@inproceedings{puang2024learning,

title = {Learning Stable Robot Grasping with Transformer-based Tactile Control Policies Authors},

author = {En Yen Puang and Zechen Li and Chee Meng Chew and Shan Luo and Yan Wu},

url = {https://arxiv.org/abs/2407.21172},

year  = {2024},

date = {2024-08-05},

urldate = {2024-08-05},

booktitle = {2024 IEEE 19th Conference on Industrial Electronics and Applications (ICIEA)},

abstract = {Measuring grasp stability is an important skill for dexterous robot manipulation tasks, which can be inferred from haptic information with a tactile sensor. Control policies have to detect rotational displacement and slippage from tactile feedback, and determine a re-grasp strategy in term of location and force. Classic stable grasp task only trains control policies to solve for re-grasp location with objects of fixed center of gravity. In this work, we propose a revamped version of stable grasp task that optimises both re-grasp location and gripping force for objects with unknown and moving center of gravity. We tackle this task with a model-free, end-to-end Transformer-based reinforcement learning framework. We show that our approach is able to solve both objectives after training in both simulation and in a real-world setup with zero-shot transfer. We also provide performance analysis of different models to understand the dynamics of optimizing two opposing objectives.},

keywords = {},

pubstate = {forthcoming},

tppubtype = {inproceedings}

}

@inproceedings{zhu2024gaussian,

title = {Gaussian Process Model Predictive Admittance Control for Compliant Tool-Environment Interaction},

author = {Wenxin Zhu and Wenyu Liang and Qinyuan Ren and Yan Wu},

year  = {2024},

date = {2024-08-05},

urldate = {2024-08-05},

booktitle = {2024 IEEE 19th Conference on Industrial Electronics and Applications (ICIEA)},

keywords = {},

pubstate = {forthcoming},

tppubtype = {inproceedings}

}

@inproceedings{zhao2024unknown,

title = {Unknown Object Retrieval in Confined Space through Reinforcement Learning with Tactile Exploration},

author = {Xinyuan Zhao and Wenyu Liang and Xiaoshi Zhang and Chee Meng Chew and Yan Wu},

url = {https://ieeexplore.ieee.org/document/10611541},

doi = {10.1109/ICRA57147.2024.10611541},

isbn = {979-8-3503-8457-4},

year  = {2024},

date = {2024-05-13},

booktitle = {2024 IEEE International Conference on Robotics and Automation (ICRA)},

pages = {10881-10887},

publisher = {IEEE},

abstract = {The potential of tactile sensing for dexterous robotic manipulation has been demonstrated by its ability to enable nuanced real-world interactions. In this study, the retrieval of unknown objects from confined spaces, which is unsuitable for conventional visual perception and gripper-based manipulation, is identified and addressed. Specifically, a tactile-sensorized tool stick that well fits in the narrow space is utilized to provide multi-point contact sensing for object manipulation. A reinforcement learning (RL) agent with a hybrid action space is then proposed to acquire the optimal policy for manipulating the objects without prior knowledge of their physical properties. To accelerate on-hardware training, a focused training strategy is adopted with the hypothesis that an agent trained on a small set of representative shapes can be generalized to a wide range of everyday objects. Additionally, a curriculum on terminal goals is designed to further accelerate the hardware-based training process. Comparative experiments and ablation studies have been conducted to evaluate the effectiveness and robustness of the proposed approach, which highlights the high success rate of our solution for retrieving everyday objects.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{language2023sui,

title = {Language Models can do Zero-Shot Visual Referring Expression Comprehension},

author = {Xiuchao Sui and Shaohua Li and Hong Yang and Hongyuan Zhu and Yan Wu},

url = {https://openreview.net/forum?id=F7mdgA7c2zD},

year  = {2023},

date = {2023-05-01},

urldate = {2023-05-01},

booktitle = {Eleventh International Conference on Learning Representations (ICLR) 2023},

series = {Tiny Paper},

abstract = {The use of visual referring expressions is an important aspect of human-robot interactions.  Comprehending referring expressions (ReC) like ``the brown cookie near the cup'' requires to understand both self-referential expressions, ``brown cookie", and relational referential expressions, ``near the cup''.  Large pretrained Vision-Language models like CLIP excel at handling self-referential expressions, while struggle with the latter. In this work, we reframe ReC as a language reasoning task and explore whether it can be addressed using large pretrained language models (LLMs), including Bing Chat and ChatGPT. Given the textual attribute tuples {object category, color, center location, size}, ChatGPT performs unstably on understanding spatial relationships even with heavy prompt engineering, while Bing Chat shows strong and stable zero-shot relation reasoning. Evaluation on RefCOCO/g datasets and scenarios of interactive robot grasping shows that LLMs can do ReC with decent performance. It suggests a vast potential of using LLMs to enhance the reasoning in vision tasks.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{wang2022end,

title = {End-to-end Reinforcement Learning of Robotic Manipulation with Robust Keypoints Representation},

author = {Tianying Wang and En Yen Puang and Marcus Lee and Yan Wu and Wei Jing},

doi = {10.23919/APSIPAASC55919.2022.9980136},

isbn = {978-616-590-477-3},

year  = {2022},

date = {2022-11-10},

urldate = {2022-11-10},

booktitle = {2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)},

abstract = {We present an end-to-end Reinforcement Learning (RL) framework for robotic manipulation tasks, using a robust and efficient keypoints representation. The proposed method learns keypoints from camera images as the state representation, through a self-supervised autoencoder architecture. The key-points encode the geometric information, as well as the relationship of the tool and target in a compact representation to ensure efficient and robust learning. After keypoints learning, the RL step then learns the robot motion from the extracted keypoints state representation. The keypoints and RL learning processes are entirely done in the simulated environment. We demonstrate the effectiveness of the proposed method on robotic manipulation tasks including grasping and pushing, in different scenarios. We also investigate the generalization capability of the trained model. In addition to the robust keypoints representation, we further apply domain randomization and adversarial training examples to achieve zero-shot sim-to-real transfer in real-world robotic manipulation tasks.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{liang2022tactile,

title = {Tactile-Guided Dynamic Object Planar Manipulation},

author = {Boyuan Liang and Wenyu Liang and Yan Wu},

doi = {10.1109/IROS47612.2022.9981270},

isbn = {978-1-6654-7927-1},

year  = {2022},

date = {2022-10-23},

urldate = {2022-10-31},

booktitle = {2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},

pages = {3203-3209},

publisher = {IEEE},

abstract = {Planar pushing is a fundamental robot manipulation task with most algorithms built upon the quasi-static as-sumption. Under this assumption the end-effector should apply force on the pushed object along the full moving trajectory. This means that the target position must lie in the robot's workspace. To enable a robot to deliver objects outside of its workspace and facilitate faster delivery, the quasi-static assumption should be lifted in favour of dynamical manipulation. In this work, we propose a two-staged data-driven manipulation method to hit an unknown object to reach a target position. This expands the reachability of the manipulated object beyond the robot's workspace. The robot equipped with a tactile sensor first explores for the stable pushing region (SPR) on the given object by using a gain-scheduling PD control with the contact centre estimated to maintain full contact between the object and the end-effector. In the second stage, a learning-based approach is used to generate the impulse the object should receive at the SPR to reach a target sliding distance. The performance of proposed method is evaluated on a KUKA LBR iiwa 14 R820 robot manipulator and a XELA tactile sensor.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{fang2022improving,

title = {Improving Generalization of Reinforcement Learning Using a Bilinear Policy Network},

author = {Fen Fang and Wenyu Liang and Yan Wu and Qianli Xu and Joo-Hwee Lim},

doi = {10.1109/ICIP46576.2022.9897349},

isbn = {978-1-6654-9620-9},

year  = {2022},

date = {2022-10-18},

urldate = {2022-10-18},

booktitle = {2022 IEEE International Conference on Image Processing (ICIP)},

pages = {991-995},

publisher = {IEEE},

abstract = {In deep reinforcement learning (DRL), the agent is usually trained on seen environments by optimizing a policy network. However, it is difficult to be generalized to unseen environments properly, even when the environmental variations are insignificant. This is partly because the policy network cannot effectively learn the representation of visual difference that is subtle among highly similar states in the environments. Because a bilinear structured model containing two feature extractors allows pairwise feature interactions in a translation-ally invariant manner which makes it particularly useful for subtle difference recognition among highly similar states, in this work, a bilinear policy network is employed to enhance representation learning, and thus to improve generalization of the DRL. The proposed bilinear policy network is tested on various DRL task, including a control task on path planning for active object detection, and Grid World, an AI game task. The test results show that the generalization of DRL can be improved by the proposed network.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{fang2022self,

title = {Self-Supervised Reinforcement Learning for Active Object Detection},

author = {Fen Fang and Wenyu Liang and Yan Wu and Qianli Xu and Joo Hwee Lim},

doi = {10.1109/LRA.2022.3193019},

issn = {2377-3766},

year  = {2022},

date = {2022-07-21},

urldate = {2022-10-31},

booktitle = {2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},

volume = {7},

number = {4},

pages = {10224-10231},

publisher = {IEEE},

abstract = {Active object detection (AOD) offers significant advantage in expanding the perceptual capacity of a robotics system. AOD is formulated as a sequential action decision process to determine optimal viewpoints to identify objects of interest in a visual scene. While reinforcement learning (RL) has been successfully used to solve many AOD problems, conventional RL methods suffer from (i) sample inefficiency, and (ii) unstable outcome due to inter-dependencies of action type (direction of view change) and action range (step size of view change). To address these issues, we propose a novel self-supervised RL method, which employs self-supervised representations of viewpoints to initialize the policy network, and a self-supervised loss on action range to enhance the network parameter optimization. The output and target pairs of self-supervised learning loss are automatically generated from the policy network online prediction and a range shrinkage algorithm (RSA), respectively. The proposed method is evaluated and benchmarked on two public datasets (T-LESS and AVD) using on-policy and off-policy RL algorithms. The results show that our method enhances detection accuracy and achieves faster convergence on both datasets. By evaluating on a more complex environment with a larger state space (where viewpoints are more densely sampled), our method achieves more robust and stable performance. Our experiment on real robot application scenario to disambiguate similar objects in a cluttered scene has also demonstrated the effectiveness of the proposed method.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{sui2022craft,

title = {CRAFT: Cross-Attentional Flow Transformers for Robust Optical Flow},

author = {Xiuchao Sui and Shaohua Li and Xue Geng and Yan Wu and Xinxing Xu and Yong Liu and Rick Goh and Hongyuan Zhu },

url = {https://openaccess.thecvf.com/content/CVPR2022/papers/Sui_CRAFT_Cross-Attentional_Flow_Transformer_for_Robust_Optical_Flow_CVPR_2022_paper.pdf},

year  = {2022},

date = {2022-06-24},

urldate = {2022-06-24},

booktitle = {2022 Conference on Computer Vision and Pattern Recognition (CVPR)},

abstract = {Optical flow estimation aims to find the 2D motion field by identifying corresponding pixels between two images. Despite the tremendous progress of deep learning-based optical flow methods, it remains a challenge to accurately estimate large displacements with motion blur. This is mainly because the correlation volume, the basis of pixel matching, is computed as the dot product of the convolutional features of the two images. The locality of convolutional features makes the computed correlations susceptible to various noises. On large displacements with motion blur, noisy correlations could cause severe errors in the estimated flow. To overcome this challenge, we propose a new architecture" CRoss-Attentional Flow Transformer"(CRAFT), aiming to revitalize the correlation volume computation. In CRAFT, a Semantic Smoothing Transformer layer transforms the features of one frame, making them more global and semantically stable. In addition, the dot-product correlations are replaced with transformer Cross-Frame Attention. This layer filters out feature noises through the Query and Key projections, and computes more accurate correlations. On Sintel (Final) and KITTI (foreground) benchmarks, CRAFT has achieved new state-of-the-art performance. Moreover, to test the robustness of different models on large motions, we designed an image shifting attack that shifts input images to generate large artificial motions. Under this attack, CRAFT performs much more robustly than two representative methods, RAFT and GMA. The code of CRAFT is is available at https://github. com/askerlee/craft.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{gao2021explainability,

title = {On Explainability and Sensor-Adaptability of a Robot Tactile Texture Representation Using a Two-Stage Recurrent Networks},

author = {Ruihan Gao and Tian Tian and Zhiping Lin and Yan Wu},

year  = {2021},

date = {2021-09-27},

booktitle = {2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) },

abstract = {The ability to simultaneously distinguish objects, materials, and their associated physical properties is one fundamental function of the sense of touch. Recent advances in the development of tactile sensors and machine learning techniques allow more accurate and complex modelling of robotic tactile sensations. However, many state-of-the-art (SotA) approaches focus solely on constructing black-box models to achieve ever higher classification accuracy and fail to adapt across sensors with unique spatial-temporal data formats. In this work, we propose an Explainable and Sensor-Adaptable Recurrent Networks (ExSARN) model for tactile texture representation. The ExSARN model consists of a two-stage recurrent networks fed by a sensor-specific header network. The first stage recurrent network emulates our human touch receptors and decouples sensor-specific tactile sensations into different frequency response bands, while the second stage codes the overall temporal signature as a variational recurrent autoencoder. We infuse the latent representation with ternary labels to qualitatively represent texture properties (e.g. roughness and stiffness), which facilitates representation learning and provide explainability to the latent space. The ExSARN model is tested on texture datasets collected with two different tactile sensors.  Our results show that the proposed model not only achieves higher accuracy, but also provides adaptability across sensors with different sampling frequencies and data formats. The addition of the crudely obtained qualitative property labels offers a practical approach to enhance the interpretability of the latent space, facilitate property inference on unseen materials, and improve the overall performance of the model.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{xu2021efficient,

title = {Towards Efficient Multiview Object Detection with Adaptive Action Prediction},

author = {Qianli Xu and Fen Fang and Nicolas Gauthier and Wenyu Liang and Yan Wu and Liyuan Li and Joo Hwee Lim },

url = {https://ieeexplore.ieee.org/document/9561388},

doi = {10.1109/ICRA48506.2021.9561388},

isbn = {978-1-7281-9077-8},

year  = {2021},

date = {2021-05-31},

booktitle = {2021 IEEE International Conference on Robotics and Automation (ICRA)},

publisher = {IEEE},

abstract = {Active vision is a desirable perceptual feature for robots. Existing approaches usually make strong assumptions about the task and environment, thus are less robust and efficient. This study proposes an adaptive view planning approach to boost the efficiency and robustness of active object detection. We formulate the multi-object detection task as an active multiview object detection problem given the initial location of the objects. Next, we propose a novel adaptive action prediction (A2P) method built on a deep Q-learning network with a dueling architecture. The A2P method is able to perform view planning based on visual information of multiple objects; and adjust action ranges according to the task status. Evaluated on the AVD dataset, A2P leads to 21.9% increase in detection accuracy in unfamiliar environments, while improving efficiency by 22.7%. On the T-LESS dataset, multi-object detection boosts efficiency by more than 30% while achieving equivalent detection accuracy.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{liang2021dexterous,

title = {Dexterous Manoeuvre through Touch in a Cluttered Scene},

author = {Wenyu Liang and Qinyuan Ren and Xiaoqiao Chen and Junli Gao and Yan Wu},

url = {http://yan-wu.com/wp-content/uploads/2021/03/liang2021dexterous.pdf

https://ieeexplore.ieee.org/document/9562061},

doi = {10.1109/ICRA48506.2021.9562061},

isbn = {978-1-7281-9077-8},

year  = {2021},

date = {2021-05-31},

booktitle = {2021 IEEE International Conference on Robotics and Automation (ICRA)},

publisher = {IEEE},

abstract = {Manipulation in a densely cluttered environment creates complex challenges in perception to close the control loop, many of which are due to the sophisticated physical interaction between the environment and the manipulator. Drawing from biological sensory-motor control, to handle the task in such a scenario, tactile sensing can be used to provide an additional dimension of the rich contact information from the interaction for decision making and action selection to manoeuvre towards a target. In this paper, a new tactile-based motion planning and control framework based on bioinspiration is proposed and developed for a robot manipulator to manoeuvre in a cluttered environment. An iterative two-stage machine learning approach is used in this framework: an autoencoder is used to extract important cues from tactile sensory readings while a reinforcement learning technique is used to generate optimal motion sequence to efficiently reach the given target. The framework is implemented on a KUKA LBR iiwa robot mounted with a SynTouch BioTac tactile sensor and tested with real-life experiments. The results show that the system is able to move the end-effector through the cluttered environment to reach the target effectively.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{xu2021tailor,

title = {TAILOR: Teaching with Active and Incremental Learning for Object Registration},

author = {Qianli Xu and Nicolas Gauthier and Wenyu Liang and Fen Fang and Hui Li Tan and Ying Sun and Yan Wu and Liyuan Li and Joo Hwee Lim},

url = {http://yan-wu.com/wp-content/uploads/2021/03/xu2021tailor.pdf

http://yan-wu.com/wp-content/uploads/2021/03/xu2021tailor_poster.pdf

https://twitter.com/RealAAAI/status/1364017094086389760

https://ojs.aaai.org/index.php/AAAI/article/view/18031},

year  = {2021},

date = {2021-05-01},

booktitle = {Proceedings of the Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI)},

volume = {35},

number = {18},

pages = {16120-16123},

publisher = {AAAI},

abstract = {When a robot is deployed to a new task, it often needs to be trained to detect novel objects. Using deep learning based detector, one has to collect and annotate a large number of images of the novel objects for training, which is labor intensive, time consuming and lack of scalability. We present TAILOR - a method and system for object registration with active and incremental learning. When instructed by a human teacher to register an object, TAILOR is able to automatically select viewpoints to capture informative images by actively exploring viewpoints, and employ a fast incremental learning algorithm to learn new objects without potential forgetting of previously learned objects. We demonstrate the effectiveness of our method with a KUKA robot to learn novel objects used in a real-world gearbox assembly task through natural interactions.},

note = {AAAI'21 Best Demo Award},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{6d2020cheng,

title = {6D Pose Estimation with Correlation Fusion},

author = {Yi Cheng and Hongyuan Zhu and Ying Sun and Cihan Acar and Wei Jing and Yan Wu and Liyuan Li and Cheston Tan and Joo Hwee Lim},

url = {https://arxiv.org/abs/1909.12936

https://ieeexplore.ieee.org/document/9412238},

doi = {10.1109/ICPR48806.2021.9412238},

isbn = {978-1-7281-8808-9},

year  = {2021},

date = {2021-01-15},

booktitle = {25th International Conference on Pattern Recognition (ICPR)},

pages = {2988-2994},

address = {Milan, Italy},

abstract = {6D object pose estimation is widely applied in robotic tasks such as grasping and manipulation. Prior methods using RGB-only images are vulnerable to heavy occlusion and poor illumination, so it is important to complement them with depth information. However, existing methods using RGB-D data don't adequately exploit consistent and complementary information between two modalities. In this paper, we present a novel method to effectively consider the correlation within and across RGB and depth modalities with attention mechanism to learn discriminative multi-modal features. Then, effective fusion strategies for intra- and inter-correlation modules are explored to ensure efficient information flow between RGB and depth. To the best of our knowledge, this is the first work to explore effective intra- and inter-modality fusion in 6D pose estimation and experimental results show that our method can help achieve the state-of-the-art performance on LineMOD and YCB-Video datasets as well as benefit robot grasping task.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{gao2020supervised,

title = {Supervised Autoencoder Joint Learning on Heterogeneous Tactile Sensory Data: Improving Material Classification Performance},

author = {Ruihan Gao and Tasbolat Taunyazov and Zhiping Lin and Yan Wu},

url = {http://yan-wu.com/wp-content/uploads/2020/08/gao2020supervised.pdf

https://ieeexplore.ieee.org/document/9341111},

doi = {10.1109/IROS45743.2020.9341111},

isbn = {978-1-7281-6212-6},

year  = {2020},

date = {2020-10-31},

booktitle = {2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},

publisher = {IEEE},

address = {Las Vegas, USA},

abstract = {The sense of touch is an essential sensing modality for a robot to interact with the environment as it provides rich and multimodal sensory information upon contact. It enriches the perceptual understanding of the environment and closes the loop for action generation. One fundamental area of perception that touch dominates over other sensing modalities, is the understanding of the materials that it interacts with, for example, glass versus plastic. However, unlike the senses of vision and audition which have standardized data format, the format for tactile data is vastly dictated by the sensor manufacturer, which makes it difficult for large-scale learning on data collected from heterogeneous sensors, limiting the usefulness of publicly available tactile datasets. This paper investigates the joint learnability of data collected from two tactile sensors performing a touch sequence on some common materials. We propose a supervised recurrent autoencoder framework to perform joint material classification task to improve the training effectiveness. The framework is implemented and tested on the two sets of tactile data collected in sliding motion on 20 material textures using the iCub RoboSkin tactile sensors and the SynTouch BioTac sensor respectively. Our results show that the learning efficiency and accuracy improve for both datasets through the joint learning as compared to independent dataset training. This suggests the usefulness for large-scale open tactile datasets sharing with different sensors.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{liang2020robust,

title = {Robust Force Tracking Impedance Control of an Ultrasonic Motor-actuated End-effector in a Soft Environment},

author = {Wenyu Liang and Zhao Feng and Yan Wu and Junli Gao and Qinyuan Ren and Tong Heng Lee},

url = {http://yan-wu.com/wp-content/uploads/2020/08/liang2020robust.pdf

https://ieeexplore.ieee.org/document/9340717},

doi = {10.1109/IROS45743.2020.9340717},

year  = {2020},

date = {2020-10-31},

booktitle = {2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},

publisher = {IEEE},

address = {Las Vegas, USA},

abstract = {Robotic systems are increasingly required not only to generate precise motions to complete their tasks but also to handle the interactions with the environment or human. Significantly, soft interaction brings great challenges on the force control due to the nonlinear, viscoelastic and inhomogeneous properties of the soft environment. In this paper, a robust impedance control scheme utilizing integral backstepping technology and integral terminal sliding mode control is proposed to achieve force tracking for an ultrasonic motor-actuated end-effector in a soft environment. In particular, the steady-state performance of the target impedance while in contact with soft environment is derived and analyzed with the nonlinear Hunt-Crossley model. Finally, the dynamic force tracking performance of the proposed control scheme is verified via several experiments.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{taunyazov2020fast,

title = {Fast Texture Classification Using Tactile Neural Coding and Spiking Neural Network},

author = {Tasbolat Taunyazov and Yansong Chua and Ruihan Gao and Harold Soh and Yan Wu },

url = {http://yan-wu.com/wp-content/uploads/2020/08/taunyazov2020fast.pdf

https://ieeexplore.ieee.org/document/9340693},

doi = {10.1109/IROS45743.2020.9340693},

year  = {2020},

date = {2020-10-31},

booktitle = {2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},

journal = {2020},

publisher = {IEEE},

address = {Las Vegas, USA},

abstract = {Touch is arguably the most important sensing modality in physical interactions. However, tactile sensing has been largely under-explored in robotics applications owing to the complexity in making perceptual inferences until the recent advancements in machine learning or deep learning in particular. Touch perception is strongly influenced by both its temporal dimension similar to audition and its spatial dimension similar to vision. While spatial cues can be learned episodically, temporal cues compete against the system's response/reaction time to provide accurate inferences. In this paper, we propose a fast tactile-based texture classification framework which makes use of the spiking neural network to learn from the neural coding of the conventional tactile sensor readings. The framework is implemented and tested on two independent tactile datasets collected in sliding motion on 20 material textures. Our results show that the framework is able to make much more accurate inferences ahead of time as compared to that by the state-of-the-art learning approaches.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{jing2020multi,

title = {Multi-UAV Coverage Path Planning for the Inspection of Large and Complex Structures via Path Primitive Sampling},

author = {Wei Jing and Di Deng and Yan Wu and Kenji Shimada},

url = {http://yan-wu.com/wp-content/uploads/2020/08/jing2020multi.pdf

https://ieeexplore.ieee.org/document/9341089},

doi = {10.1109/IROS45743.2020.9341089},

year  = {2020},

date = {2020-10-31},

booktitle = {2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},

publisher = {IEEE},

address = {Las Vegas, USA},

abstract = {We present a multi-UAV  cpp framework for the inspection of large-scale, complex 3D structures. In the proposed sampling-based coverage path planning method, we formulate the multi-UAV inspection applications as a multi-agent coverage path planning problem. By combining two NP-hard problems: scp and vrp, a sc-vrp is formulated and subsequently solved by a modified brkga with novel, efficient encoding strategies and local improvement heuristics. We test our proposed method for several complex 3D structures with the 3D model extracted from OpenStreetMap. The proposed method outperforms previous methods, by reducing the length of the planned inspection path by up to 48%.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{cai2020inferring,

title = {Inferring the Geometric Nullspace of Robot Skills from Human Demonstrations},

author = {Caixia Cai and Ying Siu Liang and Nikhil Somani and Yan Wu},

url = {https://ieeexplore.ieee.org/document/9197174

https://yan-wu.com/wp-content/uploads/2020/05/cai2020inferring.pdf},

doi = {10.1109/ICRA40945.2020.9197174},

issn = {2577-087X},

year  = {2020},

date = {2020-05-31},

booktitle = {2020 IEEE International Conference on Robotics and Automation (ICRA)},

pages = {7668-7675},

publisher = {IEEE},

address = {Paris, France},

abstract = {In this paper we present a framework to learn skills from human demonstrations in the form of geometric nullspaces, which can be executed using a robot. We collect data of human demonstrations, fit geometric nullspaces to them, and also infer their corresponding geometric constraint models. These geometric constraints provide a powerful mathematical model as well as an intuitive representation of the skill in terms of the involved objects. To execute the skill using a robot, we combine this geometric skill description with the robot's kinematics and other environmental constraints, from which poses can be sampled for the robot's execution. The result of our framework is a system that takes the human demonstrations as input, learns the underlying skill model, and executes the learnt skill with different robots in different dynamic environments. We evaluate our approach on a simulated industrial robot, and execute the final task on the iCub humanoid robot.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{chi2019motion,

title = {Motion Control of a Soft Circular Crawling Robot via Iterative Learning Control},

author = {Haozhen Chi and Xuefang Li and Wenyu Liang and Yan Wu and Qinyuan Ren},

url = {https://ieeexplore.ieee.org/document/9029234 

https://yan-wu.com/wp-content/uploads/2020/05/chi2019motion.pdf},

doi = {10.1109/CDC40024.2019.9029234},

isbn = {978-1-7281-1398-2},

year  = {2019},

date = {2019-12-13},

booktitle = {2019 IEEE 58th Conference on Decision and Control (CDC)},

pages = {6524-6529},

publisher = {IEEE},

address = {Nice, France},

abstract = {Soft robots have recently attracted widespread attention due to their abilities to work effectively in unstructured environments. As an actuation technology of soft robots, dielectric elastomer actuators (DEAs) exhibit many fantastic attributes such as large strain and high energy density. However, due to nonlinear electromechanical coupling, it is challenging to model a DEA accurately, and further it is difficult to control a DEA-based soft robot. This work studies a novel DEA-based soft circular crawling robot. The kinematics of the soft robot is explored and a knowledge-based model is established to expedite the controller design. An iterative learning control (ILC) method then is applied to control the soft robot. By employing ILC, the performance of the robot motion trajectory tracking can be improved significantly without using a perfect model. Finally, several numerical studies are conducted to illustrate the effectiveness of the ILC.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{wang2019efficient,

title = {Efficient Robotic Task Generalization Using Deep Model Fusion Reinforcement Learning},

author = {Tianying Wang and Hao Zhang and Wei Qi Toh and Hongyuan Zhu and Cheston Tan and Yan Wu and Yong Liu and Wei Jing},

url = {https://ieeexplore.ieee.org/document/8961391 

https://www.yan-wu.com/wp-content/uploads/2020/05/wang2019efficient.pdf},

doi = {10.1109/ROBIO49542.2019.8961391},

isbn = {978-1-7281-6321-5},

year  = {2019},

date = {2019-12-08},

booktitle = {2019 IEEE International Conference on Robotics and Biomimetics (ROBIO)},

pages = {148-153},

publisher = {IEEE},

address = {Dali, China},

abstract = {Learning-based methods have been used to program robotic tasks in recent years. However, extensive training is usually required not only for the initial task learning but also for generalizing the learned model to the same task but in different environments. In this paper, we propose a novel Deep Reinforcement Learning algorithm for efficient task generalization and environment adaptation in the robotic task learning problem. The proposed method is able to efficiently generalize the previously learned task by model fusion to solve the environment adaptation problem. The proposed Deep Model Fusion (DMF) method reuses and combines the previously trained model to improve the learning efficiency and results. Besides, we also introduce a Multi-objective Guided Reward (MGR) shaping technique to further improve training efficiency. The proposed method was benchmarked with previous methods in various environments to validate its effectiveness.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{pang2019modelling,

title = {Modeling and Control of A Soft Circular Crawling Robot},

author = {Jiangnan Pang and Yibo Shao and Haozhen Chi and Yan Wu},

url = {https://ieeexplore.ieee.org/document/8927474 

https://www.yan-wu.com/wp-content/uploads/2020/05/pang2019modeling.pdf},

doi = {10.1109/IECON.2019.8927474},

isbn = {978-1-7281-4878-6},

year  = {2019},

date = {2019-10-17},

booktitle = {The 45th Annual Conference of the IEEE Industrial Electronics Society (IECON)},

volume = {1},

pages = {5243-5248},

publisher = {IEEE},

address = {Lisbon, Portugal},

abstract = {Soft robots have exhibited significant advantages compared to conventional rigid robots due to the high-energy density and strong environmental compliance. Among the soft materials explored for soft robots, dielectric elastomers (DEs) stand out with the muscle-like actuation behaviors. However, recently, modeling and control of a DE-based soft robot still remain a challenging because of the nonlinearity and viscoelasticity of DE actuators. This paper focuses on the design, modeling and control of a soft circular robot which is able to achieve a 2D motion. To facilitate the design of a motion controller, a dynamic model of the robot is investigated through experimental identification. Based on the model, a feedforward plus feedback control scheme is adopted for the motion control of the robot. Finally, both simulations and experiments are conducted to verify the effectiveness of the proposed model and control approach.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{taunyazov2019towards,

title = {Towards Effective Tactile Identification of Textures using a Hybrid Touch Approach},

author = {Tasbolat Taunyazov and Hui Fang Koh and Yan Wu and Caixia Cai and Harold Soh},

url = {https://ieeexplore.ieee.org/document/8793967 

http://www.yan-wu.com/docs/taunyanov2019towards.pdf},

doi = {10.1109/ICRA.2019.8793967},

isbn = {978-1-5386-6027-0},

year  = {2019},

date = {2019-05-24},

booktitle = {2019 International Conference on Robotics and Automation (ICRA)},

pages = {4269-4275},

publisher = {IEEE},

address = {Montreal, Canada},

abstract = {The sense of touch is arguably the first human sense to develop. Empowering robots with the sense of touch may augment their understanding of interacted objects and the environment beyond standard sensory modalities (e.g., vision). This paper investigates the effect of hybridizing touch and sliding movements for tactile-based texture classification. We develop three machine-learning methods within a framework to discriminate between surface textures; the first two methods use hand-engineered features, whilst the third leverages convolutional and recurrent neural network layers to learn feature representations from raw data. To compare these methods, we constructed a dataset comprising tactile data from 23 textures gathered using the iCub platform under a loosely constrained setup, i.e., with nonlinear motion. In line with findings from neuroscience, our experiments show that a good initial estimate can be obtained via touch data, which can be further refined via sliding; combining both touch and sliding data results in 98% classification accuracy over unseen test data.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{you20193d,

title = {3D Visibility Graph Based Motion Planning and Control},

author = {Yangwei You and Caixia Cai and Yan Wu},

url = {https://dl.acm.org/doi/pdf/10.1145/3373724.3373735},

doi = {10.1145/3373724.3373735},

isbn = {9781450372350},

year  = {2019},

date = {2019-01-01},

booktitle = {The 2019 5th International Conference on Robotics and Artificial Intelligence (ICRAI)},

pages = {48–53},

publisher = {ACM},

address = {Singapore},

abstract = {This paper proposed a motion planning and control approach for real-time collision avoidance of high degree-of-freedom (DoF) robots. It constructs a 3D visibility graph by representing external obstacles with discrete polyhedrons and searches for the shortest path on it. To consider the complex internal structure of high DoF robots and take full advantage of redundancy, inverse kinematics is solved as an optimization problem which allows for flexible cost functions and constraints. The distance between the robot and external obstacles is treated as an extra constraint in the optimization which works together with visibility graph for collision avoidance. At the end, a pick-place task was successfully performed among obstacles by a simulated dual-arm quadrupedal robot considering the whole-body kinematics and stability, which demonstrated the feasibility of the proposed method.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{wong2018advbot,

title = {Advbot: Towards Understanding Human Preference in a Human-Robot Interaction Scenario},

author = {Clarice Jiaying Wong and Yong Ling Tay and Lincoln W C Lew and Hui Fang Koh and Yijing Xiong and Yan Wu},

url = {https://ieeexplore.ieee.org/document/8581325 

https://www.yan-wu.com/wp-content/uploads/2020/05/wong2018advbot.pdf},

doi = {10.1109/ICARCV.2018.8581325},

year  = {2018},

date = {2018-11-01},

booktitle = {The 15th International Conference on Control, Automation, Robotics and Vision (ICARCV)},

pages = {1305-1309},

publisher = {IEEE},

address = {Singapore},

abstract = {Recent studies show that the growth of social robotics market will dominate the robotics sector by 2025. Social robots are set to enter and co-exist in daily living, making human-robot interaction studies a crucial research direction in these robotics applications. One such application is the deployment of robotics in information dissemination to augment the shrinking workforce in developed economies, such as advertisement campaigns. This work seeks to understand whether live robot advertiser can better engage audience and improve audience perception of a marketed product as compared to pre-filmed video advertising clips. In this preliminary study, a humanoid robot, the NAO, was programmed to act according to a marketing campaign script, engaging the audience through voice, simulated eye-contacts, gestures and interaction with the Keepon, a robot to be advertised. We carried out three sets of experiments in the CBD of Singapore on teenage pedestrians. The results of this study suggest that physical robot presence will enhance information dissemination and hence improve advertising result, providing value-add to the advertised product.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{tee2018experimental,

title = {Experimental Evaluation of Divisible Human-Robot Shared Control for Teleoperation Assistance},

author = {Keng Peng Tee and Yan Wu},

url = {https://ieeexplore.ieee.org/document/8650436 

https://www.yan-wu.com/wp-content/uploads/2020/05/tee2018experimental.pdf},

doi = {10.1109/TENCON.2018.8650436},

isbn = {978-1-5386-5457-6},

year  = {2018},

date = {2018-10-31},

booktitle = {2018 IEEE Region 10 Conference (TENCON)},

pages = {0182-0187},

publisher = {IEEE},

address = {Jeju, South Korea},

abstract = {This paper is concerned with divisible shared control, which decomposes the motion space into complementary subspaces and distributes the control to the human and the robot so that each can independently effect motion control in its subspace. We present a divisible shared control scheme to assist teleoperation tasks on a curved object surface, which is difficult for a human to perform without assistance. We designed and carried an experiment to investigate its effect of user performance and work load. Experimental evaluation, based on both quantitative and qualitative measures, suggests that divisible shared control improves accuracy, speed, and smoothness, while at the same time reduces cognitive load, effort, and frustration.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{wu2018multimodal,

title = {Multi-Modal Robot Apprenticeship: Imitation Learning Using Linearly Decayed DMP+ in a Human-Robot Dialogue System},

author = {Yan Wu and Ruohan Wang and Luis Fernando. D'Haro and Rafael E Banchs and Keng Peng Tee},

url = {https://ieeexplore.ieee.org/document/8593634 

https://www.yan-wu.com/wp-content/uploads/2020/05/wu2018multimodal.pdf},

doi = {10.1109/IROS.2018.8593634},

isbn = {978-1-5386-8094-0},

year  = {2018},

date = {2018-10-05},

booktitle = {2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},

pages = {1-7},

publisher = {IEEE},

address = {Madrid, Spain},

abstract = {Robot learning by demonstration gives robots the ability to learn tasks which they have not been programmed to do before. The paradigm allows robots to work in a greater range of real-world applications in our daily life. However, this paradigm has traditionally been applied to learn tasks from a single demonstration modality. This restricts the approach to be scaled to learn and execute a series of tasks in a real-life environment. In this paper, we propose a multi-modal learning approach using DMP+ with linear decay integrated in a dialogue system with speech and ontology for the robot to learn seamlessly through natural interaction modalities (like an apprentice) while learning or re-learning is done on the fly to allow partial updates to a learned task to reduce potential user fatigue and operational downtime in teaching. The performance of new DMP+ with linear decay system is statistically benchmarked against state-of-the-art DMP implementations. A gluing demonstration is also conducted to show how the system provides seamless learning of multiple tasks in a flexible manufacturing set-up.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{wu2017investigation,

title = {Investigation on the Roles of Human and Robot in Collaborative Storytelling},

author = {Yan Wu and Ruohan Wang and Yong Ling Tay and Clarice Jiaying Wong},

url = {https://ieeexplore.ieee.org/document/8282003 

http://www.yan-wu.com/docs/wu2017investigation.pdf},

doi = {10.1109/APSIPA.2017.8282003},

isbn = {978-1-5386-1542-3},

year  = {2017},

date = {2017-12-15},

booktitle = {2017 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)},

pages = {063-068},

publisher = {IEEE},

address = {Kuala Lumpur, Malaysia},

abstract = {The tradition of storytelling is a distinctive human experience. There have been increasing evidences from social sciences that storytelling facilitates skill development in children. In this paper, we explore the use of humanoid robots to augment traditional storytelling, an area still understudied, to investigate the effectiveness of human-robot collaboration (HRC) in storytelling. In this preliminary experiment, a humanoid robot, Aldebaran's Nao, was programmed to recite a mystery story to 60 students aged 14 to 15, while engaging them through gestures, simulated eye contacts and varied voices. Nao delivered the performance as either an independent storyteller, or as a collaborator with a human storyteller. We assessed the effectiveness of HRC by comparing the participants' preference over the two settings. We found that 1) most participants believe that HRC is effective and prefer HRC over the robot-only performance; and 2) the complementarily effective roles of human and robot storytellers in interacting with the participants provide a glimpse of evidence on the preference towards HRC. These results, provide a first step towards effective use of robots for collaborative storytelling in day-to-day situations, which is well placed to guide the design of robots for education and entertainment purposes, e.g. for improving family bonding between parents and children while allowing parents to more effectively tell stories.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{catenacci2017towards,

title = {Towards Event-based MCTS for Autonomous Cars},

author = {Nicola Catenacci Volpi and Yan Wu and Dimitri Ognibene},

url = {https://ieeexplore.ieee.org/document/8282068 

http://www.yan-wu.com/docs/catenacci2017towards.pdf},

doi = {10.1109/APSIPA.2017.8282068},

isbn = {978-1-5386-1542-3},

year  = {2017},

date = {2017-12-15},

booktitle = {2017 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)},

pages = {420-427},

publisher = {IEEE},

address = {Kuala Lumpur, Malaysia},

abstract = {Uncertainty in the behaviours of vehicles surrounding a self- driving car introduces substantial computational complexity in autonomous driving. In this study 1 , a data-driven approach was used to extract probabilistic models of the behaviours of other cars and exploit them to support a driving system based on Monte Carlo Tree Search (MCTS). The model selection component of the architecture infers which models better explain the current behaviours of the other vehicles using maximum likelihood estimation for Bayesian model comparison. The inferred behaviours are then used for MCTS- based control to prevent rollouts on models that are not relevant in the current context. While the use of multiple models allows improved efficiency and higher flexibility, it also introduces identification-related issues, which were solved here using Bayesian machine learning. The results obtained from the performed simulations are presented comparing the proposed MCTS architecture when employing multiple models with a naive model of the other vehicles.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{bu2017towards,

title = {Towards a Human-Robot Teaming System for Exploration of Environment},

author = {Fan Bu and Yan Wu},

url = {https://ieeexplore.ieee.org/document/8336074/ 

https://www.yan-wu.com/wp-content/uploads/2020/05/2017ICOT.pdf},

doi = {10.1109/ICOT.2017.8336074},

isbn = {978-1-5386-3276-5},

year  = {2017},

date = {2017-12-10},

booktitle = {2017 International Conference on Orange Technologies (ICOT)},

pages = {1-6},

publisher = {IEEE},

address = {Singapore},

abstract = {The use of robots has been proposed as promising human assistants in accomplishing tasks that are otherwise too tough or dangerous to be executed by human alone. As robots in different forms have their differentiating strengths, a system that uses a combination of these robots can be more robust and resilient in achieving its mission. For example, UAVs can be used to quickly explore an unknown environment while ground robots can be used to navigate through the environment, clearing obstacles and potential dangers for human to enter. In this work, we design a unifying framework to allow human to control and accomplish an exploration mission using heterogeneous robot types. A prototype system is implemented using the Google Tango to control an UAV and a wheeled robot for the human to perform environment mapping. Preliminary experiments have been carried out to study the feasibility of the system.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{su2017upper,

title = {An Upper Limb Rehabilitation System with Motion Intention Detection},

author = {Yanyu Su and Yan Wu and Yongzhuo Gao and Wei Dong and Yixuan Sun and Zhijiang Du},

url = {https://ieeexplore.ieee.org/document/8273215 

http://www.yan-wu.com/docs/su2017upper.pdf},

doi = {10.1109/ICARM.2017.8273215},

isbn = {978-1-5386-3260-4},

year  = {2017},

date = {2017-08-31},

booktitle = {The 2nd International Conference on Advanced Robotics and Mechatronics (ICARM)},

pages = {510-516},

publisher = {IEEE},

address = {Hefei, China},

abstract = {Upper limb rehabilitation robotic systems have been studied for decades to improve the rehabilitation training of hemiplegic patients. Ideally such robotic systems should be capable of detecting the intentions of the patients and assisting them as needed. In this paper, we integrate a virtual environment and a low-cost motion sensor, the Microsoft Kinect II, into a robotic upper limb rehabilitation system to detect user motion intentions and generate motion commands for a rehabilitation robot. The system requires users to mimic a pre-programmed bimanual motion sequence shown in the virtual environment. A Gaussian Process based predictive controller uses the motion of the unaffected arm and the programmed motion sequence to estimate the motion intentions of the affected arm. We also adopt this controller in Mirror Therapy, a widely-practised therapeutic intervention method. Two preliminary experiments have been conducted to validate the proposed controller and test proper function and safety of the system including a patient case-study.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{ko2016lap,

title = {LAP: A Human-in-the-loop Adaptation Approach for Industrial Robots},

author = {Wilson Kien Ho Ko and Yan Wu and Keng Peng Tee},

url = {http://doi.acm.org/10.1145/2974804.2974805 

http://www.yan-wu.com/docs/ko2016lap.pdf},

doi = {10.1145/2974804.2974805},

isbn = {978-1-4503-4508-8},

year  = {2016},

date = {2016-10-15},

booktitle = {The Fourth International Conference on Human Agent Interaction (HAI)},

pages = {313--319},

publisher = {ACM},

address = {Singapore},

abstract = {In the last few years, a shift from mass production to mass customisation is observed in the industry. Easily reprogrammable robots that can perform a wide variety of tasks are desired to keep up with the trend of mass customisation while saving costs and development time. Learning by Demonstration (LfD) is an easy way to program the robots in an intuitive manner and provides a solution to this problem. In this work, we discuss and evaluate LAP, a three-stage LfD method that conforms to the criteria for the high-mix-low-volume (HMLV) industrial settings. The algorithm learns a trajectory in the task space after which small segments can be adapted on-the-fly by using a human-in-the-loop approach. The human operator acts as a high-level adaptation, correction and evaluation mechanism to guide the robot. This way, no sensors or complex feedback algorithms are needed to improve robot behaviour, so errors and inaccuracies induced by these subsystems are avoided. After the system performs at a satisfactory level after the adaptation, the operator will be removed from the loop. The robot will then proceed in a feed-forward fashion to optimise for speed. We demonstrate this method by simulating an industrial painting application. A KUKA LBR iiwa is taught how to draw an eight figure which is reshaped by the operator during adaptation.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{shen2016investigation,

title = {Investigation of Practical Use of Humanoid Robots in Elderly Care Centres},

author = {Zhuoyu Shen and Yan Wu},

url = {http://doi.acm.org/10.1145/2974804.2980485 

http://www.yan-wu.com/docs/shen2016investigation.pdf},

doi = {10.1145/2974804.2980485},

isbn = {978-1-4503-4508-8},

year  = {2016},

date = {2016-10-15},

booktitle = {The Fourth International Conference on Human Agent Interaction (HAI)},

pages = {63--66},

publisher = {ACM},

address = {Singapore},

abstract = {The global trend of population ageing has magnified the shortage of qualified staff in the elderly care industry. This study evaluates the feasibility and user experience of introducing robots in elderly care services. A robot instructor was being benchmarked against a human instructor administering two types of activities with 41 elderly participants. The results show that robot was more effective and better preferred by users over human instructor on instructing physical exercise, while reaching similar level of effectiveness and user acceptance on information delivery. Additionally, user perception of robots improved after the robot experiment session. These findings could be useful for future design of robots for elderly users and for social robots in general.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{wang2016dynamic,

title = {Dynamic Movement Primitives Plus: For enhanced reproduction quality and efficient trajectory modification using truncated kernels and Local Biases},

author = {Ruohan Wang and Yan Wu and Wei Liang Chan and Keng Peng Tee},

url = {https://ieeexplore.ieee.org/document/7759554 

http://www.yan-wu.com/docs/wang2016dynamic.pdf},

doi = {10.1109/IROS.2016.7759554},

isbn = {978-1-5090-3762-9},

year  = {2016},

date = {2016-10-14},

booktitle = {2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},

pages = {3765-3771},

publisher = {IEEE},

address = {Daejeon, South Korea},

abstract = {Dynamic Movement Primitives (DMPs) are a generic approach for trajectory modeling in an attractor land-scape based on differential dynamical systems. DMPs guarantee stability and convergence properties of learned trajectories, and scale well to high dimensional data. In this paper, we propose DMP+, a modified formulation of DMPs which, while preserving the desirable properties of the original, 1) achieves lower mean square error (MSE) with equal number of kernels, and 2) allows learned trajectories to be efficiently modified by updating a subset of kernels. The ability to efficiently modify learned trajectories i) improves reusability of existing primitives, and ii) reduces user fatigue during imitation learning as errors during demonstration may be corrected later without requiring another complete demonstration. In addition, DMP+ may be used with existing DMP techniques for trajectory generalization and thus complements them. We compare the performance of our proposed approach against DMPs in learning trajectories of handwritten characters, and show that DMP+ achieves lower MSE in position deviation. We demonstrate in a second experiment that DMP+ can efficiently update a learned trajectory by updating only a subset of kernels. The update algorithm achieves modeling accuracy comparable to learning the adapted trajectory with the original DMPs.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}