End-to-end learning for high-precision lane keeping via multi-state model

High-precision lane keeping is essential for the future autonomous driving.However,due to the imbalanced and inaccurate datasets collected by human drivers,current end-to-end driving models have poor lane keeping the effect.To improve the precision of lane keeping,this study presents a novel multi-state model-based end-to-end lane keeping method.First,three driving states will be defined:going straight,turning right and turning left.Second,the finite-state machine(FSM)table as well as three kinds of training datasets will be generated based on the three driving states.Instead of collecting the dataset by human drivers,the accurate dataset will be collected by the high-performance path following controller.Third,three sets of parameters based on 3DCNN-LSTM model will be trained for going straight,turning left and turning right,which will be combined with FSM table to form a multi-state model.This study evaluates the multi-state model by testing it on five tracks and recording the lane keeping error.The result shows the multi-state model-based end-to-end method performs the higher precision of lane keeping than the traditional single end-to-end model.

Airway Obstruction Caused by Neck Teratoma in an Infant: Case Report and Literature Review

Teratomas are congenital tumors that derive from all 3 germ layers. Most teratomas of the head and neck present in the pediatricage group. We report a case of a 5-month-old infant who presented with a teratoma showing possible obstruction of the airway. The tumor was successfully resected under general anesthesia. The postoperative course was uneventful, and she was fine at a 3-year follow-up.

Fusing Geometrical and Visual Information via Superpoints for the Semantic Segmentation of 3D Road Scenes

This paper addresses the problem of the semantic segmentation of large-scale 3D road scenes by incorporating the complementary advantages of point clouds and images.To make full use of geometrical and visual information,this paper extracts 3D geometric features from a point cloud using a deep neural network for 3D semantic segmentation and extracts 2D visual features from images using a Convolutional Neural Network(CNN)for 2D semantic segmentation.In order to bridge the features of the two modalities,this paper uses superpoints as an intermediate representation to connect the 2D features with the 3D features.A superpoint-based pooling method is proposed to fuse the features from the two different modalities for joint learning.To evaluate the approach,the paper generates 3D scenes from the Virtual KITTI dataset.The results of the experiments demonstrate that the proposed approach is capable of segmenting large-scale 3D road scenes based on the compact and semantically homogeneous superpoints,and that it achieves considerable improvements over the 2D image and 3D point cloud semantic segmentation methods.

Substantially Enhanced Properties of 2D WS_(2) by High Concentration of Erbium Doping against Tungsten Vacancy Formation

Doping in 2D materials is an important method for tuning of band structures.For this purpose,it is important to develop controllable doping techniques.Here,we demonstrate a substitutional doping strategy by erbium(Er)ions in the synthesis of monolayer WS2 by chemical vapor deposition.Substantial enhancements in photoluminescent and photoresponsive properties are achieved,which indicate a tungsten vacancy suppression mechanism by Er filling.Er ion doping in the monolayer WS2 is proved by X-ray diffraction(XRD)and X-ray photoelectron spectra(XPS),fluorescence,absorption,excitation,and Raman spectra.11.5 at%of the maximum Er concentration is examined by energy dispersive X-ray spectroscopy(EDX).Over 6 times enhancement of intensities with 7.9 nm redshift in peaks are observed from the fluorescent spectra of Er-doped WS2 monolayers compared with their counterparts of the pristine WS2 monolayers,which agrees well with the density functional theory calculations.In addition,over 11 times of dark current,469 times of photocurrents,photoresponsivity,and external quantum efficiency,and two orders of photoresponse speed are demonstrated from the Er-doped WS2 photodetector compared with those of the pristine WS2 device.Our findings prove rare-earth doping in 2D materials,the exciting and ideal technique for substantially enhanced photoluminescent and photoresponsive properties.

Giant renal angiomyolipoma with tuberous sclerosis complex

Though rare,angiomyolipomas(AMLs)are the most common mesenchymal tumors of kidney.In general,AMLs can always be associated with two conditions affecting other organ systems:tuberous sclero-sis complex(TSC)and sporadic lymphangioleiomyoma-tosis.This article presents a case of renal AML occurring in a 14-year-old girl with a definite diagnosis of TSC.She had been diagnosed with TSC at the age of three,and a schedule for close observation was disobeyed.At this time,she underwent a series of examinations:physical examina-tion,ultrasonography,angiography,computed tomography(CT),and magnetic resonance imaging(MRI)scans.The physical examination showed adenoma sebaceum in a butterfly paranasal distribution,and a mass was palpated in the left upper quadrant.There were no neurological deficits.Imaging studies(including ultrasonography,angiography,CT,and MRI)of the abdomen showed a large heterogeneous mass arising from the left kidney.Partial nephrectomy was performed.The pathological diagnosis was hemorrhagic renal AML.No recurrence was found in the three-year follow-up.We concluded that schedule of close observation on patients with TSC should be strictly abided by for the high morbidity of AMLs.The specific risks of renal AMLs are spontaneous hemorrhage and rupture.Treatment options for AMLs include conservative and interventional(total/partial nephrectomy,cryoptherapy,and embolization)treatments.