Speech recognition, as the man-machine interface, plays a very important role in field of artificial intelligence. Traditional speech recognition methods are shallow learning structure, and have their limitations. This paper uses Convolution Neural Networks (CNNs) to realize recognition. It is an alternative type neural network that can reduce spectral variation model correlations which exist signals. Besides Back Propagation train network. During whole experiment, group recorded by...

Mobile robot olfaction of toxic and hazardous odor sources is great significance in anti-terrorism, disaster prevention, control scenarios. Aiming at the problems low search efficiency easily falling into a local optimum current source localization strategies, paper proposes adaptive space-aware Infotaxis II algorithm. To improve tracking robots, new reward function designed by considering space information emphasizing exploration behavior robots. Considering enhancement exploratory...

Owing to insufficient illumination of the space station, image information collected by intelligent robot will be degraded, and it not able accurately identify tools required for robot’s on-orbit maintenance. This situation increases difficulty maintenance in a low-illumination environment. We proposes novel enhancement method images under low-illumination, namely, deep learning algorithm based on combination convolutional Wasserstein generative adversarial networks (DC-WGAN) CIELAB color...

Glucose concentration measurement is the basis of noninvasive detection blood glucose concentration. It significant in scientific research. In this study, Near Infrared Spectrum (NIRS) and regression analysis methodology were combined to measure The spectrum solutions was obtained with Fourier Transformed Spectrometer, then data used for analysis. addition, method Paritial Least Squares (PLS) achieve principle components various spectral preprocessing methods discussed. During PLS modeling,...

In order to guarantee the safety of prosthesis users, effective identification stumble situation in walking process is key point. this paper, movement information human body collected by using accelerator and gyroscope installed on prosthesis. And correlation coefficients different signals are obtained via analysis. The two kinds sensors fused Dempster-Shafer evidential theory. experiments show that recognition based D-S theory method accurate. It has great significance for control

Motor function assessment of patients and the elderly is crucial to gait rehabilitation. Accuracy affected by clinician’s experience. To solve problem, this article proposes motor index assess patients. VICON system collects video subjects when they are walking. And original videos pre-processed pixel-based adaptive segmenter extracted convolutional neural network. The kernel entropy component analysis local tangent space alignment reduced dimensions features, obtained. Pearson correlation...

With the widespread use of unmanned aerial vehicles (UAVs), segmentation their images has been widely applied. However, fast semantic high-resolution UAV still faces several challenges: limited hardware processing resources and high real-time requirements platforms, balance between model accuracy efficiency, poor inference ability models for small target classes in images. To address these issues, this paper proposes a lightweight remote sensing image bilateral network (RS-BiseNet) that...

This paper investigates the fixed-time tracking control strategy for a wearable upper-limb exoskeleton under bounded disturbances. A 5 DOF (degree of freedom) model is firstly derived by Lagrange method. Then controller with novel sliding surface constructed which can ensure that both reaching time and settling closed-system are independent initial system conditions. In order to guarantee performance disturbances, nonlinear disturbance observer utilized estimate exact values disturbances...

The formation of the complexation behaviour host molecules water-soluble p-sulfonatocalix[6]arene (C6AS) with amino benzoic acid (ABA) isomers including o-amino (OABA), m-amino (MABA), and p-amino (PABA) – three guest has been studied by fluorescence spectrophotometric nuclear magnetic resonance (NMR) spectroscopy. Experimental conditions concentration C6AS medium acidity were investigated in detail. results showed that forms 1:1 complexes water. Their stability constants determined...

In this paper, we propose a novel framework for face recognition based on manifold ranking algorithm and PCA algorithm. First, is used to explore the relationship among all data points in feature space, then constructs weighted graph with respect intrinsic structure collectively revealed by space data. So images can be measured ranked similarity of image The front r one category retrieval result are selected because no accord between such semantic concept belong same image. Finally, applied...

Inspired by the programmed cell death theory, which is an advanced research achievement in bionomic field, process simulated genetic algorithm presenting man-made control genes. The gene controls balance alive-death of information cells. standard optimized such bionics way. application and its simulation result path planning for mobile robot proves feasibility advantage based on idea control. Some improvements are so gained. However, some adjustments still needed to do this newly presented algorithm.