journal article·2024-06-06Open access

A Trajectory Optimisation-Based Incremental Learning Strategy for Learning from Demonstration

Yuqi Wang, Weidong Li, Yuchen Liang

Applied Sciences· Volume 14· Issue 11

Abstract

The insufficient generalisation capability of the conventional learning from demonstration (LfD) model necessitates redemonstrations. In addition, retraining the model can overwrite existing knowledge, making it impossible to perform previously acquired skills in new application scenarios. These are not economical and efficient. To address the issues, in this study, a broad learning system (BLS) and probabilistic roadmap (PRM) are integrated with dynamic movement primitive (DMP)-based LfD. Three key innovations are proposed in this paper: (1) segmentation and extended demonstration: a 1D-based topology trajectory segmentation algorithm (1D-SEG) is designed to divide the original demonstration into several segments. Following the segmentation, a Gaussian probabilistic roadmap (G-PRM) is proposed to generate an extended demonstration that retains the geometric features of the original demonstration. (2) DMP modelling and incremental learning updating: BLS-based incremental learning for DMP (Bi-DMP) is performed based on the constructed DMP and extended demonstration. With this incremental learning approach, the DMP is capable of self-updating in response to task demands, preserving previously acquired skills and updating them without training from scratch. (3) Electric vehicle (EV) battery disassembly case study: this study developed a solution suitable for EV battery disassembly and established a decommissioned battery disassembly experimental platform. Unscrewing nuts and battery cell removal are selected to verify the effectiveness of the proposed algorithms based on the battery disassembly experimental platform. In this study, the effectiveness of the algorithms designed in this paper is measured by the success rate and error of the task execution. In the task of unscrewing nuts, the success rate of the classical DMP is 57.14% and the maximum error is 2.760 mm. After the optimisation of 1D-SEG, G-PRM, and Bi-DMP, the success rate of the task is increased to 100% and the maximum error is reduced to 1.477 mm.
Source: crossref · Quality: baseline

Research context

OpenAlex topics: Machine Learning and Data Classification · Data Stream Mining Techniques · Intelligent Tutoring Systems and Adaptive Learning

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

Affiliations

School of Transportation and Logistics Engineering, Wuhan University of Technology, Wuhan 430063, China

School of Engineering, University of Birmingham, Birmingham B15 2TT, UK

School of Mechanical Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China

School of Mechanical Engineering, Jiangsu University, Zhenjiang 212013, China

Funding

National Natural Science Foundation of China · International Cooperative Project of the Ministry of Science and Technology of China · Science and Technology Commission of Shanghai Municipality · Engineering and Physical Sciences Research Council, UK · China Scholarship Council · China Postdoctoral Science Foundation

References (39)

Show source-provided references
1. iTP-LfD: Improved task parametrised learning from demonstration for adaptive path generation of cobot10.1016/j.rcim.2020.102109 ↗
2. Optimised Learning from Demonstrations for Collaborative Robots10.1016/j.rcim.2021.102169 ↗
3. Learning Task-Parameterized Skills From Few Demonstrations10.1109/LRA.2022.3150013 ↗
4. Chi, M., Yao, Y., Liu, Y., and Zhong, M. (2019). Learning, Generalization, and Obstacle Avoidance with Dynamic Movement Primitives and Dynamic Potential Fields. Appl. Sci., 9.10.3390/app9081535 ↗
5. A Motion Planning Method for Robots Based on DMPs and Modified Obstacle-Avoiding Algorithm10.1109/TASE.2022.3202351 ↗
6. Residual Learning from Demonstration: Adapting DMPs for Contact-Rich Manipulation10.1109/LRA.2022.3150024 ↗
7. DMPs-based skill learning for redundant dual-arm robotic synchronized cooperative manipulation10.1007/s40747-021-00429-3 ↗
8. Robot Learning System Based on Adaptive Neural Control and Dynamic Movement Primitives10.1109/TNNLS.2018.2852711 ↗
9. Dynamic Skill Learning from Human Demonstration Based on the Human Arm Stiffness Estimation Model and Riemannian DMP10.1109/TMECH.2022.3212707 ↗
10. Arguz, S.H., Ertugrul, S., and Altun, K. (2022, January 17–20). Experimental Evaluation of the Success of Peg-in-Hole Tasks Learned from Demonstration. Proceedings of the 2022 8th International Conference on Control, Decision and Information Technologies (CoDIT), Istanbul, Turkey.10.1109/CoDIT55151.2022.9804111 ↗
11. An imitation learning framework for generating multi-modal trajectories from unstructured demonstrations10.1016/j.neucom.2022.05.076 ↗
12. Smooth and collision-free trajectory generation in cluttered environments using cubic B-spline form10.1016/j.mechmachtheory.2021.104606 ↗
13. Hüppi, M., Bartolomei, L., Mascaro, R., and Chli, M. (2022, January 23–27). T-PRM: Temporal Probabilistic Roadmap for Path Planning in Dynamic Environments. Proceedings of the 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Kyoto, Japan.10.1109/IROS47612.2022.9981739 ↗
14. Dynamic movement primitives in robotics: A tutorial survey10.1177/02783649231201196 ↗
15. Composite dynamic movement primitives based on neural networks for human–robot skill transfer10.1007/s00521-021-05747-8 ↗
16. Noohian, A., Raisi, M., and Khodaygan, S. (2022, January 15–18). A Framework for Learning Dynamic Movement Primitives with Deep Reinforcement Learning. Proceedings of the 2022 10th RSI International Conference on Robotics and Mechatronics (ICRoM), Tehran, Islamic Republic of Iran.10.1109/ICRoM57054.2022.10025346 ↗
17. Kim, W., Lee, C., and Kim, H.J. (2018, January 1–5). Learning and Generalization of Dynamic Movement Primitives by Hierarchical Deep Reinforcement Learning from Demonstration. Proceedings of the 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Madrid, Spain.10.1109/IROS.2018.8594476 ↗
18. Incremental Motor Skill Learning and Generalization From Human Dynamic Reactions Based on Dynamic Movement Primitives and Fuzzy Logic System10.1109/TFUZZ.2021.3136933 ↗
19. A trajectory and force dual-incremental robot skill learning and generalization framework using improved dynamical movement primitives and adaptive neural network control10.1016/j.neucom.2022.11.076 ↗
20. Broad Learning System: An Effective and Efficient Incremental Learning System without the Need for Deep Architecture10.1109/TNNLS.2017.2716952 ↗
21. Trajectory generation with multi-stage cost functions learned from demonstrations10.1016/j.robot.2019.04.006 ↗
22. Online obstacle avoidance path planning and application for arc welding robot10.1016/j.rcim.2022.102413 ↗
23. Park, D.H., Hoffmann, H., Pastor, P., and Schaal, S. (2008, January 3). Movement reproduction and obstacle avoidance with dynamic movement primitives and potential fields. Proceedings of the 2008 8th IEEE-RAS International Conference on Humanoid Robots (Humanoids 2008), Daejeon, Republic of Korea.
24. Dynamical Movement Primitives: Learning Attractor Models for Motor Behaviors10.1162/NECO_a_00393 ↗
25. Fuzzy dynamical system for robot learning motion skills from human demonstration10.1016/j.robot.2023.104406 ↗
26. Ding, G., Liu, Y., Zang, X., Zhang, X., Liu, G., and Zhao, J. (2020). A Task-Learning Strategy for Robotic Assembly Tasks from Human Demonstrations. Sensors, 20.10.3390/s20195505 ↗
27. Si, W., Yue, T., Guan, Y., Wang, N., and Yang, C. (2022, January 20–24). A Novel Robot Skill Learning Framework Based on Bilateral Teleoperation. Proceedings of the 2022 IEEE 18th International Conference on Automation Science and Engineering (CASE), Mexico City, Mexico.10.1109/CASE49997.2022.9926526 ↗
28. Iturrate, I., Roberge, E., Ostergaard, E.H., Duchaine, V., and Savarimuthu, T.R. (2019, January 22–26). Improving the Generalizability of Robot Assembly Tasks Learned from Demonstration via CNN-based Segmentation. Proceedings of the 2019 IEEE 15th International Conference on Automation Science and Engineering (CASE), Vancouver, BC, Canada.10.1109/COASE.2019.8843029 ↗
29. UAV trajectory planning based on bi-directional APF-RRT* algorithm with goal-biased10.1016/j.eswa.2022.119137 ↗
30. Topology-based Smoothing of 2D Scalar Fields with C1-Continuity10.1111/j.1467-8659.2009.01702.x ↗
31. Ichter, B., Schmerling, E., Lee, T.-W.E., and Faust, A. (August, January 31). Learned Critical Probabilistic Roadmaps for Robotic Motion Planning. Proceedings of the 2020 IEEE International Conference on Robotics and Automation (ICRA), Paris, France.10.1109/ICRA40945.2020.9197106 ↗
32. Eiter, T., and Mannila, H. (2024, June 02). Computing Discrete Fréchet Distance. Available online: https://www.researchgate.net/profile/Thomas-Eiter-2/publication/228723178_Computing_Discrete_Frechet_Distance/links/5714d93908aebda86c0d1a7b/Computing-Discrete-Frechet-Distance.pdf.
33. Wang, R., Wu, Y., Chan, W.L., and Tee, K.P. (2016, January 9–14). Dynamic Movement Primitives Plus: For enhanced reproduction quality and efficient trajectory modification using truncated kernels and Local Biases. Proceedings of the 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Daejeon, Republic of Korea.
34. (2024, June 02). Khansari, Billard, TRO 2011, LASA Handwriting Dataset, Stable Estimator of Dynamical Systems (SEDS). 24 March 2015. Available online: https://cs.stanford.edu/people/khansari/download.html#LearningLyapunovFunctions.
35. Avaei, A., Van Der Spaa, L., Peternel, L., and Kober, J. (2023). An Incremental Inverse Reinforcement Learning Approach for Motion Planning with Separated Path and Velocity Preferences. Robotics, 12.10.3390/robotics12020061 ↗
36. Xing, H., Torabi, A., Ding, L., Gao, H., Li, W., Mushahwar, V.K., and Tavakoli, M. (October, January 27). Human-Robot Collaboration for Heavy Object Manipulation: Kinesthetic Teaching of the Role of Wheeled Mobile Manipulator. Proceedings of the 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Prague, Czech Republic.10.1109/IROS51168.2021.9635910 ↗
37. Fog computing and convolutional neural network enabled prognosis for machining process optimization10.1016/j.jmsy.2019.05.003 ↗
38. Computational complexity measures for multi-objective optimization problems10.1016/j.procs.2014.09.077 ↗
39. On the complexity of neural network classifiers: A comparison between shallow and deep architectures10.1109/TNNLS.2013.2293637 ↗