Search Results

Abstract Lane changes are stressful maneuvers for drivers, particularly during high-speed traffic flows. However, modeling driver’s lane-changing decision and implementation process is challenging due to the complexity and uncertainty of driving behaviors. To address this issue, this article presents a personalized Lane-Changing Model (LCM) for Advanced Driver Assistance System (ADAS) based on deep learning method. The LCM contains three major computational components. Firstly, with abundant inputs of Root Residual Network (Root-ResNet), LCM is able to exploit more local information from the front view video data. Secondly, the LCM has an ability of learning the global spatial-temporal information via Temporal Modeling Blocks (TMBs). Finally, a two-layer Long Short-Term Memory (LSTM) network is used to learn video contextual features combined with lane boundary based distance features in lane change events.

Abstract Real-time reconstruction of 3D environment attributed with semantic information is significant for a variety of applications, such as obstacle detection, traffic scene comprehension and autonomous navigation. The current approaches to achieve it are mainly using stereo vision, Structure from Motion (SfM) or mobile LiDAR sensors. Each of these approaches has its own limitation, stereo vision has high computational cost, SfM needs accurate calibration between a sequences of images, and the onboard LiDAR sensor can only provide sparse points without color information. This paper describes a novel method for traffic scene semantic segmentation by combining sparse LiDAR point cloud (e.g. from Velodyne scans), with monocular color image. The key novelty of the method is the semantic coupling of stereoscopic point cloud with color lattice from camera image labelled through a Convolutional Neural Network (CNN).

Abstract A kinematic path-following model is developed based on an existing modeling framework established by the authors [1, 2] for prediction of the paths of towing vehicle systems. The presented path-following model determines the path of the towing vehicle using the vehicle’s speed and acceleration data collected by an inertial measurement unit (IMU). An Ackerman steering model was presented to calculate instantaneous directional angles and radii for each towed vehicle based on its geometric data and steering angle. In that model the off-tracking effect is properly captured. A 1:4 scale model for a towing vehicle system was built to test the developed steering model, and it was found that the angles and radii of the towing vehicle and each towed unit calculated using the Ackerman steering model agreed very well with those measured from the scale model.

Abstract The main focus of this article is the online estimation of the terminal airspace sector capacity from the Air Traffic Controller 0ATC) dynamical neural model using Neural Partial Differentiation (NPD) with permissible safe separation and affordable workload. For this purpose, a primarily neural model of a multi-input-single-output (MISO) ATC dynamical system is established, and the NPD method is used to estimate the model parameters from the experimental data. These estimated parameters have a less relative standard deviation, and hence the model validation results show that the predicted neural model response is well matched with the intervention of the ATC workload. Moreover, the proposed neural network-based approach works well with the experimental data online as it does not require the initial values of model parameters, which are unknown in practice.

Abstract An investigation into the capability of passive porosity to reduce the drag of a bluff-body is presented. This initial work involves integrating varying degrees of porosity into the side and back faces of a small-scale model to determine optimum conditions for maximum drag reduction. Both force and pressure measurements at differing degrees of model yaw are presented, with the conditions for optimum performance, identified. At a length-based Reynolds number of 2.3 × 106, results showed a maximum drag reduction of 12% at zero yaw when the ratio of the open area on the back face relative to the side faces was between two and four. For all non-zero yaw angles tested, this ratio reduced to approximately two, with the drag benefit reducing to 6% at 10.5 degrees. From a supplementary theoretical analysis, calculated optimum bleed rate into the base for maximum drag reduction, also showed reasonable agreement to other results reported previously.

Abstract Finite Element Analysis (FEA) of metal cutting is largely the domain of research organizations. Despite significant advances towards accurately modelling metal machining processes, industrial adoption of these advances has been limited. Academic studies, which mainly focused on orthogonal cutting, fail to address this discrepancy. This paper bridges the gap between simplistic orthogonal cutting models and the complex components typical in the manufacturing sector. This paper outlines how to utilize results from orthogonal cutting simulations to predict industrially relevant performance measures efficiently. In this approach, using 2D FEA cutting models a range of feed, speed and rake angles are simulated. Cutting force coefficients are then fit to the predicted cutting forces. Using these coefficients, forces for 3D cutting geometries are calculated.

Abstract This paper presents a machine learning application of the force/torque sensor in a human-robot collaborative manufacturing scenario. The purpose is to simplify the programming for physical interactions between the human operators and industrial robots in a hybrid manufacturing cell which combines several robotic applications, such as parts manipulation, assembly, sealing and painting, etc. A multiclass classifier using Light Gradient Boosting Machine (LightGBM) is first introduced in a robotic application for discriminating five different contact states w.r.t. the force/torque data. A systematic approach to train machine-learning based classifiers is presented, thus opens a door for enabling LightGBM with robotic data process. The total task time is reduced largely because force transitions can be detected on-the-fly. Experiments on an ABB force sensor and an industrial robot demonstrate the feasibility of the proposed method.

Abstract In an automated machining process, monitoring the conditions of the tool is essential for deciding to replace or repair the tool without any manual intervention. Intelligent models built with sensor information and machine learning techniques are predicting the condition of the tool with good accuracy. In this study, statistical models are developed to identify the conditions of the abrasive grinding wheel using the Acoustic Emission (AE) signature acquired during the surface grinding operation. Abrasive grinding wheel conditions are identified using the abrasive wheel wear plot established by conducting experiments. The piezoelectric sensor is used to capture the AE from the grinding process, and statistical features of the abrasive wheel conditions are extracted in time and wavelet domains of the signature. Machine learning algorithms, namely, Classification and Regression Trees (CART) and Support Vector Classifiers (SVC), are used to build statistical models.

The CDIF Family of Standards is primarily designed to be used as a description of a mechanism for transferring information between CASE tools. It facilitates a successful transfer when the authors of the importing and exporting tools have nothing in common except an agreement to conform to CDIF. The language that is defined for the Transfer Format also has applicability as a general language for Import/Export from repositories. The CDIF Integrated Meta-model defined for CASE also has applicability as the basis of standard definitions for use in repositories. The standards that form the complete family of CDIF Standards are documented in EIA/IS-106 CDIF - CASE Data Interchange Format - Overview. These standards cover the overall framework, the transfer format and the CDIF Integrated Meta-model. The diagram in Figure 1 depicts the various standards that comprise the CDIF Family of Standards. The shaded box depicts this Standard and its position in the CDIF Family of Standards.

Abstract Management of expressway networks has been mainly focused on defect management without looking at the correlations with accidental risks. This causes unsustainability in expressway infrastructure maintenance since such defects may not be a contributing factor toward public safety. Thus it is necessary to incorporate accidental events for decision-making in infrastructure management. This study has developed a novel approach to machine learning (ML) that incorporates actual primary data from the last 10 years of single-vehicle accidents (SVA) by collisions with motorway facilities, or so-called single-vehicle collisions with fixed objects. The ML is firstly aimed at identifying the influential factors of SVA in relation to finding effective countermeasures for accidents by integrating the correlation analysis, multiple regression analysis, and ML techniques. The study reveals that wet pavement conditions have a significant effect on SVA.

Engine cooling module air flows depend on package components and vehicle front end geometry. For years, in the early stages of vehicle development, front end geometry air flows were determined from 3/8 scale models or retrofit of similar existing vehicles. As time to market has become much shorter, finite element modeling of air flows is the only tool available. This paper describes how finite element simulations of front end air flows can be run early in the development program independent of any specific engine cooling module configuration and then coupled with traditional one-dimensional component performance models to predict cooling module air flows. The CFD simulation thus replaces the previous scale model testing process. The CFD simulations are used to determine the two parameters that characterize the front end geometry flow resistance (recovery coefficient and internal loss coefficient).

With recent increases in global fuel prices there has become a growing interest in expanding the use of diesel engines in the transportation industry. However, new engine development is costly and time intensive, requiring many hours of expensive engine tests. The ability to accurately predict an engine's performance based on existing models would reduce the expense involved in creating a new engine of different size. In the present study experimental results from two single-cylinder direct injection diesel engines were used to examine previously developed engine scaling models. The first scaling model was based on an equal spray penetration correlation. The second model considered both equal spray penetration and flame lift-off length. The engines used were a heavy-duty Caterpillar engine with a 2.44L displacement and a light-duty GM engine with a 0.48L displacement.

Empirical models for engine-out oxides of Nitrogen (NOx) and smoke emissions have been developed for the purpose of minimizing transient emissions while maintaining transient response. Three major issues have been addressed: data acquisition, data processing and modeling method. Real and virtual transient parameters have been identified for acquisition. Accounting for the phase shift between transient engine events and transient emission measurements has been shown to be very important to the quality of model predictions. Several methods have been employed to account for the transient transport delays and sensor lags which constitute the phase shift. Finally several different empirical modeling methods have been used to determine the most suitable modeling method for transient emissions. These modeling methods include several kinds of neural networks, global regression and localized regression.

This recommended practice applies to vibration monitoring systems for rotorcraft and fixed-wing drive trains, airframes, propulsion systems, electric power generators, and flight control systems. It addresses all aspects of metrics, including what to measure, how to measure, and how to evaluate the results.

With more stringent emissions regulations, effective emission modeling on NOx and soot for both on-road and off-road diesel engines becomes increasingly important for diesel engine system design and real-time engine controls. In this paper, a heuristic macro-parameter-dependent approach is proposed by combining theoretical analysis with semi-empirical method. The proposed modeling approach is different from the existing methods, such as empirical modeling, phenomenological modeling, and three-dimensional KIVA modeling. The proposed model uses the macro parameters of engine performance, both cycle-average (e.g., air-to-fuel ratio, EGR rate) and in-cylinder instantaneous data (e.g., cylinder pressure trace) as input. The model computes NOx and soot as a function of crank angle. A concept of “time-variant virtual space zones (burning, burned, and unburned)” is proposed based on the fraction of fuel burnt.

For many suppliers in the aerospace value chain, business commences when the customer shares the Technical Data Package (TDP) that defines the detailed requirements for a specific part. To convert the customer TDP into the necessary internal documentation, suppliers must expend large amounts of effort. This generally involves passing along copies of the TDP to each functional discipline, which not only results in redundant and laborious work, but it introduces technical risk. There are now software tools available that enable an intelligent TDP that provides more value than just sharing a 3D CAD model. These tools electronically organize and integrate all elements of the TDP independent of the PLM software in use. The application of the intelligent TDP has enabled a 30% reduction in supply chain inefficiencies.

The commercial turbofan trend of increasing bypass ratio and decreasing fan pressure ratio has seen its latest market entry in Pratt & Whitney's PurePower™ product line, which will power regional aircraft for the Bombardier and Mitsubishi corporations, starting in 2013. The high-bypass-ratio, low-fan-pressure-ratio trend, which is aimed at diminishing noise while increasing propulsive efficiency, combines with contemporary business factors including the escalating cost of testing and limited availability of simulated altitude test sites to pose formidable challenges for engine certification and performance validation. Most fundamentally, high bypass ratio and low fan pressure ratio drive increased gross-to-net thrust ratio and decreased fan temperature rise, magnifying by a factor of two or more the sensitivity of in-flight thrust and low spool efficiency to errors of measurement and assumption, i.e., physical modeling.

The authors measured the vibratory forces acting on an airfoil model by performing a ground vibration test (GVT). The airfoil model was manufactured using rapid prototyping. In the experiments, the airfoil model's structural response was also recorded and described. This paper detailedly introduces the entire experiment process and the obtained experimental data agreed well to the actual values.

The objective of this study is to evaluate ventilation efficiency regarding to the International Space Station (ISS) cabin ventilation during the ISS assembly mission 1J. The focus is on carbon dioxide spatial/temporal variations within the Node 2 and attached modules. An integrated model for CO2 transport analysis that combines 3D CFD modeling with the lumped parameter approach has been implemented. CO2 scrubbing from the air by means of two ISS removal systems is taken into account. It has been established that the ventilation scheme with an ISS Node 2 bypass duct reduces short-circuiting effects and provides less CO2 gradients when the Space Shuttle Orbiter is docked to the ISS. This configuration results in reduced CO2 level within the ISS cabin.

Diesel spray experimentation at controlled high-temperature and high-pressure conditions is intended to provide a more fundamental understanding of diesel combustion than can be achieved in engine experiments. This level of understanding is needed to develop the high-fidelity multi-scale CFD models that will be used to optimize future engine designs. Several spray chamber facilities capable of high-temperature, high-pressure conditions typical of engine combustion have been developed, but because of the uniqueness of each facility, there are uncertainties about their operation. For this paper, we describe results from comparative studies using constant-volume vessels at Sandia National Laboratories and IFP.