RRT-CoLearn: Towards Kinodynamic Planning Without Numerical Trajectory Optimization
| العنوان: | RRT-CoLearn: Towards Kinodynamic Planning Without Numerical Trajectory Optimization |
|---|---|
| المؤلفون: | Thomas M. Moerland, Wouter Wolfslag, Martijn Wisse, Mukunda Bharatheesha |
| المصدر: | IEEE Robotics and Automation Letters. 3:1655-1662 |
| بيانات النشر: | Institute of Electrical and Electronics Engineers (IEEE), 2018. |
| سنة النشر: | 2018 |
| مصطلحات موضوعية: | FOS: Computer and information sciences, 0209 industrial biotechnology, Mathematical optimization, Control and Optimization, Test data generation, Computer science, Computation, Biomedical Engineering, 02 engineering and technology, Computer Science::Robotics, Computer Science - Robotics, 020901 industrial engineering & automation, Artificial Intelligence, FOS: Mathematics, 0202 electrical engineering, electronic engineering, information engineering, Mathematics - Optimization and Control, Mechanical Engineering, Supervised learning, Sampling (statistics), Trajectory optimization, Optimal control, Computer Science Applications, Human-Computer Interaction, Kinodynamic planning, Optimization and Control (math.OC), Control and Systems Engineering, Metric (mathematics), 020201 artificial intelligence & image processing, Computer Vision and Pattern Recognition, Robotics (cs.RO) |
| الوصف: | Sampling-based kinodynamic planners, such as Rapidly-exploring Random Trees (RRTs), pose two fundamental challenges: computing a reliable (pseudo-)metric for the distance between two randomly sampled nodes, and computing a steering input to connect the nodes. The core of these challenges is a Two Point Boundary Value Problem, which is known to be NP-hard. Recently, the distance metric has been approximated using supervised learning, reducing computation time drastically. The previous work on such learning RRTs use direct optimal control to generate the data for supervised learning. This paper proposes to use indirect optimal control instead, because it provides two benefits: it reduces the computational effort to generate the data, and it provides a low dimensional parametrization of the action space. The latter allows us to learn both the distance metric and the steering input to connect two nodes. This eliminates the need for a local planner in learning RRTs. Experimental results on a pendulum swing up show 10-fold speed-up in both the offline data generation and the online planning time, leading to at least a 10-fold speed-up in the overall planning time. Comment: This paper is currently under review at IEEE RA-L |
| تدمد: | 2377-3774 |
| DOI: | 10.1109/lra.2018.2801470 |
| URL الوصول: | https://explore.openaire.eu/search/publication?articleId=doi_dedup___::d417f11373a20496227804941e1b810c https://doi.org/10.1109/lra.2018.2801470 |
| Rights: | OPEN |
| رقم الانضمام: | edsair.doi.dedup.....d417f11373a20496227804941e1b810c |
| قاعدة البيانات: | OpenAIRE |
| تدمد: | 23773774 |
|---|---|
| DOI: | 10.1109/lra.2018.2801470 |