Inverse Kinematics for Robotic Manipulators via Deep Neural Networks: Experiments and Results
Source abstract
Abstract
This paper explores the application of Deep Neural Networks (DNNs) to solve the Inverse Kinematics (IK) problem in robotic manipulators. The IK problem, crucial for ensuring precision in robotic movements, involves determining joint configurations for a manipulator to reach a desired position or orientation. Traditional methods, such as analytical and numerical approaches, have limitations, especially for redundant manipulators, or involve high computational costs. Recent advances in machine learning, particularly with DNNs, have shown promising results and seem fit for addressing these challenges. This study investigates several DNN architectures, namely Feed-Forward Multilayer Perceptrons (MLPs), Convolutional Neural Networks (CNNs), and Recurrent Neural Networks (RNNs), for solving the IK problem, using the TIAGo robotic arm with seven Degrees of Freedom (DOFs). Different training datasets, normalization techniques, and orientation representations are tested, and custom metrics are introduced to evaluate position and orientation errors. The performance of these models is compared, with a focus on curriculum learning to optimize training. The results demonstrate the potential of DNNs to efficiently solve the IK problem while avoiding issues such as singularities, competing with traditional methods in precision and speed.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Robotic Mechanisms and Dynamics · Robot Manipulation and Learning · Human Pose and Action Recognition
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
RoboticsLab, Systems and Automation Engineering Department, University Carlos III of Madrid, 28911 Leganés, Spain
Funding
Comunidad de Madrid · Instituto de Salud Carlos III · EU structural funds
References (35)
Show source-provided references
1. Kucuk, S., and Bingul, Z. (2006). Robot Kinematics: Forward and Inverse Kinematics, INTECH Open Access.10.5772/5015 ↗
2. Aristidou, A., and Lasenby, J. (2009). Inverse Kinematics: A Review of Existing Techniques and Introduction of a New Fast Iterative Solver, University of Cambridge, Department of Engineering.
3. Calzada-Garcia, A., Victores, J.G., Naranjo-Campos, F.J., and Balaguer, C. (2025). A Review on Inverse Kinematics, Control and Planning for Robotic Manipulators With and Without Obstacles via Deep Neural Networks. Algorithms, 18.10.3390/a18010023 ↗
4. Inverse kinematics techniques in computer graphics: A survey10.1111/cgf.13310 ↗
5. Performance analysis of data-driven techniques for solving inverse kinematics problems
6. Forward and Inverse Kinematics Solution of A 3-DOF Articulated Robotic Manipulator Using Artificial Neural Network
7. Solving inverse kinematics of a 7-DOF manipulator using convolutional neural network
8. Kumhar, H.S., and Kukshal, V. (2022). Inverse kinematic solution for 6-r industrial robot manipulator using convolution neural network. Recent Trends in Product Design and Intelligent Manufacturing Systems, Springer. Lecture Notes in Mechanical Engineering.10.1007/978-981-19-4606-6_84 ↗
9. Shaar, A., and Ghaeb, J.A. (2023, January 22–24). Intelligent Solution for Inverse Kinematic of Industrial Robotic Manipulator Based on RNN. Proceedings of the 2023 IEEE Jordan International Joint Conference on Electrical Engineering and Information Technology (JEEIT), Amman, Jordan.10.1109/JEEIT58638.2023.10185778 ↗
10. Wang, S., Zhang, Y., Chen, S., Xu, M., Yu, Y., and Liu, P. (2023, January 16–18). Inverse Kinematics Analysis of 5-DOF Cooperative Robot Based on Long Short-Term Memory Network. Proceedings of the 2023 IEEE 3rd International Conference on Software Engineering and Artificial Intelligence (SEAI), Xiamen, China.10.1109/SEAI59139.2023.10217721 ↗
11. An analytical and a Deep Learning model for solving the inverse kinematic problem of an industrial parallel robot10.1016/j.cie.2020.106682 ↗
12. Analytical and deep learning approaches for solving the inverse kinematic problem of a high degrees of freedom robotic arm10.1016/j.engappai.2023.106301 ↗
13. Learning constrained distributions of robot configurations with generative adversarial network10.1109/LRA.2021.3068671 ↗
14. A deep learning approach to navigating the joint solution space of redundant inverse kinematics and its applications to numerical IK computations10.1109/ACCESS.2023.3234104 ↗
15. Generative Graphical Inverse Kinematics10.1109/TRO.2024.3521862 ↗
16. Kim, J.T., Park, J., Choi, S., and Ha, S. (2021). Learning robot structure and motion embeddings using graph neural networks. arXiv.
17. Iqdymat, A., and Stamatescu, G. (2025). Reinforcement Learning of a Six-DOF Industrial Manipulator for Pick-and-Place Application Using Efficient Control in Warehouse Management. Sustainability, 17.10.3390/su17020432 ↗
18. Honelign, L., Abebe, Y., Jung, Y.S., Tullu, A., and Jung, S. (2025). Deep Reinforcement Learning-Based Enhancement of Robotic Arm Target-Reaching Performance. Algorithms, 14.10.20944/preprints202501.1917.v1 ↗
19. Kumar, V., Hoeller, D., Sundaralingam, B., Tremblay, J., and Birchfield, S. (October, January 27). Joint space control via deep reinforcement learning. Proceedings of the RSJ International Conference on Intelligent Robots and Systems (IROS), Prague, Czech Republic.
20. Alkhodary, A., and Gur, B. (2023, January 12–14). KineFormer: Solving the Inverse Modeling Problem of Soft Robots Using Transformers. Proceedings of the International Conference on Robotics and Networks, Nanjing, China.10.1007/978-3-031-64495-5_3 ↗
21. Bensadoun, R., Gur, S., Blau, N., and Wolf, L. (2022, January 17–23). Neural inverse kinematic. Proceedings of the International Conference on Machine Learning, Baltimore, MD, USA.
22. Xing, D., Xia, W., and Xu, B. (2022, January 23–27). Kinematics learning of massive heterogeneous serial robots. Proceedings of the 2022 International Conference on Robotics and Automation (ICRA), Philadelphia, PA, USA.10.1109/ICRA46639.2022.9812021 ↗
23. Pages, J., Marchionni, L., and Ferro, F. (2016, January 9–14). TIAGo: The modular robot that adapts to different research needs. Proceedings of the International Workshop on Robot Modularity, IROS, Daejeon, Republic of Korea.
24. Marchionni, L., Pages, J., Adell, J., Capriles, J.R., and Tomé, H. (2013, January 10–14). REEM service robot: How may I help you?. Proceedings of the International Work-Conference on the Interplay Between Natural and Artificial Computation, Mallorca, Spain.10.1007/978-3-642-38637-4_13 ↗
25. Łukawski, B., Montesino, I., Oña, E.D., Victores, J.G., Balaguer, C., and Jardon, A. (2025, January 2–3). Towards the Development of Telepresence Applications with TIAGo and TIAGo++ Using a Virtual Reality Headset. Proceedings of the 2025 IEEE International Conference on Autonomous Robot Systems and Competitions (ICARSC), Funchal, Portugal.10.1109/ICARSC65809.2025.10970173 ↗
26. Koenig, N., and Howard, A. (October, January 28). Design and use paradigms for Gazebo, an open-source multi-robot simulator. Proceedings of the 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (IEEE Cat. No.04CH37566), Sendai, Japan.
27. Chitta, S. (2017). ros_control: A generic and simple control framework for ROS. J. Open Source Softw.10.21105/joss.00456 ↗
28. The choice of scaling technique matters for classification performance10.1016/j.asoc.2022.109924 ↗
29. Deep neural networks in the cloud: Review, applications, challenges and research directions10.1016/j.neucom.2023.126327 ↗
30. Artificial Neural Networks for inverse kinematics problem in articulated robots10.1016/j.engappai.2023.107175 ↗
31. Metrics for 3D rotations: Comparison and analysis10.1007/s10851-009-0161-2 ↗
32. Arithmetic and geometric solutions for average rigid-body rotation10.1016/j.mechmachtheory.2010.05.002 ↗
33. Means and averaging in the group of rotations10.1137/S0895479801383877 ↗
34. Grassmann, R., and Burgner-Kahrs, J. (2019, January 22–26). On the Merits of Joint Space and Orientation Representations in Learning the Forward Kinematics in SE (3). Proceedings of the Robotics: Science and Systems, Freiburg im Breisgau, Germany.
35. A Distance Metric for Finite Sets of Rigid-Body Displacements via the Polar Decomposition10.1115/1.2735640 ↗