Next Article in Journal
A Novel Approach to the Job Shop Scheduling Problem Based on the Deep Q-Network in a Cooperative Multi-Access Edge Computing Ecosystem
Next Article in Special Issue
A Survey of Deep Convolutional Neural Networks Applied for Prediction of Plant Leaf Diseases
Previous Article in Journal
Significant Geo-Social Group Discovery over Location-Based Social Network
Previous Article in Special Issue
Neural Network for Metal Detection Based on Magnetic Impedance Sensor
Article

Action Generative Networks Planning for Deformable Object with Raw Observations

School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou 510006, China
*
Author to whom correspondence should be addressed.
These authors contributed equally to this work and should be regarded as co-first authors.
Academic Editors: Abdeldjalil Ouahabi, Amir Benzaoui and Sébastien Jacques
Sensors 2021, 21(13), 4552; https://doi.org/10.3390/s21134552
Received: 14 May 2021 / Revised: 25 June 2021 / Accepted: 26 June 2021 / Published: 2 July 2021
Synthesizing plans for a deformable object to transit from initial observations to goal observations, both of which are represented by high-dimensional data (namely “raw” data), is challenging due to the difficulty of learning abstract state representations of raw data and transition models of continuous states and continuous actions. Even though there have been some approaches making remarkable progress regarding the planning problem, they often neglect actions between observations and are unable to generate action sequences from initial observations to goal observations. In this paper, we propose a novel algorithm framework, namely AGN. We first learn a state-abstractor model to abstract states from raw observations, a state-generator model to generate raw observations from states, a heuristic model to predict actions to be executed in current states, and a transition model to transform current states to next states after executing specific actions. After that, we directly generate plans for a deformable object by performing the four models. We evaluate our approach in continuous domains and show that our approach is effective with comparison to state-of-the-art algorithms. View Full-Text
Keywords: AI planning; contrastive learning; action model AI planning; contrastive learning; action model
Show Figures

Figure 1

MDPI and ACS Style

Sheng, Z.; Jin, K.; Ma, Z.; Zhuo, H.-H. Action Generative Networks Planning for Deformable Object with Raw Observations. Sensors 2021, 21, 4552. https://doi.org/10.3390/s21134552

AMA Style

Sheng Z, Jin K, Ma Z, Zhuo H-H. Action Generative Networks Planning for Deformable Object with Raw Observations. Sensors. 2021; 21(13):4552. https://doi.org/10.3390/s21134552

Chicago/Turabian Style

Sheng, Ziqi, Kebing Jin, Zhihao Ma, and Hankz-Hankui Zhuo. 2021. "Action Generative Networks Planning for Deformable Object with Raw Observations" Sensors 21, no. 13: 4552. https://doi.org/10.3390/s21134552

Find Other Styles
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.

Article Access Map by Country/Region

1
Back to TopTop