APPLICATION OF ARTIFICIAL INTELLIGENCE FOR PLANNING TRAJECTORIES OF AN INDUSTRIAL MANIPULATOR ROBOT IN CONSTRAINED CONDITIONS
DOI:
https://doi.org/10.55287/22275398_2026_58_58-66Keywords:
industrial manipulator robot, trajectory planning, artificial intelligence, deep learning, RRT, constrained environments, collision avoidance, ROSAbstract
The article addresses the urgent task of increasing the autonomy and efficiency of industrial manipulator robots operating in limited space cluttered with obstacles. An analysis of classical trajectory planning methods (RRT, PRM, potential fields) is carried out, and their key shortcomings in constrained environments are identified: high computational complexity, susceptibility to local minima, and insufficient adaptability to dynamic environmental changes. The scientific novelty of the research lies in the development of a hybrid AI-RRT algorithm that combines a modified rapidly-exploring random tree (RRT*) method with a deep neural network (DNN) for predicting heuristics and correcting trajectories in real time. The algorithm is supplemented by a scene semantic analysis system based on a convolutional neural network (CNN), which classifies obstacle types (static rigid bodies, deformable objects, zones with adjustable clearance) to optimize bypass maneuvers. This allows for assigning different penalty coefficients in the path cost function depending on the obstacle category, thereby facilitating the search for more rational and safe routes.
Practical significance is confirmed by the results of simulation modeling in the ROS and Gazebo environment for a UR5 manipulator. Compared to the baseline RRT*, the proposed hybrid algorithm demonstrated a reduction in average planning time by 42%, a decrease in final trajectory length by 18%, and an increase in the success rate of finding a collision-free path in complex scenarios to 98.5%. The developed approach enables effective planning in high-clutter environments and provides a foundation for creating adaptive robotic systems for assembly, welding, and logistics.
References
1. Elbanhawi M., Simic M. Sampling-based robot motion planning: A review. IEEE Access. 2014;2:56–77.
2. Karaman S., Frazzoli E. Optimal kinodynamic motion planning using incremental sampling-based methods. Proceedings of the IEEE Conference on Decision and Control. 2010:7681–7687.
3. Kuffner J. J., LaValle S. M. RRT-connect: An efficient approach to single-query path planning. Proceedings of the IEEE International Conference on Robotics and Automation. 2000:995–1001.
4. Gammell J. D., Srinivasa S. S., Barfoot T. D. Informed RRT*: Optimal sampling-based path planning focused via direct sampling of an admissible ellipsoidal heuristic. Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems. 2014:2997–3004.
5. Chamzas C., Shrivastava A., Kavraki L. E. Using local experiences for global motion planning. Proceedings of the IEEE International Conference on Robotics and Automation. 2019:8606–8612.
6. Ichter B., Schmerling E., Lee T.-W. E., Faust A. Learned critical probabilistic roadmaps for robotic motion planning. Proceedings of the IEEE International Conference on Robotics and Automation. 2020:9535–9541.
7. Li Y., Wu J., Chen X., Lu J. Path planning for robotic manipulators using deep reinforcement learning. IEEE Access. 2021;9:7632–7641.
8. Qureshi A. H., Simeonov A., Bency M. J., Yip M. C. Motion planning networks. Proceedings of the IEEE International Conference on Robotics and Automation. 2019:2118–2124.
9. Chiang H.-T. L., Faust A., Fiser M., Francis A. Learning navigation behaviors end-to-end with AutoRL. IEEE Robotics and Automation Letters. 2019;4(2):2007–2014.
10. Chen Y., Zhou B., Yang Y., Lu W., Li H. Semantic-aware motion planning for robot manipulation in crowded environments. IEEE Transactions on Industrial Informatics. 2022;18(5):2986–2995.
11. Shmal’ko E. Yu., Petrov A. A., Ivanov D. S. Identifikatsiya neyrosetevoy modeli robota dlya resheniya zadachi optimal’nogo upravleniya [Identification of a neural network model of a robot for optimal control]. Informatika i avtomatizatsiya. 2021;20(6):1234–1278. (In Russian).
12. Zhang F., Leitão J., Stolt A., Linderoth M., Robertsson A., Johansson R. Geometric reinforcement learning for path planning of UAVs. Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems. 2020:5638–5645.
13. Janson L., Schmerling E., Pavone M. Monte Carlo motion planning for robot trajectory optimization under uncertainty. International Journal of Robotics Research. 2018;37(13–14):1719–1753.
14. Xie L., Wang S., Rosa S., Markham A., Trigoni N. Learning with training wheels: Speeding up training with a simple controller for deep reinforcement learning. Proceedings of the IEEE International Conference on Robotics and Automation. 2018:6276–6283.
15. Everett M., Chen Y. F., How J. P. Motion planning among dynamic, decision-making agents with deep reinforcement learning. Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems. 2018:3052–3059.
16. Wang C., Wang J., Zhang X., Zhang Y. Autonomous vehicle path planning based on deep reinforcement learning. IEEE Transactions on Intelligent Transportation Systems. 2022;23(7):7945–7957.
17. Aljalbout E., Chen J., Ritt K., Ulbrich S., Lienkamp M. Learning to drive from a world on rails. Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems. 2021:3492–3499.
18. Li S., Zhang Y., Li H., Han L. A survey of deep learning for robot motion planning. IEEE/CAA Journal of Automatica Sinica. 2021;8(2):303–324.
19. Ganesan S., Natarajan S. K., Thondiyath A. G-RRT: Goal-biased sampling-based RRT algorithm for mobile robot navigation with improved convergence rate. Proceedings of Advances in Robotics. 2021:1–6.
20. Jiang L., Liu S., Cui Y., Jiang H. Trajectory planning of robotic manipulators in a complex obstacle environment based on improved RRT algorithm. IEEE/ASME Transactions on Mechatronics. 2022;27(6):4774–4785.
