@article{Dong2024OptimizingRobotic,

title = {Optimizing Robotic Manipulation with Decision-RWKV: A Recurrent Sequence Modeling Approach for Lifelong Learning},

author = {Yujian Dong and Tianyu Wu and Chaoyang Song},

url = {https://doi.org/10.48550/arXiv.2407.16306},

doi = {10.1115/1.4067524},

year  = {2025},

date = {2025-01-27},

urldate = {2025-01-27},

journal = {Journal of Computing and Information Science in Engineering},

volume = {23},

number = {3},

pages = {031004},

abstract = {Models based on the Transformer architecture have seen widespread application across fields such as natural language processing (NLP), computer vision, and robotics, with large language models (LLMs) like ChatGPT revolutionizing machine understanding of human language and demonstrating impressive memory and reproduction capabilities. Traditional machine learning algorithms struggle with catastrophic forgetting, which is detrimental to the diverse and generalized abilities required for robotic deployment. This paper investigates the Receptance Weighted Key Value (RWKV) framework, known for its advanced capabilities in efficient and effective sequence modeling, integration with the decision transformer (DT), and experience replay architectures. It focuses on potential performance enhancements in sequence decision-making and lifelong robotic learning tasks. We introduce the Decision-RWKV (DRWKV) model and conduct extensive experiments using the D4RL database within the OpenAI Gym environment and on the D’Claw platform to assess the DRWKV model's performance in single-task tests and lifelong learning scenarios, showcasing its ability to handle multiple subtasks efficiently. The code for all algorithms, training, and image rendering in this study is open-sourced at https://github.com/ancorasir/DecisionRWKV. },

keywords = {Authorship - Corresponding, JCR Q2, Jour - J. Comput. Inf. Sci. Eng. (JCISE)},

pubstate = {published},

tppubtype = {article}

}

@article{DeGusseme2024BenchmarkingGrasp,

title = {A Dataset and Benchmark for Robotic Cloth Unfolding Grasp Selection: The ICRA 2024 Cloth Competition},

author = {Victor-Louis De Gusseme and Thomas Lips and Remko Proesmans and Julius Hietala and Giwan Lee and Jiyoung Choi and Jeongil Choi and Geon Kim and Phayuth Yonrith and Domen Tabernik and Andrej Gams and Peter Nimac and Matej Urbas and Jon Muhovic and Danijel Skocaj and Matija Mavsar and Hyojeong Yu and Minseo Kwon and Young J. Kim and Yang Cong and Ronghan Chen and Yu Ren and Supeng Diao and Jiawei Weng and Jiayue Liu and Haoran Sun and Linhan Yang and Zeqing Zhang and Ning Guo and Lei Yang and Fang Wan and Chaoyang Song and Jia Pan and Yixiang Jin and Yong A and Jun Shi and Dingzhe Li and Yong Yang and Kakeru Yamasaki and Takumi Kajiwara and Yuki Nakadera and Krati Saxena and Tomohiro Shibata and Chongkun Xia and Kai Mo and Yanzhao Yu and Qihao Lin and Binqiang Ma and Uihun Sagong and JungHyun Choi and JeongHyun Park and Dongwoo Lee and Yeongmin Kim and Myun Joong Hwang and Yusuke Kuribayashi and Naoki Hiratsuka and Daisuke Tanaka and Solvi Arnold and Kimitoshi Yamazaki and Carlos Mateo-Agullo and Andreas Verleysen and Francis wyffels},

url = {https://airo.ugent.be/cloth_competition/

https://doi.org/10.48550/arXiv.2508.16749},

doi = {10.1177/02783649251414885},

year  = {2025},

date = {2025-01-10},

urldate = {2025-01-10},

journal = {The International Journal of Robotics Research},

volume = {0},

number = {0},

pages = {0},

abstract = {Robotic cloth manipulation suffers from a lack of standardized benchmarks and shared datasets for evaluating and comparing different approaches. To address this, we organized the ICRA 2024 Cloth Competition, a unique head-to-head evaluation focused on grasp pose selection for cloth unfolding. Eleven diverse teams competed with a shared dual-arm robot, utilizing our publicly released dataset of real-world robotic cloth unfolding attempts. We expanded this dataset with 176 live evaluation trials, which now encompasses 679 unfolding demonstrations across 34 garments. The competition established a key benchmark and reference for robotic cloth manipulation. Analysis revealed a significant discrepancy between competition performance and prior work, underscoring the importance of independent out-of-the-lab evaluation in robotic cloth manipulation. The resulting dataset, one of the most comprehensive collections of real-world robotic cloth manipulation data, is a valuable resource for developing and evaluating grasp selection methods, particularly for learning-based approaches. It can serve as a foundation for future benchmarks and drive further progress in data-driven robotic cloth manipulation.},

note = {Accepted},

keywords = {Authorship - Co-Author, JCR Q2, Jour - Int. J. Robot. Res. (IJRR)},

pubstate = {forthcoming},

tppubtype = {article}

}

@article{Liu20242024ProprioceptiveLearning,

title = {Proprioceptive Learning with Soft Polyhedral Networks},

author = {Xiaobo Liu and Xudong Han and Wei Hong and Fang Wan and Chaoyang Song},

doi = {10.1177/02783649241238765},

year  = {2024},

date = {2024-10-07},

urldate = {2024-10-07},

journal = {The International Journal of Robotics Research},

volume = {43},

number = {12},

pages = {1916-1935},

abstract = {Proprioception is the “sixth sense” that detects limb postures with motor neurons. It requires a natural integration between the musculoskeletal systems and sensory receptors, which is challenging among modern robots that aim for lightweight, adaptive, and sensitive designs at low costs in mechanical design and algorithmic computation. Here, we present the Soft Polyhedral Network with an embedded vision for physical interactions, capable of adaptive kinesthesia and viscoelastic proprioception by learning kinetic features. This design enables passive adaptations to omni-directional interactions, visually captured by a miniature high-speed motion-tracking system embedded inside for proprioceptive learning. The results show that the soft network can infer real-time 6D forces and torques with accuracies of 0.25/0.24/0.35 N and 0.025/0.034/0.006 Nm in dynamic interactions. We also incorporate viscoelasticity in proprioception during static adaptation by adding a creep and relaxation modifier to refine the predicted results. The proposed soft network combines simplicity in design, omni-adaptation, and proprioceptive sensing with high accuracy, making it a versatile solution for robotics at a low material cost with more than one million use cycles for tasks such as sensitive and competitive grasping and touch-based geometry reconstruction. This study offers new insights into vision-based proprioception for soft robots in adaptive grasping, soft manipulation, and human-robot interaction.},

keywords = {Authorship - Corresponding, JCR Q2, Jour - Int. J. Robot. Res. (IJRR)},

pubstate = {published},

tppubtype = {article}

}

@article{Wang2022DeepClaw2.0,

title = {DeepClaw 2.0: A Data Collection Platform for Learning Human Manipulation},

author = {Haokun Wang and Xiaobo Liu and Nuofan Qiu and Ning Guo and Fang Wan and Chaoyang Song},

url = {Sec. Computational Intelligence in Robotics},

doi = {10.3389/frobt.2022.787291},

year  = {2022},

date = {2022-03-15},

urldate = {2022-03-15},

journal = {Frontiers in Robotics and AI},

volume = {9},

pages = {787291},

abstract = {Besides direct interaction, human hands are also skilled at using tools to manipulate objects for typical life and work tasks. This paper proposes DeepClaw 2.0 as a low-cost, open-sourced data collection platform for learning human manipulation. We use an RGB-D camera to visually track the motion and deformation of a pair of soft finger networks on a modified kitchen tong operated by human teachers. These fingers can be easily integrated with robotic grippers to bridge the structural mismatch between humans and robots during learning. The deformation of soft finger networks, which reveals tactile information in contact-rich manipulation, is captured passively. We collected a comprehensive sample dataset involving five human demonstrators in ten manipulation tasks with five trials per task. As a low-cost, open-sourced platform, we also developed an intuitive interface that converts the raw sensor data into state-action data for imitation learning problems. For learning-by-demonstration problems, we further demonstrated our dataset’s potential by using real robotic hardware to collect joint actuation data or using a simulated environment when limited access to the hardware.},

keywords = {Authorship - Corresponding, JCR Q2, Jour - Front. Robot. AI. (FROBT)},

pubstate = {published},

tppubtype = {article}

}

@article{Jiang2021RigidSoft,

title = {Rigid-Soft Interactive Design of a Lobster-Inspired Finger Surface for Enhanced Grasping Underwater},

author = {Haiyang Jiang and Xudong Han and Yonglin Jing and Ning Guo and Fang Wan and Chaoyang Song},

url = {Sec. Soft Robotics},

doi = {10.3389/frobt.2021.787187},

year  = {2021},

date = {2021-12-22},

urldate = {2021-12-22},

issuetitle = {Section Soft Robotics},

journal = {Frontiers in Robotics and AI},

volume = {8},

pages = {787187},

abstract = {Bio-inspirations from soft-bodied animals provide a rich design source for soft robots, yet limited literature explored the potential enhancement from rigid-bodied ones. This paper draws inspiration from the tooth profiles of the rigid claws of the Boston Lobster, aiming at an enhanced soft finger surface for underwater grasping using an iterative design process. The lobsters distinguish themselves from other marine animals with a pair of claws capable of dexterous object manipulation both on land and underwater. We proposed a 3-stage design iteration process that involves raw imitation, design parametric exploration, and bionic parametric exploitation on the original tooth profiles on the claws of the Boston Lobster. Eventually, 7 finger surface designs were generated and fabricated with soft silicone. We validated each design stage through many vision-based robotic grasping attempts against selected objects from the Evolved Grasping Analysis Dataset (EGAD). Over 14,000 grasp attempts were accumulated on land (71.4%) and underwater (28.6%), where we selected the optimal design through an on-land experiment and further tested its capability underwater. As a result, we observed an 18.2% improvement in grasping success rate at most from a resultant bionic finger surface design, compared with those without the surface, and a 10.4% improvement at most compared with the validation design from the previous literature. Results from this paper are relevant and consistent with the bioresearch earlier in 1911, showing the value of bionics. The results indicate the capability and competence of the optimal bionic finger surface design in an amphibious environment, which can contribute to future research in enhanced underwater grasping using soft robots.},

keywords = {Authorship - Corresponding, JCR Q2, Jour - Front. Robot. AI. (FROBT)},

pubstate = {published},

tppubtype = {article}

}

@article{Wang2021VolumetricallyEnhanced,

title = {Volumetrically Enhanced Soft Actuator with Proprioceptive Sensing},

author = {Baiyue Wang and Weijie Guo and Shihao Feng and Hongdong Yi and Fang Wan and Chaoyang Song},

doi = {10.1109/LRA.2021.3072859},

year  = {2021},

date = {2021-07-01},

urldate = {2021-07-01},

journal = {IEEE Robotics and Automation Letters},

volume = {6},

number = {3},

issue = {July},

pages = {5284-5291},

abstract = {Soft robots often show a superior power-to-weight ratio using highly compliant, light-weight material, which leverages various bio-inspired body designs to generate desirable deformations for life-like motions. In this letter, given that most material used for soft robots is light-weight in general, we propose a volumetrically enhanced design strategy for soft robots, providing a novel design guideline to govern the form factor of soft robots. We present the design, modeling, and optimization of a volumetrically enhanced soft actuator (VESA) with linear and rotary motions, respectively, achieving superior force and torque output, linear and rotary displacement, and overall extension ratio per unit volume. We further explored VESA's proprioceptive sensing capability by validating the output force and torque through analytical modeling and experimental verification. Our results show that the volumetric metrics hold the potential to be used as a practical design guideline to optimize soft robots’ engineering performance.},

keywords = {Authorship - Corresponding, JCR Q1, JCR Q2, Jour - IEEE Robot. Autom. Lett. (RA-L), Special - Dual-Track},

pubstate = {published},

tppubtype = {article}

}

@article{Wan2020FlangeBased,

title = {Flange-Based Hand-Eye Calibration Using a 3D Camera with High Resolution, Accuracy, and Frame Rate},

author = {Fang Wan and Chaoyang Song},

doi = {10.3389/frobt.2020.00065},

year  = {2020},

date = {2020-05-29},

urldate = {2020-05-29},

journal = {Frontiers in Robotics and AI},

volume = {7},

pages = {65},

abstract = {Point cloud data provides three-dimensional (3D) measurement of the geometric details in the physical world, which relies heavily on the quality of the machine vision system. In this paper, we explore the potentials of a 3D scanner of high quality (15 million points per second), accuracy (up to 0.150 mm), and frame rate (up to 20 FPS) during static and dynamic measurements of the robot flange for direct hand-eye calibration and trajectory error tracking. With the availability of high-quality point cloud data, we can exploit the standardized geometric features on the robot flange for 3D measurement, which are directly accessible for hand-eye calibration problems. In the meanwhile, we tested the proposed flange-based calibration methods in a dynamic setting to capture point cloud data in a high frame rate. We found that our proposed method works robustly even in dynamic environments, enabling a versatile hand-eye calibration during motion. Furthermore, capturing high-quality point cloud data in real-time opens new doors for the use of 3D scanners, capable of detecting sensitive anomalies of refined details even in motion trajectories. Codes and sample data of this calibration method is provided at Github (https://github.com/ancorasir/flange_handeye_calibration).},

keywords = {Authorship - Corresponding, JCR Q2, Jour - Front. Robot. AI. (FROBT)},

pubstate = {published},

tppubtype = {article}

}

@article{HolttaOtto2018TheCharacteristics,

title = {The Characteristics of Innovative, Mechanical Products—10 Years Later},

author = {Katja Hölttä-Otto and Kevin Otto and Chaoyang Song and Jianxi Luo and Timothy Li and Carolyn C. Seepersad and Warren Seering},

doi = {10.1115/1.4039851},

year  = {2018},

date = {2018-08-01},

urldate = {2018-08-01},

journal = {Journal of Mechanical Design},

volume = {140},

number = {8},

issue = {August},

pages = {084501},

abstract = {Ten years prior to this paper, innovative mechanical products were analyzed and found to embody multiple innovation characteristics—an average of two more than competing products in the marketplace. At the time, it was not known whether these products would be successful over time and whether the number or type of innovation characteristics would be related with success. In this work, products from the previous study were categorized into well- and under-adopted products. Also, each product was categorized according to the type of firm that launched it: a new venture or an established firm. The innovative products enjoyed a success rate of 77% on average. The success was not dependent on the number or type of innovation characteristics embodied by the product. However, products developed in new ventures embody, on average, one more innovation characteristic and enjoy a slightly higher success rate than those launched by established firms.},

keywords = {Authorship - Co-Author, JCR Q2, Jour - J. Mech. Des. (JMD)},

pubstate = {published},

tppubtype = {article}

}

@article{Wan2018ANeural,

title = {A Neural Network with Logical Reasoning based on Auxiliary Inputs},

author = {Fang Wan and Chaoyang Song},

url = {Sec. Computational Intelligence in Robotics},

doi = {10.3389/frobt.2018.00086},

year  = {2018},

date = {2018-07-30},

urldate = {2018-07-30},

journal = {Frontiers in Robotics and AI},

volume = {5},

pages = {86},

abstract = {This paper describes a neural network design using auxiliary inputs, namely the indicators, that act as the hints to explain the predicted outcome through logical reasoning, mimicking the human behavior of deductive reasoning. Besides the original network input and output, we add an auxiliary input that reflects the specific logic of the data to formulate a reasoning process for cross-validation. We found that one can design either meaningful indicators, or even meaningless ones, when using such auxiliary inputs, upon which one can use as the basis of reasoning to explain the predicted outputs. As a result, one can formulate different reasonings to explain the predicted results by designing different sets of auxiliary inputs without the loss of trustworthiness of the outcome. This is similar to human explanation process where one can explain the same observation from different perspectives with reasons. We demonstrate our network concept by using the MNIST data with different sets of auxiliary inputs, where a series of design guidelines are concluded. Later, we validated our results by using a set of images taken from a robotic grasping platform. We found that our network enhanced the last 1–2% of the prediction accuracy while eliminating questionable predictions with self-conflicting logics. Future application of our network with auxiliary inputs can be applied to robotic detection problems such as autonomous object grasping, where the logical reasoning can be introduced to optimize robotic learning.},

keywords = {Authorship - Corresponding, JCR Q2, Jour - Front. Robot. AI. (FROBT)},

pubstate = {published},

tppubtype = {article}

}

@article{Song2014KinematicStudy,

title = {Kinematic Study of the Original and Revised General Line-Symmetric Bricard 6R Linkages},

author = {Chaoyang Song and Yan Chen and I-Ming Chen},

doi = {10.1115/1.4026339},

year  = {2014},

date = {2014-08-01},

urldate = {2014-08-01},

journal = {Journal of Mechanisms and Robotics},

volume = {6},

number = {3},

issue = {August},

pages = {031002},

abstract = {In this paper, the solutions to closure equations of the original general line-symmetric Bricard 6R linkage are derived through matrix method. Two independent linkage closures are found in the original general line-symmetric Bricard 6R linkage, which are line-symmetric in geometry conditions, kinematic variables and spatial configurations. The revised general line-symmetric Bricard 6R linkage differs from the original linkage with negatively equaled offsets on the opposite joints. Further analysis shows that the revised linkage is equivalent to the original linkage with different setups on joint axis directions. As a special case of the general line-symmetric Bricard linkage, the line-symmetric octahedral Bricard linkage also has two forms in the closure equations. Their closure curves are not independent but joined into a full circle. This work offers an in-depth understanding about the kinematics of the general line-symmetric Bricard linkages.},

keywords = {Authorship - First Author, JCR Q2, Jour - J. Mech. Robot. (JMR)},

pubstate = {published},

tppubtype = {article}

}

@article{Song2012AFamily,

title = {A Family of Mixed Double-Goldberg 6R Linkages},

author = {Chaoyang Song and Yan Chen},

doi = {10.1098/rspa.2011.0345},

year  = {2012},

date = {2012-03-08},

urldate = {2012-03-08},

journal = {Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences},

volume = {468},

number = {2139},

pages = {871-890},

abstract = {A complete family of double-Goldberg 6R linkages is reported in this article by combining a subtractive Goldberg 5R linkage and a Goldberg 5R linkage through the common link-pair or common Bennett-linkage method. A number of distinct types of overconstrained linkages are built, namely the mixed double-Goldberg 6R linkages. They all have one degree of freedom and their closure equations are derived in detail. One of them degenerates into a Goldberg 5R linkage. From the construction process and geometry conditions, the corresponding relationship between the newly found 6R linkages and the double-Goldberg 6R linkages, constructed from two Goldberg 5R linkages or two subtractive Goldberg 5R linkages, has been established.},

keywords = {Authorship - First Author, JCR Q2, Jour - Proc. Math. Phys. Eng. Sci. (RoyalSocA)},

pubstate = {published},

tppubtype = {article}

}