Theory Of Wing Sections Including A Summary Of Airfoil Data

Concise deep reinforcement learning obstacle avoidance for underactuated unmanned marine vessels

The development of magnetic field sensors for biomedical applications primarily focuses on equivalent magnetic noise reduction or overall design improvement in order to make them smaller and cheaper while keeping the required values of a limit of detection. One of the cutting-edge topics today is the use of magnetic field sensors for applications such as magnetocardiography, magnetotomography, magnetomyography, magnetoneurography, or their application in point-of-care devices. This introductory review focuses on modern magnetic field sensors suitable for biomedicine applications from a physical point of view and provides an overview of recent studies in this field. Types of magnetic field sensors include direct current superconducting quantum interference devices, search coil, fluxgate, magnetoelectric, giant magneto-impedance, anisotropic/giant/tunneling magnetoresistance, optically pumped, cavity optomechanical, Hall effect, magnetoelastic, spin wave interferometry, and those based on the behavior of nitrogen-vacancy centers in the atomic lattice of diamond.

Ultrasensitive Magnetic Field Sensors for Biomedical Applications.

In this paper, a dynamic model of a wheeled mobile robotic (WMR) system with coupled control input is developed, which will increase the complexity of its tracking control with time-varying advance angle. To deal with this problem, a partial reinforcement learning neural network (PRLNN)-based tracking algorithm is proposed for the WMR systems. The main contributions of the PRLNN adaptive tracking control method is that it is the first control method to introduce the PRLNN adaptive control to the WMR system, which determines to solve the WMR tracking control with the time-varying advance angle. The critic neural network (NN) and action NN adaptive laws for the decoupled controllers are designed using the standard gradient-based adaptation method. According to the Lyapunov stability analysis theorem, the uniform ultimate boundedness of all signals in the WMR system can be guaranteed with the design parameters chose properly, and the tracking error converge to a small compact set nearby zero. A numerical simulation is presented to verify the effectiveness of the proposed control algorithm.

Adaptive Partial Reinforcement Learning Neural Network-Based Tracking Control for Wheeled Mobile Robotic Systems

Otolithic afferents with regular resting discharge respond to gravity or low frequency linear accelerations, and we term these the static or sustained otolithic system. However in the otolithic sense organs there is anatomical differentiation across the maculae and corresponding physiological differentiation. A specialized band of receptors called the striola consists of mainly type I receptors whose hair bundles are weakly tethered to the overlying otolithic membrane. The afferent neurons which form calyx synapses on type I striolar receptors have irregular resting discharge and have low thresholds to high frequency (e.g. 500Hz) bone conducted vibration and air conducted sound. High frequency sound and vibration likely causes fluid displacement which deflects the weakly tethered hair bundles of the very fast type I receptors. Irregular vestibular afferents show phase locking, similar to cochlear afferents, up to stimulus frequencies of kilohertz. We term these irregular afferents the transient system signalling dynamic otolithic stimulation. 500Hz vibration preferentially activates the otolith irregular afferents, since regular afferents are not activated at intensities used in clinical testing whereas irregular afferents have low thresholds. We show how this sustained and transient distinction applies at the vestibular nuclei. The two systems have differential response to vibration and sound, to ototoxic antibiotics, to galvanic stimulation and to natural linear acceleration, and such differential sensitivity allows probing of the two systems. 500Hz vibration which selectively activates irregular otolithic afferents results in stimulus-locked eye movements in animals and humans. The preparatory myogenic potentials for these eye movements are measured in the new clinical test of otolith function – ocular vestibular evoked myogenic potentials. We suggest 500Hz vibration may identify the contribution of the transient system to vestibular controlled responses such as vestibulo-ocular, vestibulo-spinal and vestibulo-sympathetic responses. The prospect of particular treatments targeting one or the other of the transient or sustained systems is now being realized in the clinic by the use of intratympanic gentamicin which preferentially attacks type I receptors. We suggest that it is valuable to view vestibular responses by this sustained-transient distinction.

/pdf/sustained-and-transient-vestibular-systems-a-physiological-1eus11pa4a.pdf

Sustained and Transient Vestibular Systems: A Physiological Basis for Interpreting Vestibular Function.

Recently, real-time obstacle detection by monocular vision exhibits a promising prospect in enhancing the safety of unmanned surface vehicles (USVs). Since the obstacles that may threaten USVs generally appear below the water edge, most existing methods first detect the horizon line and then search for obstacles below the estimated horizon line. However, these methods detect horizon line only using edge or line features, which are susceptible to interference edges from clouds, waves, and land, eventually resulting in poor obstacle detection. To avoid being affected by interference edges, in this paper, we propose a novel horizon line detection method based on semantic segmentation. The method assumes a Gaussian mixture model (GMM) with spatial smoothness constraints to fit the semantic structure of marine images and simultaneously generate a water segmentation mask. The horizon line is estimated from the water boundary points via straight line fitting. Further, inspired by human visual attention mechanisms, an efficient saliency detection method based on background prior and contrast prior is presented to detect obstacles below the estimated horizon line. To reduce false positives caused by sun glitter, waves and foam, the continuity of the adjacent frames is employed to filter the detected obstacles. An extensive evaluation was conducted on a large marine image dataset collected by our ‘Jinghai VIII’ USV. The experimental results show that the proposed method significantly outperformed the recent state-of-the-art marine obstacle method by 22.07% in terms of F-score while running over 24 fps on an NVIDIA GTX1080Ti GPU.

Real-Time Monocular Obstacle Detection Based on Horizon Line and Saliency Estimation for Unmanned Surface Vehicles

Nowadays, a significant portion of road accidents is attributed to dirver’s inattention or dangerous behavior such as call or smoke. Monitoring the driver’s status through computer vision techniques to detect driving errors in real time to avoid vehicle collisions is one of the most important issues in machine vision and Advanced Driver Assistant System(ADAS). In this paper, we propose a novel facial landmark model and a lightweight architecture of Driver Status Monitoring(DSM) by integrating several deep learning models in order to monitor the driver’s inattention caused by drowsiness, fatigue and distraction in real time. With inverted residual block and multi-scale feature map, the proposed facial landmark model achieve 4.90% normalized mean error on common set of 300W dataset. The proposed detection model achieved a mean average precise(mAP@0.5) of 90% on our validation dataset. Several results presented in this paper may provide meaningful insights for the autonomous driving research community and automotive industry for future algorithm development.

A Lightweight Architecture For Driver Status Monitoring Via Convolutional Neural Networks

This paper presents the design and construction of a photovoltaic system (PV system) using dual axis solar tracking and maximum power point tracking (MPPT) controller to improve the efficiency of PV cell output power. The dual axis solar tracking technique is designed to track the sun both the azimuth angle and altitude angle to enable the solar panel perpendicular to the illumination of the sun. The MPPT algorithm is designed to keep the PV system operating at maximum power point, a modified variable step size incremental conductance algorithm has been proposed in this paper to meet the requirements of response speed and dynamic stability. Both the solar tracking and MPPT algorithms are implemented into TMD320F2812 controller. Simulations and practical experiments have been made through Matlab/Simulink and prototype device respectively. The results show that both the solar tracking and MPPT algorithms could improve the efficiency of PV array output power.

An improved PV system based on dual axis solar tracking and MPPT

This paper, wind power generation system experimental platform is established, and wind generator external characteristics are measured. Vertical axis wind turbine system features parameters are measured directly by the experimental platform, so output power, input wind energy, tip speed ratio and wind energy utilization coefficient et al. parameters are calculated by the measured data, and draw the appropriate output characteristic curve and analyze the experimental results.

Experiment validation of vertical axis wind turbine control system based on wind energy utilization coefficient characteristics

With the help of a multiview system, an unmanned vehicle system can better understand the surrounding environment and choose a more accurate and safer path to avoid obstacles. However, due to the interference of the signal or the loss of part of the signal during the acquisition, processing, compression, transmission and decompression of the video image signal, the local area of the image is abnormal, which affects the perception and decision of the system. This article addresses the problems of inaccurate restored images and noise in the restored images by proposing an image restoration method that is applied to a multicamera system. We utilize different perspective images captured by different cameras to assist and constrain the restoration of the damaged image. This method restores the image by combining sample representations and sample distribution models which respectively based on self-encoder reconstruction loss learning and generative adversarial networks. In this method, the infrastructure is a conditional generative adversarial network, the condition is the images that are from the other perspectives, and the generator is a self-encoder structure with cross-layer connection, group convolution and feature map channel exchanged. This method was carried out on a dataset recorded in Zurich using a pair of cameras mounted on a mobile platform. The experimental results demonstrate that the proposed method is superior to the existing methods in terms of mean L1 Loss, mean L2 Loss and the peak signal to noise ratio (PSNR).

Hengyu Li

Papers

Real-Time Monocular Obstacle Detection Based on Horizon Line and Saliency Estimation for Unmanned Surface Vehicles

A Lightweight Architecture For Driver Status Monitoring Via Convolutional Neural Networks

An improved PV system based on dual axis solar tracking and MPPT

Experiment validation of vertical axis wind turbine control system based on wind energy utilization coefficient characteristics

Multiview Scene Image Inpainting Based on Conditional Generative Adversarial Networks