-
Article
A stable data-augmented reinforcement learning method with ensemble exploration and exploitation
Learning from visual observations is a significant yet challenging problem in Reinforcement Learning (RL). Two respective problems, representation learning and task learning, need to solve to infer an optimal ...
-
Article
A Graph-Based Deep Reinforcement Learning Approach to Gras** Fully Occluded Objects
Gras** in cluttered scenes is an important issue in robotic manipulation. The cooperation of gras** and pushing actions based on reinforcement learning is an effective means to obtain the target object whe...
-
Article
Offline reinforcement learning with anderson acceleration for robotic tasks
Offline reinforcement learning (RL) can learn effective policy from a fixed batch of data without interaction. However, the real-world requirements, such as better performance and high sample efficiency, put s...
-
Chapter and Conference Paper
Attention Guided 6D Object Pose Estimation with Multi-constraints Voting Network
For visual-based robotic manipulation, it has always been a challenging task to perform real-time and accurate pose estimation of target objects under cluttered background, illumination variations, occlusion, ...
-
Chapter and Conference Paper
Sample-Efficient Reinforcement Learning Based on Dynamics Models via Meta-policy Optimization
Model-based reinforcement learning (RL) can acquire remarkable sample efficiency, which makes it a suitable choice for applications where experiment data is hard to collect. However, it is difficult to learn a...
-
Article
SOAR Improved Artificial Neural Network for Multistep Decision-making Tasks
Recently, artificial neural networks (ANNs) have been applied to various robot-related research areas due to their powerful spatial feature abstraction and temporal information prediction abilities. Decision-m...
-
Article
Fine semantic map** based on dense segmentation network
This paper proposes a fine semantic map** method using dense segmentation network (DS-Net) to obtain good performance of semantic map** fusion. First, the RGB image and the depth image are used to generate...
-
Chapter and Conference Paper
Multi-robot Formation Control Using Reinforcement Learning Method
Formation is a good example of the research for multi-robot cooperation. Many different ways can be used to accomplish this task, but the main drawbacks of most of these methods are that robots can’t self-lear...
-
Chapter and Conference Paper
Mandarin Voice Conversion Using Tone Codebook Map**
A tone codebook map** method is proposed to obtain a better performance in voice conversion of Mandarin speech than the conventional conversion method which deals mainly with short-time spectral envelopes. T...
-
Chapter and Conference Paper
The Cognitive Behaviors of a Spiking-Neuron Based Classical Conditioning Model
A spiking-neuron based cognitive model with classical conditioning behaviors is proposed. With a reflex arc structure and a reinforcement learning method based on the Hebb rule, the cognitive model possesses t...