Proceedings Article10.1109/ROBIO.2018.8665229
Complex Sequential Tasks Learning with Bayesian Inference and Gaussian Mixture Model
Huiwen Zhang,Xiaoning Han,Wei Zhang,Weijia Zhou +3 more
- 01 Dec 2018
- pp 1927-1934
7
TL;DR: A segmentation algorithm which can segment unstructured demonstrations into movement primitives (MPs) with minimal prior knowledge requirements needs to be proposed and a representation model used to jointly extract tasks constraints from the discovered MPs is proposed.
read more
Abstract: Transferring skills to robots by demonstrations has been extensively researched for decades. However, the majority of the work focuses on individual or low-level task learning. Theories and applications for learning complex sequential tasks are not well-investigated. For this reason, this paper presents a unified top-down framework for complex tasks learning. Specifically, we conclude two critical objectives. First, a segmentation algorithm which can segment unstructured demonstrations into movement primitives (MPs) with minimal prior knowledge requirements needs to be proposed. Second, choosing a representation model used to jointly extract tasks constraints from the discovered MPs. To achieve the first goal, a change-point detection algorithm based on Bayesian inference is used. It can segment unstructured demonstrations online. Then, we propose to model MPs with dynamical system approximated by the Gaussian mixture models (GMMs), which is flexible and powerful in movement representation. Finally, the whole framework is evaluated by an open-and-place task on a real robot. Experiments show the segmentation accuracy can reach to 95.6% and the task can be replayed in new contexts successfully.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
SKID RAW: Skill Discovery From Raw Trajectories
Daniel Tanneberg,Kai Ploeger,Elmar Rueckert,Jan Peters +3 more
- 01 Jul 2021
TL;DR: In this paper, a Bayesian and variational inference based approach is proposed to learn to segment trajectories into reoccurring patterns and the skills to reconstruct these patterns from unlabeled demonstrations without further supervision.
24
SHM deformation monitoring for high-speed rail track slabs and Bayesian change point detection for the measurements
TL;DR: Through posterior distribution information of the Bayesian CPD method, change points which have the highest posterior probability can be identified for the versine deformation of HSR track slabs, which is essential for investigating the cause of the change point and providing maintenance plan in time.
19
Adaptive and intelligent robot task planning for home service: A review
Haizhen Li,Xilun Ding +1 more
TL;DR: In this article , the authors explored primary sources of uncertainty in-depth and summarized three challenges, i.e., reliable planning under uncertain and incomplete information, efficient planning for complex tasks, and scalable planning for task generalization.
12
SKID RAW: Skill Discovery from Raw Trajectories
TL;DR: In this article, a Bayesian and variational inference based approach is proposed to learn to segment trajectories into reoccurring patterns and the skills to reconstruct these patterns from unlabeled demonstrations without further supervision.
8
Learning from demonstration for autonomous generation of robotic trajectory: Status quo and forward-looking overview
Weidong Li,Yuqi Wang,Yuchen Liang,Duc Truong Pham +3 more
7
References
A survey of robot learning from demonstration
TL;DR: A comprehensive survey of robot Learning from Demonstration (LfD), a technique that develops policies from example state to action mappings, which analyzes and categorizes the multiple ways in which examples are gathered, as well as the various techniques for policy derivation.
4.2K
Dynamical movement primitives: Learning attractor models for motor behaviors
TL;DR: Dynamical movement primitives is presented, a line of research for modeling attractor behaviors of autonomous nonlinear dynamical systems with the help of statistical learning techniques, and its properties are evaluated in motor control and robotics.
Is imitation learning the route to humanoid robots
TL;DR: In this article, a review of recent developments in artificial intelligence and neural computation: learning from imitation and the development of humanoid robots is presented. But the authors focus on three important issues: efficient motor learning, the connection between action and perception, and modular motor control in the form of movement primitives.
1.5K
Learning and generalization of motor skills by learning from demonstration
Peter Pastor,Heiko Hoffmann,Tamim Asfour,Stefan Schaal +3 more
- 12 May 2009
TL;DR: A general approach for learning robotic motor skills from human demonstration is provided and how this framework extends to the control of gripper orientation and finger position and the feasibility of this approach is demonstrated.
Learning Stable Nonlinear Dynamical Systems With Gaussian Mixture Models
TL;DR: A learning method is proposed, which is called Stable Estimator of Dynamical Systems (SEDS), to learn the parameters of the DS to ensure that all motions closely follow the demonstrations while ultimately reaching and stopping at the target.