 |
Description
We present a novel motion planning algorithm for transferring a liquid body from a source to a target container. Our approach uses a receding-horizon optimization strategy that takes into account fluid constraints and avoids collisions. In order to efficiently handle the high-dimensional configuration space of a liquid body, we use system identification to learn its dynamics characteristics using a neural network. We generate the training dataset using stochastic optimization in a transfer-problem-specific search space. The runtime feedback motion planner is used for real-time planning and we observe high success rate in our simulated 2D and 3D fluid transfer benchmarks.
Publication
Feedback Motion Planning for Liquid Transfer using Supervised Learning
Zherong Pan, and Dinesh Manocha
The 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS),
[arXiv:1609.03433] and [PDF]
Video
Download Video: [MP4]
Dataset
Here are the two datasets we use to train the neural network for predicting liquid parameters. We generate 1000 random optimal pouring trajectories using stochastic optimization. The specification of the 1000 problems in TRANSFER dataset can be found here: FluidRLProbs.zip, the output of CMA-ES optimization can be found here: FluidRLData.zip, and finally the extracted features can be found here: FluidRLDataExtracted.zip.
In TRANSFER dataset, the initial liquid speed follows the speed of source container. We use this setting to simplify the problem and avoid spilling. Our second dataset, TRANSFER+SPILL, encourages spilling by setting initial liquid speed to zero. This dataset can be downloaded here: FluidRLProbs_not_follow.zip, FluidRLData_not_follow.zip, FluidRLDataExtracted_not_follow.zip.
The files in FluidRLProbs.zip are self-explanatory ASCII files. The files in FluidRLData.zip are the ASCII format output of the standard CMA-ES implementation. Finally, you could use this code and run main prob*.dat to visualize the binary data in FluidRLDataExtracted.zip using ParaView.