Rapid increase of traffic volume on urban roads over time has changed the traffic scenario globally. It has also increased the ratio of road accidents that can be severe and fatal in the worst case. ...To improve traffic safety and its management on urban roads, there is a need for prediction of severity level of accidents. Various machine learning models are being used for accident prediction. In this study, tree based ensemble models (Random Forest, AdaBoost, Extra Tree, and Gradient Boosting) and ensemble of two statistical models (Logistic Regression Stochastic Gradient Descent) as voting classifiers are compared for prediction of road accident severity. Significant features that are strongly correlated with the accident severity are identified by Random Forest. Analysis proved Random Forest as the best performing model with highest classification results with 0.974 accuracy, 0.954 precision, 0.930 recall and 0.942 F-score using 20 most significant features as compared to other techniques classification of road accidents severity.
The first goal of this project is to investigate the beneficial sessions for each student in an e-learning system. The second goal is to explore the relationships between student session difficulty, ...workload, engagement and loyalty based on the session outcome. The most difficult problem faced by e-learning instructors is finding which course sessions or what course materials are most beneficial to their students during a course. When instructors do receive insufficient feedback concerning a course session or week, the result can be that students fail the course, drop out, or receive a lower grade on the final exam. In this study, we used machine learning (ML) algorithms and regression analysis to identify beneficial sessions based on students' workload, engagement, difficulty and loyalty during the course. The results revealed that strong relationships exist between the input student features (engagement, difficulty, workload and loyalty) and the session scores. In addition, the results show that deep learning and random forest models are appropriate ML algorithms for predicting beneficial sessions.
In this paper, a low cost, high efficient, ultra-wideband (UWB), S-shaped monopole antenna is designed and implemented. It can present meander new logical structure antenna which uses a coupling ...mechanism to improve the bandwidth of operating frequency. The total size of the antenna is 41×39×1.57 mm 3 . The antenna is operating in S/C/X bands with a wide bandwidth of 12.55 GHz (2.9-15.45 GHz) and suitable for wireless local area network (WLAN), Worldwide Interoperability for Microwave Access (WiMAX), and for ultra-wideband (UWB) applications. It is a multifunctional antenna that can be used to reduce the use of multiple antennas in a wireless communication network. Monopole antenna is used because it has multiple applications for UWB. The antenna is simple in structure, wide bandwidth, acceptable omnidirectional properties, ease of fabrication, light weight and a low cost. Waveguide transformer is used in the feeding line of the antenna. The simulated and measured results show that the antenna has 137% fractional bandwidth and 5.96 dBi gains are achieved at 9 GHz in the operating frequencies.
In this paper, a compact, mechanically rugged, DC grounded Null-filled horizontally polarized antenna having low gain ripple in the omnidirectional pattern is proposed for UAV Ground Command Station. ...The antenna is based on the coaxial cylinder structures, which consists of oversize coaxial cylinders, slots and their feeding assembly and a matching unit. The horizontal polarization property is realized by four vertical slots surrounding the axis of the coaxial cylinder in the outer conductor. The antenna is dc ground which eliminates the electrostatic discharge ESD problems. Simulation results show that the antenna bandwidth is 8.1% (2.56 GHz to 2.78 GHz respectively, gain variation in the horizontal plane is less than ±1 dB while the cross-polarization level is below 15 dB over the whole bandwidths.
In this research, a comprehensive numerical investigation of high-pressure ratio turbocharger centrifugal compressor impeller has been carried out. The objective is to study the influence of varying ...blade meridional contour on the total pressure ratio and isentropic efficiency from stall to choke using (3D) numerical simulations. The transonic SRV2 compressor developed by DLR (German Aerospace Center) has been used as the test case. Numerical simulations have been executed using Reynolds Averaged Navier Stokes (RANS) based k-ɛ model to predict turbulence. Y plus is kept 35 for the structured mesh near the boundaries. In first part, calculations were carried out for design conditions; speed of 50,000 1/min and mass flow rate of 2.546 kg/s to compare the results of ANSYS CFX with experimental data in the design procedure. Numerical simulations overpredicts the experimental data by 9 percent. In second part, blade meridional contour and blade thickness have been changed to see its effect on performance parameters of centrifugal compressor at high rotational speed. Being the important parameter, appropriate blade meridional design can improve pressure ratio and efficiency of centrifugal compressor. The novel design improved the stall margin by 1.1 percent while operating range from stall to choke has been upgraded by 1 percent.
Due to the high stiffness-to-weight and strength-to-weight ratios, Carbon Fiber Reinforced Polymer (CFRP) based composite materials are gaining increasing attention and growingly applied in modern ...aircraft structures. As one of the load-carrying members, the stiffened panels are widely used in the aircraft structures, such as: fuselage, wing and tail. These thin walled panels will buckle under the compressive or shear loading, which is a major consideration in the design process. However, in order to fully exploit the large post buckling reserves of CFRP stiffened panels, their damage behaviour and failure modes must be well understood. Being the anisotropic nature of strength/stiffness properties of CFRP structures and highly nonlinear nature of post buckling analysis brings unique challenges for the design and analysis. In this research, a parametric model based on the Global Local Analysis has been developed to determine damage tolerance of the CFRP panels. A Progressive Damage Model and Bilinear Cohesive Zone Model (CZM) with linear softening characterization have been used for the intralaminar and interlaminar damage (skin/stiffener de-bond) analysis respectively. The model has been validated by analysing Hat and T stringer stiffened Single Stringer Compression Specimens (SSCS) and a fair comparison has been found with literature available numerical and experimental data. Meanwhile, a simplified model has been proposed for Hat stringer stiffened SSCS panel, which make the parametric model computationally efficient. For the parametric design of the panels like the de-bond size/location, mesh sensitivity study, stiffener size etc. the simplified model increases the computational efficiency with satisfactory accuracy.
In this paper, a low profile and simple circular polarized (CP), terahertz (THz) planer mirostrip antenna is designed. Due to the high directivity properties this antenna will be used in short range ...and for fast communication applications. It will be also used for imaging and sensing applications. S-shaped slot are used in microstrip line, for circular polarization purpose. The range of frequency is from 0.258 THz to 0.355 THz with bandwidth of 97 GHz, at -10 dB return loss. The bandwidth of the axial ratio at 3 dB is 11% (0.277-0.309 THz). It has left hand circular polarized (LHCP) antenna at these frequencies. A simple 50 Ω feeding line is used with waveguide port.