Luefeng Chen

Softmax regression based deep sparse autoencoder network for facial emotion recognition in human-robot interaction

Deep neural network (DNN) has been used as a learning model for modeling the hierarchical architecture of human brain. However, DNN suffers from problems of learning efficiency and computational complexity. To address these problems, deep sparse autoencoder network (DSAN) is used for learning facial features, which considers the sparsity of hidden units for learning high-level structures. Meanwhile, Softmax regression (SR) is used to classify expression feature. In this paper, Softmax regression-based deep sparse autoencoder network (SRDSAN) is proposed to recognize facial emotion in human-robot interaction. It aims to handle large data in the output of deep learning by using SR, moreover, to overcome local extrema and gradient diffusion problems in the training process, the overall network weights are fine-tuned to reach the global optimum, which makes the entire depth of the neural network more robust, thereby enhancing the performance of facial emotion recognition. Results show that the average recognition accuracy of SRDSAN is higher than that of the SR and the convolutional neural network. The preliminarily application experiments are performed in the developing emotional social robot system (ESRS) with two mobile robots, where emotional social robot is able to recognize emotions such as happiness and angry.

A facial expression emotion recognition based human-robot interaction system

A facial expression emotion recognition based human-robot interaction U+0028 FEER-HRI U+0029 system is proposed, for which a four-layer system framework is designed. The FEER-HRI system enables the robots not only to recognize human emotions, but also to generate facial expression for adapting to human emotions. A facial emotion recognition method based on 2D-Gabor, uniform local binary pattern U+0028 LBP U+0029 operator, and multiclass extreme learning machine U+0028 ELM U+0029 classifier is presented, which is applied to real-time facial expression recognition for robots. Facial expressions of robots are represented by simple cartoon symbols and displayed by a LED screen equipped in the robots, which can be easily understood by human. Four scenarios, i.e., guiding, entertainment, home service and scene simulation are performed in the human-robot interaction experiment, in which smooth communication is realized by facial expression recognition of humans and facial expression generation of robots within 2 seconds. As a few prospective applications, the FEER-HRI system can be applied in home service, smart home, safe driving, and so on.

A quick control strategy based on hybrid intelligent optimization algorithm for planar n-link underactuated manipulators

Abstract This paper presents a quick two-stage position control strategy based on a hybrid intelligent optimization algorithm for a planar n-link underactuated manipulator with a passive first joint. In stage 1, the system is directly reduced to a planar virtual Acrobot by controlling n-2 active links to their target angles. A hybrid intelligent optimization algorithm, which includes genetic algorithm (GA) and particle swarm optimization algorithm (PSO), is used to solve all link target angles according to the target position of the system. By coordinating GA and PSO, the hybrid intelligent optimization algorithm ensures that all link target angles, the angle of the passive link at the end of stage 1, and the initial angle of the active link of the planar virtual Acrobot meet the angle constraint of the planar virtual Acrobot. So, the position control objective of the planar n-link underactuated manipulator is realized by controlling the active link of the planar virtual Acrobot to its target angle in stage 2.