings{10.1007/978-3-031-32606-6_49,
title = {Adaptive Robotic Levering for Recycling Tasks},
author = {Boris Kuster and Mihael Simonič and Aleš Ude},
editor = {Tadej Petrič and Aleš Ude and Leon Žlajpah},
isbn = {978-3-031-32606-6},
year = {2023},
date = {2023-01-01},
booktitle = {Advances in Service and Industrial Robotics},
pages = {417–425},
publisher = {Springer Nature Switzerland},
address = {Cham},
abstract = {A common step in autonomous robotic disassembly (recycling) of electronics is levering, which allows the robot to apply greater forces when removing parts of the devices. In practical applications, the robot should be able to adapt a levering action to different device types without an operator specifically recording a trajectory for each device. A method to generalize the existing levering actions to new devices is thus needed. In this paper we present a parameterized algorithm for performing robotic levering using feedback-based control to determine contact points and a sinusoidal pattern to realize adaptive levering motion. The algorithm can deal with devices of different shapes. After the initial adaptation process, the subsequent executions of the learnt levering action can be sped up to improve performance.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
A common step in autonomous robotic disassembly (recycling) of electronics is levering, which allows the robot to apply greater forces when removing parts of the devices. In practical applications, the robot should be able to adapt a levering action to different device types without an operator specifically recording a trajectory for each device. A method to generalize the existing levering actions to new devices is thus needed. In this paper we present a parameterized algorithm for performing robotic levering using feedback-based control to determine contact points and a sinusoidal pattern to realize adaptive levering motion. The algorithm can deal with devices of different shapes. After the initial adaptation process, the subsequent executions of the learnt levering action can be sped up to improve performance.
2022
Simonič M, Hrovat M M, Džeroski S, Ude A, Nemec B
Determining Exception Context in Assembly Operations from Multimodal Data Journal Article
In: Sensors, vol. 22, no. 20, 2022.
@article{Simonic2022,
title = {Determining Exception Context in Assembly Operations from Multimodal Data},
author = {Mihael Simonič and Matevž Majcen Hrovat and Sašo Džeroski and Aleš Ude and Bojan Nemec},
doi = {10.3390/s22207962},
year = {2022},
date = {2022-01-01},
journal = {Sensors},
volume = {22},
number = {20},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Gams A, Petrič T, Nemec B, Ude A
Manipulation learning on humanoid robots Journal Article
In: Current Robotics Reports, vol. 3, no. 3, pp. 97–109, 2022.
@article{gams2022,
title = {Manipulation learning on humanoid robots},
author = {Andrej Gams and Tadej Petrič and Bojan Nemec and Aleš Ude},
year = {2022},
date = {2022-01-01},
journal = {Current Robotics Reports},
volume = {3},
number = {3},
pages = {97–109},
publisher = {Springer DOI = 10.1007/s43154-022-00082-9},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Dežman M, Asfour T, Ude A, Gams A
Mechanical design and friction modelling of a cable-driven upper-limb exoskeleton Journal Article
In: Mechanism and Machine Theory, vol. 171, pp. 104746, 2022.
@article{Dezman2022,
title = {Mechanical design and friction modelling of a cable-driven upper-limb exoskeleton},
author = {Miha Dežman and Tamim Asfour and Aleš Ude and Andrej Gams},
url = {https://www.sciencedirect.com/science/article/pii/S0094114X22000234},
doi = {https://doi.org/10.1016/j.mechmachtheory.2022.104746},
issn = {0094-114X},
year = {2022},
date = {2022-01-01},
journal = {Mechanism and Machine Theory},
volume = {171},
pages = {104746},
abstract = {This paper presents a lightweight and low-inertia cable-driven upper-limb exoskeleton powerful enough to meet the requirements for activities of daily living. It presents the mechanical design, kinematic structure,the underlying actuation system, sensors, other electronic components as well as the controller of the exoskeleton. The extensive effect of friction on cable-driven designs, such as the one presented in this paper, requires proper mathematical modelling for controller design. Thus, we propose a current actuator model that describes the relationship between the motor current, velocity, and external load. The model relies on an underlying Stribeck+Coulomb friction representation and an additional parameter that modifies its Coulomb friction representation with an offset to represent adhesion between a cable and sheath. The model has been validated based on experimental data collected with the exoskeleton. The results show that the proposed model better captures the non-linear behaviour of the exoskeleton's actuation system, increasing overall descriptive performance by 15%. However, adding the adhesion offset to extend the relation of static friction, does not improve the model.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
This paper presents a lightweight and low-inertia cable-driven upper-limb exoskeleton powerful enough to meet the requirements for activities of daily living. It presents the mechanical design, kinematic structure,the underlying actuation system, sensors, other electronic components as well as the controller of the exoskeleton. The extensive effect of friction on cable-driven designs, such as the one presented in this paper, requires proper mathematical modelling for controller design. Thus, we propose a current actuator model that describes the relationship between the motor current, velocity, and external load. The model relies on an underlying Stribeck+Coulomb friction representation and an additional parameter that modifies its Coulomb friction representation with an offset to represent adhesion between a cable and sheath. The model has been validated based on experimental data collected with the exoskeleton. The results show that the proposed model better captures the non-linear behaviour of the exoskeleton's actuation system, increasing overall descriptive performance by 15%. However, adding the adhesion offset to extend the relation of static friction, does not improve the model.
Nimac P, Krpič A, Batagelj B, Gams A
Pedestrian Traffic Light Control with Crosswalk FMCW Radar and Group Tracking Algorithm Journal Article
In: Sensors, vol. 22, no. 5, 2022.
@article{Nimac2022,
title = {Pedestrian Traffic Light Control with Crosswalk FMCW Radar and Group Tracking Algorithm},
author = {Peter Nimac and Andrej Krpič and Boštjan Batagelj and Andrej Gams},
url = {https://www.mdpi.com/1424-8220/22/5/1754},
doi = {10.3390/s22051754},
issn = {1424-8220},
year = {2022},
date = {2022-01-01},
journal = {Sensors},
volume = {22},
number = {5},
abstract = {The increased mobility requirements of modern lifestyles put more stress on existing traffic infrastructure, which causes reduced traffic flow, especially in peak traffic hours. This calls for new and advanced solutions in traffic flow regulation and management. One approach towards optimisation is a transition from static to dynamic traffic light intervals, especially in spots where pedestrian crossing cause stops in road traffic flow. In this paper, we propose a smart pedestrian traffic light triggering mechanism that uses a Frequency-modulated continuous-wave (FMCW) radar for pedestrian detection. Compared to, for example, camera-surveillance systems, radars have advantages in the ability to reliably detect pedestrians in low-visibility conditions and in maintaining privacy. Objects within a radar's detection range are represented in a point cloud structure, in which pedestrians form clusters where they lose all identifiable features. Pedestrian detection and tracking are completed with a group tracking (GTRACK) algorithm that we modified to run on an external processor and not integrated into the used FMCW radar itself. The proposed prototype has been tested in multiple scenarios, where we focused on removing the call button from a conventional pedestrian traffic light. The prototype responded correctly in practically all cases by triggering the change in traffic signalization only when pedestrians were standing in the pavement area directly in front of the zebra crossing.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The increased mobility requirements of modern lifestyles put more stress on existing traffic infrastructure, which causes reduced traffic flow, especially in peak traffic hours. This calls for new and advanced solutions in traffic flow regulation and management. One approach towards optimisation is a transition from static to dynamic traffic light intervals, especially in spots where pedestrian crossing cause stops in road traffic flow. In this paper, we propose a smart pedestrian traffic light triggering mechanism that uses a Frequency-modulated continuous-wave (FMCW) radar for pedestrian detection. Compared to, for example, camera-surveillance systems, radars have advantages in the ability to reliably detect pedestrians in low-visibility conditions and in maintaining privacy. Objects within a radar's detection range are represented in a point cloud structure, in which pedestrians form clusters where they lose all identifiable features. Pedestrian detection and tracking are completed with a group tracking (GTRACK) algorithm that we modified to run on an external processor and not integrated into the used FMCW radar itself. The proposed prototype has been tested in multiple scenarios, where we focused on removing the call button from a conventional pedestrian traffic light. The prototype responded correctly in practically all cases by triggering the change in traffic signalization only when pedestrians were standing in the pavement area directly in front of the zebra crossing.
Nemec B, Mavsar M, Simonič M, Hrovat M M, Škrabar J, Ude A
Integration of a reconfigurable robotic workcell for assembly operations in automotive industry Proceedings Article
In: 2022 IEEE/SICE International Symposium on System Integration (SII), pp. 778-783, 2022.
@inproceedings{nemec2022,
title = {Integration of a reconfigurable robotic workcell for assembly operations in automotive industry},
author = {Bojan Nemec and Matija Mavsar and Mihael Simonič and Matevž Majcen Hrovat and Jure Škrabar and Aleš Ude},
doi = {10.1109/SII52469.2022.9708896},
year = {2022},
date = {2022-01-01},
booktitle = {2022 IEEE/SICE International Symposium on System Integration (SII)},
pages = {778-783},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
2021
Simonič M, Petrič T, Ude A, Nemec B
Analysis of Methods for Incremental Policy Refinement by Kinesthetic Guidance Journal Article
In: Journal of Intelligent & Robotic Systems, vol. 102, no. 1, 2021.
@article{Simonic2021,
title = {Analysis of Methods for Incremental Policy Refinement by Kinesthetic Guidance},
author = {Mihael Simonič and Tadej Petrič and Aleš Ude and Bojan Nemec},
year = {2021},
date = {2021-01-01},
journal = {Journal of Intelligent & Robotic Systems},
volume = {102},
number = {1},
publisher = {Springer},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Simonič M, Pahič R, Gašpar T, Abdolshah S, Haddadin S, Catalano M G, Wörgötter F, Ude A
Modular ROS-based software architecture for reconfigurable, Industry 4.0 compatible robotic workcells Proceedings Article
In: 20th International Conference on Advanced Robotics (ICAR), pp. 44-51, IEEE Ljubljana, Slovenia, 2021.
@inproceedings{Simonic2021b,
title = {Modular ROS-based software architecture for reconfigurable, Industry 4.0 compatible robotic workcells},
author = {Mihael Simonič and Rok Pahič and Timotej Gašpar and Saeed Abdolshah and Sami Haddadin and Manuel G. Catalano and Florentin Wörgötter and Aleš Ude},
year = {2021},
date = {2021-01-01},
booktitle = {20th International Conference on Advanced Robotics (ICAR)},
pages = {44-51},
address = {Ljubljana, Slovenia},
organization = {IEEE},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Simonič M, Deniša M, Ude A, Nemec B
A New Phase Determination Algorithm for Iterative Learning of Human-Robot Collaboration Proceedings Article
In: 20th International Conference on Advanced Robotics (ICAR), pp. 480–485, IEEE Ljubljana, Slovenia, 2021.
@inproceedings{Simonic2021c,
title = {A New Phase Determination Algorithm for Iterative Learning of Human-Robot Collaboration},
author = {Mihael Simonič and Miha Deniša and Aleš Ude and Bojan Nemec},
year = {2021},
date = {2021-01-01},
booktitle = {20th International Conference on Advanced Robotics (ICAR)},
pages = {480–485},
address = {Ljubljana, Slovenia},
organization = {IEEE},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Lončarević Z, Pahič R, Ude A, Gams A
Generalization-Based Acquisition of Training Data for Motor Primitive Learning by Neural Networks Journal Article
In: Applied Sciences, vol. 11, no. 3, 2021.
@article{Loncarevic2021a,
title = {Generalization-Based Acquisition of Training Data for Motor Primitive Learning by Neural Networks},
author = {Zvezdan Lončarević and Rok Pahič and Aleš Ude and Andrej Gams},
url = {https://www.mdpi.com/2076-3417/11/3/1013},
doi = {10.3390/app11031013},
issn = {2076-3417},
year = {2021},
date = {2021-01-01},
journal = {Applied Sciences},
volume = {11},
number = {3},
abstract = {Autonomous robot learning in unstructured environments often faces the problem that the dimensionality of the search space is too large for practical applications. Dimensionality reduction techniques have been developed to address this problem and describe motor skills in low-dimensional latent spaces. Most of these techniques require the availability of a sufficiently large database of example task executions to compute the latent space. However, the generation of many example task executions on a real robot is tedious, and prone to errors and equipment failures. The main result of this paper is a new approach for efficient database gathering by performing a small number of task executions with a real robot and applying statistical generalization, e.g., Gaussian process regression, to generate more data. We have shown in our experiments that the data generated this way can be used for dimensionality reduction with autoencoder neural networks. The resulting latent spaces can be exploited to implement robot learning more efficiently. The proposed approach has been evaluated on the problem of robotic throwing at a target. Simulation and real-world results with a humanoid robot TALOS are provided. They confirm the effectiveness of generalization-based database acquisition and the efficiency of learning in a low-dimensional latent space.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Autonomous robot learning in unstructured environments often faces the problem that the dimensionality of the search space is too large for practical applications. Dimensionality reduction techniques have been developed to address this problem and describe motor skills in low-dimensional latent spaces. Most of these techniques require the availability of a sufficiently large database of example task executions to compute the latent space. However, the generation of many example task executions on a real robot is tedious, and prone to errors and equipment failures. The main result of this paper is a new approach for efficient database gathering by performing a small number of task executions with a real robot and applying statistical generalization, e.g., Gaussian process regression, to generate more data. We have shown in our experiments that the data generated this way can be used for dimensionality reduction with autoencoder neural networks. The resulting latent spaces can be exploited to implement robot learning more efficiently. The proposed approach has been evaluated on the problem of robotic throwing at a target. Simulation and real-world results with a humanoid robot TALOS are provided. They confirm the effectiveness of generalization-based database acquisition and the efficiency of learning in a low-dimensional latent space.
Gašpar T, Kovač I, Ude A
Optimal layout and reconfiguration of a fixturing system constructed from passive Stewart platforms Journal Article
In: Journal of Manufacturing Systems, vol. 60, pp. 226-238, 2021.
@article{Gaspar2021,
title = {Optimal layout and reconfiguration of a fixturing system constructed from passive Stewart platforms},
author = {Timotej Gašpar and Igor Kovač and Aleš Ude},
url = {https://www.sciencedirect.com/science/article/pii/S0278612521001242},
doi = {https://doi.org/10.1016/j.jmsy.2021.05.020},
issn = {0278-6125},
year = {2021},
date = {2021-01-01},
journal = {Journal of Manufacturing Systems},
volume = {60},
pages = {226-238},
abstract = {The distinguishing property of Reconfigurable Manufacturing Systems (RMS) is that they can rapidly and efficiently adapt to new production requirements, both in terms of their capacity and functionalities. For this type of systems to achieve the desired efficiency, it should be possible to easily and quickly setup and reconfigure all of their components. This includes fixturing jigs that are used to hold workpieces firmly in place to enable a robot to carry out the desired production processes. In this paper, we formulate a constrained nonlinear optimization problem that must be solved to determine an optimal layout of reconfigurable fixtures for a given set of workpieces. The optimization problem takes into account the kinematic limitations of the fixtures, which are built in shape of Sterwart platforms, and the characteristics of the workpieces that need to be fastened into the fixturing system. Experimental results are presented that demonstrate that the automatically computed fixturing system layouts satisfy different constraints typically imposed in production environments.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The distinguishing property of Reconfigurable Manufacturing Systems (RMS) is that they can rapidly and efficiently adapt to new production requirements, both in terms of their capacity and functionalities. For this type of systems to achieve the desired efficiency, it should be possible to easily and quickly setup and reconfigure all of their components. This includes fixturing jigs that are used to hold workpieces firmly in place to enable a robot to carry out the desired production processes. In this paper, we formulate a constrained nonlinear optimization problem that must be solved to determine an optimal layout of reconfigurable fixtures for a given set of workpieces. The optimization problem takes into account the kinematic limitations of the fixtures, which are built in shape of Sterwart platforms, and the characteristics of the workpieces that need to be fastened into the fixturing system. Experimental results are presented that demonstrate that the automatically computed fixturing system layouts satisfy different constraints typically imposed in production environments.
Radanovič P, Jereb J, Kovač I, Ude A
Design of a Modular Robotic Workcell Platform Enabled by Plug & Produce Connectors Proceedings Article
In: 2021 20th International Conference on Advanced Robotics (ICAR), pp. 304-309, 2021.
@inproceedings{Radanovic2020,
title = {Design of a Modular Robotic Workcell Platform Enabled by Plug & Produce Connectors},
author = {Primož Radanovič and Jaka Jereb and Igor Kovač and Aleš Ude},
doi = {10.1109/ICAR53236.2021.9659345},
year = {2021},
date = {2021-01-01},
booktitle = {2021 20th International Conference on Advanced Robotics (ICAR)},
pages = {304-309},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Mavsar M, Deniša M, Nemec B, Ude A
Intention Recognition with Recurrent Neural Networks for Dynamic Human-Robot Collaboration Proceedings Article
In: 2021 20th International Conference on Advanced Robotics (ICAR), pp. 208-215, 2021.
@inproceedings{9659473,
title = {Intention Recognition with Recurrent Neural Networks for Dynamic Human-Robot Collaboration},
author = {Matija Mavsar and Miha Deniša and Bojan Nemec and Aleš Ude},
doi = {10.1109/ICAR53236.2021.9659473},
year = {2021},
date = {2021-01-01},
booktitle = {2021 20th International Conference on Advanced Robotics (ICAR)},
pages = {208-215},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Deniša M, Schwaner K L, Iturrate I, Savarimuthu T R
Semi-Autonomous Cooperative Tasks in a Multi-Arm Robotic Surgical Domain Proceedings Article
In: 2021 20th International Conference on Advanced Robotics (ICAR), pp. 134-141, 2021.
@inproceedings{9659445,
title = {Semi-Autonomous Cooperative Tasks in a Multi-Arm Robotic Surgical Domain},
author = {Miha Deniša and Kim Lindberg Schwaner and Iñigo Iturrate and Thiusius Rajeeth Savarimuthu},
doi = {10.1109/ICAR53236.2021.9659445},
year = {2021},
date = {2021-01-01},
booktitle = {2021 20th International Conference on Advanced Robotics (ICAR)},
pages = {134-141},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Tirmizi A, Dianatfar M, Lanz M, Pieters R, Pitkäaho T, Kaarlela T, Pieskä S, Kousi N, Zoga A, Valjus V, Deniša M, Ude A
Technical Maturity for Industrial Deployment of Robot Demonstrators Proceedings Article
In: 2021 20th International Conference on Advanced Robotics (ICAR), pp. 310-317, 2021.
@inproceedings{9659436,
title = {Technical Maturity for Industrial Deployment of Robot Demonstrators},
author = {Asad Tirmizi and Morteza Dianatfar and Minna Lanz and Roel Pieters and Tomi Pitkäaho and Tero Kaarlela and Sakari Pieskä and Niki Kousi and Adamantia Zoga and Veikko Valjus and Miha Deniša and Aleš Ude},
doi = {10.1109/ICAR53236.2021.9659436},
year = {2021},
date = {2021-01-01},
booktitle = {2021 20th International Conference on Advanced Robotics (ICAR)},
pages = {310-317},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Pahič R, Lončarević Z, Gams A, Ude A
Robot skill learning in latent space of a deep autoencoder neural network Journal Article
In: Robotics and Autonomous Systems, vol. 135, no. 1, pp. 103690, 2021.
@article{Pahic2021,
title = {Robot skill learning in latent space of a deep autoencoder neural network},
author = {Rok Pahič and Zvezdan Lončarević and Andrej Gams and Aleš Ude},
url = {https://www.sciencedirect.com/science/article/pii/S0921889020305303},
doi = {https://doi.org/10.1016/j.robot.2020.103690},
issn = {0921-8890},
year = {2021},
date = {2021-01-01},
journal = {Robotics and Autonomous Systems},
volume = {135},
number = {1},
pages = {103690},
abstract = {Just like humans, robots can improve their performance by practicing, i.e. by performing the desired behavior many times and updating the underlying skill representation using the newly gathered data. In this paper, we propose to implement robot practicing by applying statistical and reinforcement learning (RL) in a latent space of the selected skill representation. The latent space is computed by a deep autoencoder neural network, with the data to train the network generated in simulation. However, we show that the resulting latent space representation is useful also for learning on a real robot. Our simulation and real-world results demonstrate that by exploiting the latent space of the underlying motor skill representation, a significant reduction of the amount of data needed for effective learning by Gaussian Process Regression (GPR) can be achieved. Similarly, the number of RL epochs can be significantly reduced. Finally, it is evident from our results that an autoencoder-based latent space is more effective for these purposes than a latent space computed by principal component analysis.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Just like humans, robots can improve their performance by practicing, i.e. by performing the desired behavior many times and updating the underlying skill representation using the newly gathered data. In this paper, we propose to implement robot practicing by applying statistical and reinforcement learning (RL) in a latent space of the selected skill representation. The latent space is computed by a deep autoencoder neural network, with the data to train the network generated in simulation. However, we show that the resulting latent space representation is useful also for learning on a real robot. Our simulation and real-world results demonstrate that by exploiting the latent space of the underlying motor skill representation, a significant reduction of the amount of data needed for effective learning by Gaussian Process Regression (GPR) can be achieved. Similarly, the number of RL epochs can be significantly reduced. Finally, it is evident from our results that an autoencoder-based latent space is more effective for these purposes than a latent space computed by principal component analysis.
Lanz M, Reimann J, Ude A, Kousi N, Pieters R, Dianatfar M, Makris S
Digital innovation hubs for robotics–TRINITY approach for distributing knowledge via modular use case demonstrations Journal Article
In: Procedia CIRP, vol. 97, pp. 45–50, 2021.
@article{Lanz2021,
title = {Digital innovation hubs for robotics–TRINITY approach for distributing knowledge via modular use case demonstrations},
author = {Minna Lanz and Jan Reimann and Aleš Ude and Niki Kousi and Roel Pieters and Morteza Dianatfar and Sotiris Makris},
year = {2021},
date = {2021-01-01},
journal = {Procedia CIRP},
volume = {97},
pages = {45–50},
publisher = {Elsevier},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Lončarević Z, Ude A, Gams A
Accelerated Robot Skill Acquisition by Reinforcement Learning-Aided Sim-to-Real Domain Adaptation Proceedings Article
In: 2021 20th International Conference on Advanced Robotics (ICAR), pp. 269-274, 2021.
@inproceedings{9659351,
title = {Accelerated Robot Skill Acquisition by Reinforcement Learning-Aided Sim-to-Real Domain Adaptation},
author = {Zvezdan Lončarević and Aleš Ude and Andrej Gams},
doi = {10.1109/ICAR53236.2021.9659351},
year = {2021},
date = {2021-01-01},
booktitle = {2021 20th International Conference on Advanced Robotics (ICAR)},
pages = {269-274},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Lončarević Z, Gams A, Reberšek S, Nemec B, Škrabar J, Skvarč J, Ude A
Specifying and optimizing robotic motion for visual quality inspection Journal Article
In: Robotics and Computer-Integrated Manufacturing, vol. 72, pp. 102200, 2021.
@article{Loncarevic2021,
title = {Specifying and optimizing robotic motion for visual quality inspection},
author = {Zvezdan Lončarević and Andrej Gams and Simon Reberšek and Bojan Nemec and Jure Škrabar and Jure Skvarč and Aleš Ude},
url = {https://www.sciencedirect.com/science/article/pii/S0736584521000831},
doi = {https://doi.org/10.1016/j.rcim.2021.102200},
issn = {0736-5845},
year = {2021},
date = {2021-01-01},
journal = {Robotics and Computer-Integrated Manufacturing},
volume = {72},
pages = {102200},
abstract = {Installation or even just modification of robot-supported production and quality inspection is a tedious process that usually requires full-time human expert engagement. The resulting parameters, e.g. robot velocities specified by an expert, are often subjective and produce suboptimal results. In this paper, we propose a new approach for specifying visual inspection trajectories based on CAD models of workpieces to be inspected. The expert involvement is required only to select â in a CAD system â the desired points on the inspection path along which the robot should move the camera. The rest of the approach is fully automatic. From the selected path data, the system computes temporal parametrization of the path, which ensures smoothness of the resulting robot trajectory for visual inspection. We then apply a new learning method for the optimization of robot speed along the specified path. The proposed approach combines iterative learning control and reinforcement learning. It takes a numerical estimate of image quality as input and produces the fastest possible motion that does not result in the degradation of image quality as output. In our experiments, the algorithm achieved up to 53% cycle time reduction from an initial, manually specified motion, without degrading the image quality. We show experimentally that the proposed algorithm achieves better results compared to some other policy learning approaches. The described approach is general and can be used with different types of learning and feedback signals.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Installation or even just modification of robot-supported production and quality inspection is a tedious process that usually requires full-time human expert engagement. The resulting parameters, e.g. robot velocities specified by an expert, are often subjective and produce suboptimal results. In this paper, we propose a new approach for specifying visual inspection trajectories based on CAD models of workpieces to be inspected. The expert involvement is required only to select â in a CAD system â the desired points on the inspection path along which the robot should move the camera. The rest of the approach is fully automatic. From the selected path data, the system computes temporal parametrization of the path, which ensures smoothness of the resulting robot trajectory for visual inspection. We then apply a new learning method for the optimization of robot speed along the specified path. The proposed approach combines iterative learning control and reinforcement learning. It takes a numerical estimate of image quality as input and produces the fastest possible motion that does not result in the degradation of image quality as output. In our experiments, the algorithm achieved up to 53% cycle time reduction from an initial, manually specified motion, without degrading the image quality. We show experimentally that the proposed algorithm achieves better results compared to some other policy learning approaches. The described approach is general and can be used with different types of learning and feedback signals.
Pahič R, Gams A, Ude A
Reconstructing Spatial Aspects of Motion by Image-to-Path Deep Neural Networks Journal Article
In: IEEE Robotics and Automation Letters, vol. 6, no. 1, pp. 255-262, 2021.
@article{Pahic2021a,
title = {Reconstructing Spatial Aspects of Motion by Image-to-Path Deep Neural Networks},
author = {Rok Pahič and Andrej Gams and Aleš Ude},
doi = {10.1109/LRA.2020.3039937},
year = {2021},
date = {2021-01-01},
journal = {IEEE Robotics and Automation Letters},
volume = {6},
number = {1},
pages = {255-262},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Nemec B, Yasuda K, Ude A
A Virtual Mechanism Approach for Exploiting Functional Redundancy in Finishing Operations Journal Article
In: IEEE Transactions on Automation Science and Engineering, vol. 18, no. 4, pp. 2048-2060, 2021.
@article{9246671,
title = {A Virtual Mechanism Approach for Exploiting Functional Redundancy in Finishing Operations},
author = {Bojan Nemec and Kenichi Yasuda and Aleš Ude},
doi = {10.1109/TASE.2020.3032075},
year = {2021},
date = {2021-01-01},
journal = {IEEE Transactions on Automation Science and Engineering},
volume = {18},
number = {4},
pages = {2048-2060},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
2020
Pahič R, Ridge B, Gams A, Morimoto J, Ude A
Training of deep neural networks for the generation of dynamic movement primitives Journal Article
In: Neural Networks, vol. 127, pp. 121-131, 2020.
@article{pahic2020,
title = {Training of deep neural networks for the generation of dynamic movement primitives},
author = {Rok Pahič and Barry Ridge and Andrej Gams and Jun Morimoto and Aleš Ude},
doi = {10.1016/j.neunet.2020.04.010},
year = {2020},
date = {2020-01-01},
journal = {Neural Networks},
volume = {127},
pages = {121-131},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Gašpar T, Deniša M, Radanovič P, Ridge B, Savarimuthu T R, Kramberger A, Priggemeyer M, Rossmann J, Wörgötter F, Ivanovska T, Parizi S, Gosar Ž, Kovač I, Ude A
Smart hardware integration with advanced robot programming technologies for efficient reconfiguration of robot workcells Journal Article
In: Robotics and Computer-Integrated Manufacturing, vol. 66, pp. 101979, 2020.
@article{Gaspar2020,
title = {Smart hardware integration with advanced robot programming technologies for efficient reconfiguration of robot workcells},
author = {Timotej Gašpar and Miha Deniša and Primož Radanovič and Barry Ridge and T. Rajeeth Savarimuthu and Aljaž Kramberger and Marc Priggemeyer and J:urgen Rossmann and Florentin Wörgötter and Tatyana Ivanovska and Shahab Parizi and Žiga Gosar and Igor Kovač and Aleš Ude},
url = {https://www.sciencedirect.com/science/article/pii/S0736584519306726},
doi = {https://doi.org/10.1016/j.rcim.2020.101979},
issn = {0736-5845},
year = {2020},
date = {2020-01-01},
journal = {Robotics and Computer-Integrated Manufacturing},
volume = {66},
pages = {101979},
abstract = {The manufacturing industry is seeing an increase in demand for more custom-made, low-volume production. This type of production is rarely automated and is to a large extent still performed manually. To keep up with the competition and market demands, manufacturers will have to undertake the effort to automate such manufacturing processes. However, automating low-volume production is no small feat as the solution should be adaptable and future proof to unexpected changes in customer's demands. In this paper, we propose a reconfigurable robot workcell aimed at automating low-volume production. The developed workcell can adapt to the changes in manufacturing processes by employing a number of passive, reconfigurable hardware elements, supported by the ROS-based, modular control software. To further facilitate and expedite the setup process, we integrated intuitive, user-friendly robot programming methods with the available hardware. The system was evaluated by implementing five production processes from different manufacturing industries.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The manufacturing industry is seeing an increase in demand for more custom-made, low-volume production. This type of production is rarely automated and is to a large extent still performed manually. To keep up with the competition and market demands, manufacturers will have to undertake the effort to automate such manufacturing processes. However, automating low-volume production is no small feat as the solution should be adaptable and future proof to unexpected changes in customer's demands. In this paper, we propose a reconfigurable robot workcell aimed at automating low-volume production. The developed workcell can adapt to the changes in manufacturing processes by employing a number of passive, reconfigurable hardware elements, supported by the ROS-based, modular control software. To further facilitate and expedite the setup process, we integrated intuitive, user-friendly robot programming methods with the available hardware. The system was evaluated by implementing five production processes from different manufacturing industries.
Nemec B, Simonič M, Ude A
Learning of Exception Strategies in Assembly Tasks Proceedings Article
In: 2020 IEEE International Conference on Robotics and Automation (ICRA), pp. 6521-6527, 2020.
@inproceedings{Nemec2020,
title = {Learning of Exception Strategies in Assembly Tasks},
author = {Bojan Nemec and Mihael Simonič and Aleš Ude},
doi = {10.1109/ICRA40945.2020.9197480},
year = {2020},
date = {2020-01-01},
booktitle = {2020 IEEE International Conference on Robotics and Automation (ICRA)},
pages = {6521-6527},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Zorko M, Nemec B, Matjačić Z, Olenšek A, Tomazin K, Supej M
Wide skis as a potential knee injury risk factor in alpine skiing Journal Article
In: Frontiers in sports and active living, vol. 2, pp. 7, 2020.
@article{zorko2020wide,
title = {Wide skis as a potential knee injury risk factor in alpine skiing},
author = {Martin Zorko and Bojan Nemec and Zlatko Matjačić and Andrej Olenšek and Katja Tomazin and Matej Supej},
year = {2020},
date = {2020-01-01},
journal = {Frontiers in sports and active living},
volume = {2},
pages = {7},
publisher = {Frontiers Media SA},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
2019
Simonič M, Žlajpah L, Ude A, Nemec B
Autonomous Learning of Assembly Tasks from the Corresponding Disassembly Tasks Proceedings Article
In: IEEE-RAS 19th International Conference on Humanoid Robots (Humanoids), pp. 230–236, IEEE Toronto, Canada, 2019.
@inproceedings{Simonic2019,
title = {Autonomous Learning of Assembly Tasks from the Corresponding Disassembly Tasks},
author = {Mihael Simonič and Leon Žlajpah and Aleš Ude and Bojan Nemec},
year = {2019},
date = {2019-01-01},
booktitle = {IEEE-RAS 19th International Conference on Humanoid Robots (Humanoids)},
pages = {230–236},
address = {Toronto, Canada},
organization = {IEEE},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Dežman M, Gams A
Optimization and analysis of the modified plvl-variable stiffness actuator Journal Article
In: International Journal of Mechanics and Control, vol. 20, no. 01, pp. 23–33, 2019.
@article{dezman2019,
title = {Optimization and analysis of the modified plvl-variable stiffness actuator},
author = {Miha Dežman and Andrej Gams},
year = {2019},
date = {2019-01-01},
journal = {International Journal of Mechanics and Control},
volume = {20},
number = {01},
pages = {23–33},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Dežman M, Asfour T, Ude A, Gams A
Exoskeleton arm pronation/supination assistance mechanism with a guided double rod system Proceedings Article
In: IEEE-RAS International Conference on Humanoid Robots (Humanoids), pp. 590-595, Toronto, ON, Canada, 2019.
@inproceedings{dezman2017hum,
title = {Exoskeleton arm pronation/supination assistance mechanism with a guided double rod system},
author = {Miha Dežman and Tamim Asfour and Aleš Ude and Andrej Gams},
doi = {10.1109/Humanoids43949.2019.9034992},
year = {2019},
date = {2019-01-01},
booktitle = {IEEE-RAS International Conference on Humanoid Robots (Humanoids)},
pages = {590-595},
address = {Toronto, ON, Canada},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Tamosiunaite M, Aein M J, Braun J M, Kulvicius T, Markievicz I, Kapociute-Dzikiene J, Valteryte R, Haidu A, Chrysostomou D, Ridge B, Krilavicius T, Vitkute-Adzgauskiene D, Beetz M, Madsen O, Ude A, Krüger N, Wörgötter F
Cut & recombine: reuse of robot action components based on simple language instructions Journal Article
In: The International Journal of Robotics Research, vol. 38, no. 10-11, pp. 1179-1207, 2019.
@article{doi:10.1177/0278364919865594,
title = {Cut & recombine: reuse of robot action components based on simple language instructions},
author = {Minija Tamosiunaite and Mohamad Javad Aein and Jan Matthias Braun and Tomas Kulvicius and Irena Markievicz and Jurgita Kapociute-Dzikiene and Rita Valteryte and Andrei Haidu and Dimitrios Chrysostomou and Barry Ridge and Tomas Krilavicius and Daiva Vitkute-Adzgauskiene and Michael Beetz and Ole Madsen and Ales Ude and Norbert Krüger and Florentin Wörgötter},
url = {https://doi.org/10.1177/0278364919865594},
doi = {10.1177/0278364919865594},
year = {2019},
date = {2019-01-01},
journal = {The International Journal of Robotics Research},
volume = {38},
number = {10-11},
pages = {1179-1207},
abstract = {Human beings can generalize from one action to similar ones. Robots cannot do this and progress concerning information transfer between robotic actions is slow. We have designed a system that performs action generalization for manipulation actions in different scenarios. It relies on an action representation for which we perform code-snippet replacement, combining information from different actions to form new ones. The system interprets human instructions via a parser using simplified language. It uses action and object names to index action data tables (ADTs), where execution-relevant information is stored. We have created an ADT database from three different sources (KUKA LWR, UR5, and simulation) and show how a new ADT is generated by cutting and recombining data from existing ADTs. To achieve this, a small set of action templates is used. After parsing a new instruction, index-based searching finds similar ADTs in the database. Then the action template of the new action is matched against the information in the similar ADTs. Code snippets are extracted and ranked according to matching quality. The new ADT is created by concatenating code snippets from best matches. For execution, only coordinate transforms are needed to account for the poses of the objects in the new scene. The system was evaluated, without additional error correction, using 45 unknown objects in 81 new action executions, with 80% success. We then extended the method including more detailed shape information, which further reduced errors. This demonstrates that cut & recombine is a viable approach for action generalization in service robotic applications.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Human beings can generalize from one action to similar ones. Robots cannot do this and progress concerning information transfer between robotic actions is slow. We have designed a system that performs action generalization for manipulation actions in different scenarios. It relies on an action representation for which we perform code-snippet replacement, combining information from different actions to form new ones. The system interprets human instructions via a parser using simplified language. It uses action and object names to index action data tables (ADTs), where execution-relevant information is stored. We have created an ADT database from three different sources (KUKA LWR, UR5, and simulation) and show how a new ADT is generated by cutting and recombining data from existing ADTs. To achieve this, a small set of action templates is used. After parsing a new instruction, index-based searching finds similar ADTs in the database. Then the action template of the new action is matched against the information in the similar ADTs. Code snippets are extracted and ranked according to matching quality. The new ADT is created by concatenating code snippets from best matches. For execution, only coordinate transforms are needed to account for the poses of the objects in the new scene. The system was evaluated, without additional error correction, using 45 unknown objects in 81 new action executions, with 80% success. We then extended the method including more detailed shape information, which further reduced errors. This demonstrates that cut & recombine is a viable approach for action generalization in service robotic applications.
Ridge B, Pahič R, Ude A, Morimoto J
Learning to Write Anywhere with Spatial Transformer Image-to-Motion Encoder-Decoder Networks Proceedings Article
In: 2019 International Conference on Robotics and Automation (ICRA), pp. 2111-2117, 2019.
@inproceedings{8794253,
title = {Learning to Write Anywhere with Spatial Transformer Image-to-Motion Encoder-Decoder Networks},
author = {Barry Ridge and Rok Pahič and Aleš Ude and Jun Morimoto},
doi = {10.1109/ICRA.2019.8794253},
year = {2019},
date = {2019-01-01},
booktitle = {2019 International Conference on Robotics and Automation (ICRA)},
pages = {2111-2117},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Nemec B, Simonič M, Petrič T, Ude A
Incremental Policy Refinement by Recursive Regression and Kinesthetic Guidance Proceedings Article
In: 2019 19th International Conference on Advanced Robotics (ICAR), pp. 344-349, 2019.
@inproceedings{8981606,
title = {Incremental Policy Refinement by Recursive Regression and Kinesthetic Guidance},
author = {Bojan Nemec and Mihael Simonič and Tadej Petrič and Aleš Ude},
doi = {10.1109/ICAR46387.2019.8981606},
year = {2019},
date = {2019-01-01},
booktitle = {2019 19th International Conference on Advanced Robotics (ICAR)},
pages = {344-349},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Gams A, Reberšek S, Nemec B, Škrabar J, Krhlikar R, Skvarč J, Ude A
Robotic Learning for Increased Productivity: Autonomously Improving Speed of Robotic Visual Quality Inspection Proceedings Article
In: 2019 IEEE 15th International Conference on Automation Science and Engineering (CASE), pp. 1275-1281, 2019.
@inproceedings{8842851,
title = {Robotic Learning for Increased Productivity: Autonomously Improving Speed of Robotic Visual Quality Inspection},
author = {Andrej Gams and Simon Reberšek and Bojan Nemec and Jure Škrabar and Rok Krhlikar and Jure Skvarč and Aleš Ude},
doi = {10.1109/COASE.2019.8842851},
year = {2019},
date = {2019-01-01},
booktitle = {2019 IEEE 15th International Conference on Automation Science and Engineering (CASE)},
pages = {1275-1281},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
2018
Wolniakowski A, Gams A, Kramberger A, Ude A
Compensating pose uncertainties through appropriate gripper finger cutoutS Journal Article
In: Acta Mechanica et Automatica, vol. 12, no. 1, pp. 78-83, 2018.
@article{wolniakowski2018,
title = {Compensating pose uncertainties through appropriate gripper finger cutoutS},
author = {Adam Wolniakowski and Andrej Gams and Aljaž Kramberger and Aleš Ude},
doi = {10.2478/ama-2018-0013},
year = {2018},
date = {2018-01-01},
journal = {Acta Mechanica et Automatica},
volume = {12},
number = {1},
pages = {78-83},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Nemec B, Likar N, Gams A, Ude A
Human robot cooperation with compliance adaptation along the motion trajectory Journal Article
In: Autonomous Robots, vol. 42, no. 5, pp. 1023-1035, 2018.
@article{nemec2018,
title = {Human robot cooperation with compliance adaptation along the motion trajectory},
author = {Bojan Nemec and Nejc Likar and Andrej Gams and Aleš Ude},
doi = {10.1007/s10514-017-9676-3},
year = {2018},
date = {2018-01-01},
journal = {Autonomous Robots},
volume = {42},
number = {5},
pages = {1023-1035},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Petrič T, Gams A, Colasanto L, Ijspeert A J, Ude A
Accelerated Sensorimotor Learning of compliant movement primitives Journal Article
In: IEEE Transactions on Robotics, vol. 34, no. 6, pp. 1636-1642, 2018.
@article{petric2018,
title = {Accelerated Sensorimotor Learning of compliant movement primitives},
author = {Tadej Petrič and Andrej Gams and Luca Colasanto and Auke Jan Ijspeert and Aleš Ude},
doi = {10.1109/TRO.2018.2861921},
year = {2018},
date = {2018-01-01},
journal = {IEEE Transactions on Robotics},
volume = {34},
number = {6},
pages = {1636-1642},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Dežman M, Gams A
Rotatable cam-based variable-ratio lever compliant actuator for wearable devices Journal Article
In: Mechanism and Machine Theory, vol. 130, pp. 508-522, 2018.
@article{dezman2018,
title = {Rotatable cam-based variable-ratio lever compliant actuator for wearable devices},
author = {Miha Dežman and Andrej Gams},
doi = {10.1016/j.mechmachtheory.2018.09.006},
year = {2018},
date = {2018-01-01},
journal = {Mechanism and Machine Theory},
volume = {130},
pages = {508-522},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Gašpar T, Nemec B, Morimoto J, Ude A
Skill learning and action recognition by arc-length dynamic movement primitives Journal Article
In: Robotics and Autonomous Systems, vol. 100, pp. 225-235, 2018.
@article{GASPAR2018225,
title = {Skill learning and action recognition by arc-length dynamic movement primitives},
author = {Timotej Gašpar and Bojan Nemec and Jun Morimoto and Aleš Ude},
url = {https://www.sciencedirect.com/science/article/pii/S0921889017302695},
doi = {https://doi.org/10.1016/j.robot.2017.11.012},
issn = {0921-8890},
year = {2018},
date = {2018-01-01},
journal = {Robotics and Autonomous Systems},
volume = {100},
pages = {225-235},
abstract = {Effective robot programming by demonstration requires the availability of multiple demonstrations to learn about all relevant aspects of the demonstrated skill or task. Typically, a human teacher must demonstrate several variants of the desired task to generate a sufficient amount of data to reliably learn it. Here a problem often arises that there is a large variability in the speed of execution across human demonstrations. This can cause problems when multiple demonstrations are compared to extract the relevant information for learning. In this paper we propose an extension of dynamic movement primitives called arc-length dynamic movement primitives, where spatial and temporal components of motion are well separated. We show theoretically and experimentally that the proposed representation can be effectively applied for robot skill learning and action recognition even when there are large variations in the speed of demonstrated movements.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Effective robot programming by demonstration requires the availability of multiple demonstrations to learn about all relevant aspects of the demonstrated skill or task. Typically, a human teacher must demonstrate several variants of the desired task to generate a sufficient amount of data to reliably learn it. Here a problem often arises that there is a large variability in the speed of execution across human demonstrations. This can cause problems when multiple demonstrations are compared to extract the relevant information for learning. In this paper we propose an extension of dynamic movement primitives called arc-length dynamic movement primitives, where spatial and temporal components of motion are well separated. We show theoretically and experimentally that the proposed representation can be effectively applied for robot skill learning and action recognition even when there are large variations in the speed of demonstrated movements.