Your Selections

Technical review
Show Only

Collections

File Formats

Content Types

Dates

Sectors

Topics

Authors

Publishers

Affiliations

Committees

Events

Magazine

 

An Interactive Vehicle Recommender System Based on Decision Trees and Multi-Armed Bandits

Carnegie Mellon University-Tong Yu, Ole Mengshoel
Ford Motor Co., Ltd.-Dominique Meroux, Zhen Jiang
  • Technical Paper
  • 2019-01-1079
To be published on 2019-04-02 by SAE International in United States
Recommender systems guide a user to useful objects in a large space of possible options in a personalized way. In this paper, we study how to make recommendations for vehicle purchases. This can effectively reduce the human labor in the traditional setting, where customers get recommended vehicles through conversations with the salesmen in dealerships. Comparing a vehicle recommender system to one in other application domains (movies, music, etc.), we identify two major challenges. First, customers usually only purchase a limited number of vehicles, compared to the number of movies or songs. Thus, it is difficult to obtain rich information about a user's purchase history. Second, the content information obtained about the users (demographic, vehicle preference, etc.) is also very limited during their short stay in the dealership. To address these two challenges, we propose an interactive vehicle recommender system based on the methods of decision tree classification and multi-armed bandit. The decision tree effectively selects important questions for the user and understands the user's preference. With these preference as prior information, the multi-armed bandit algorithm…
 

Prediction of Human Actions in Assembly Process by a Spatial-Temporal End-to-End Learning Model

Clemson University-Zhujun Zhang, Weitian Wang, Yi Chen, Yunyi Jia
Harbin Institute of Technology-Zhujun Zhang, Gaoliang Peng
  • Technical Paper
  • 2019-01-0509
To be published on 2019-04-02 by SAE International in United States
It’s important to predict the future actions of human in the industry assembly process. Foreseeing future actions before they have happened is an essential part for flexible human-robot collaboration and crucial safety issues. Vision-based human actions prediction from videos provides intuitive and adequate knowledge for many complex applications. This problem can be interpreted as deducing the next action of people from a short video clip. The history information needs to be considered to learn these relations between each time step for predicting the future steps. However, it is difficult to extract the history information and use it to infer the future situation with the traditional methods. In this scenario, a model is needed to handle the spatial and temporal details stored in past human motions and construct the future action based on limited accessible human demonstrations. In this paper, we apply an autoencoder based deep learning framework for human actions construction, merging into the RNN pipeline for human future actions prediction. This contrasts with traditional approaches which use hand-crafted features and different domain output. The…
 

Accuracy and sensitivity of yaw speed analysis to available data

MEA Forensic Engineers & Scientists-Bradley Heinrichs, Janice Lee, Cole Young
  • Technical Paper
  • 2019-01-0417
To be published on 2019-04-02 by SAE International in United States
Accident reconstructions rarely have complete data with which to determine vehicle speed, and so must bracket the true value with a range. Previous work has shown the effect of friction uncertainty in determining speed from tire marks left by a vehicle in yaw. The goal of the current study was to assess improvements in the accuracy of vehicle speed estimated from yaw marks using progressively more scene and vehicle information. Data for this analysis came from staged S-turn maneuvers that in some cases led to rollover of the SUV test vehicles. Initial speeds were first calculated using the critical curve speed (CCS) formula on the yaw marks from the first portion of the S-maneuver. Then computer simulations were performed with progressively more input data: i) the complete tire marks from the whole S-maneuver, ii) measured vehicle mass, iii) measured suspension stiffness and damping, and iv) measured steering history. Simulations based on the complete tire marks improved the average error compared with the CCS equation. Adding the remaining input data to the simulations did not further…
 

On the Mechanical Design of the Rolling Road Addition to the CSUF Wind Tunnel

California State University-Ramitha Edirisinghe, Salvador Mayoral
  • Technical Paper
  • 2019-01-0651
To be published on 2019-04-02 by SAE International in United States
For ground effect aerodynamic studies in wind tunnels, boundary layer growth over the floor of the test section is a large concern. Many different techniques have been used to remove this boundary layer, of which creating a moving ground plane is considered the best option. By creating a moving floor in the test section, the ground plane speed can be set to that of the free stream air speed, thus removing the boundary layer or at the very least mitigating the boundary layer growth. Several studies detail the aerodynamic design concerns for such devices, but little information exists on the mechanical design concerns. This work reviews lessons learned during a design study of a rolling road for the wind tunnel at California State University, Fullerton (CSUF) mainly focusing on the motor specification and cooling system design. First, an overview of the major systems that comprise a rolling road apparatus are given. Then, initial constraints and free-body diagrams are presented detailing the forces experienced by the rolling road. From these free-body diagrams, analytical equations are presented…
 

A Preliminary Study of the Discharge Current and Spark Energy for the Multi-coil Offset Strategy

University of Windsor-Hua Zhu, Xiao Yu, Qingyuan Tan, Ming Zheng, Graham Reader, Li Liang
Zhuzhou Torch Spark Plug Co., LTD-Huimu Yang, Jin Qian, Tangliang Zhang
  • Technical Paper
  • 2019-01-0725
To be published on 2019-04-02 by SAE International in United States
Future clean Spark Ignited (SI) engines are on the trend of lean/diluted combustion strategies and engine downsizing to achieve the reduction of exhaust emission and fuel consumption. However, the new development in SI engines demands higher spark energy in terms of higher discharge current level and longer discharge duration to overcome the unfavorable ignition conditions caused by the diluted in-cylinder charge and the increased back pressure. The spark energy delivered from the single-coil single discharge of the conventional Transistor Coil Ignition (TCI) system is limited because of its energy storage capability. In order to realize robust ignition control, advanced discharge strategies based on the TCI system, including dual-coil or multiple coil discharge, were proposed over the past. With the use of additional coils, these ignition systems are capable to deliver more energy to the spark gap and generate controllable discharge current profiles. In this work, a three-coil ignition system with a compact package size was developed. Based on this system, the spark discharge characteristic of dual-coil and three-coil discharge strategies were studied and compared with…
 

Application of Dynamic Mode Decomposition to Influence the Driving Stability of Road Vehicles

Altair Engineering-Daiki Matsumoto, Christoph Niedermeier
Technical University of Munich-Thomas Indinger
  • Technical Paper
  • 2019-01-0653
To be published on 2019-04-02 by SAE International in United States
The recent growth of available computational resources has enabled the automotive industry to utilize unsteady Computational Fluid Dynamics (CFD) for their product development on a regular basis. Over the past years, it has been confirmed that unsteady CFD can accurately simulate the transient flow field around complex geometries. Concerning the aerodynamic properties of road vehicles, the detailed analysis of the transient flow field can help to improve the driving stability. Until now, however, there haven’t been many investigations that successfully identified a specific transient phenomenon from a simulated flow field corresponding to driving stability. This is because the unsteady flow field around a vehicle consists of various time and length scales and is therefore too complex to be analyzed with the same strategies as for steady state results. Dynamic Mode Decomposition (DMD) extracts the coherent structures from complex, transient flow fields, which can help to identify certain target phenomena. However, one issue in the practical application of DMD is the difficulty to find a connection between a computed mode and an actual aerodynamic effect on…
 

Numerical modelling of metal forming by SPH with multi-GPU acceleration

ESS Engineering Software Steyr-Md Rushdie Ibne Islam, Chong Peng, Alireza Eslamian
  • Technical Paper
  • 2019-01-1085
To be published on 2019-04-02 by SAE International in United States
Large material distortion, plastic deformation and forging make the numerical modelling of metal forming a difficult task. Grid-based methods such as the Finite Element Method (FEM) are incapable of simulating this process as these schemes suffer from mesh distortion, mesh entanglement and require mesh refinement. The mesh-based numerical frameworks with discontinuous enrichment can model finite deformation problems with limited success. Moreover, the presence of flaws, multiple crack surfaces and their interaction make the simulation even more numerically and computationally intensive. In this regard, the particle-based or meshfree methods are more relevant. There exist several mesh-free methods and among these Smoothed Particle Hydrodynamics (SPH) is a truly meshfree. In SPH the computational domain is discretised by a set of particles. A given particle interacts only with its neighbouring particles through a kernel function with a constant radius. The interaction between particles stops when these move out of each other influence domain. Due to the absence of mesh/grids, SPH is naturally equipped to handle large material deformation. Based on the SPH, a solver with multi GPU acceleration…
 

Numerical Analysis using Fast RANS simulations and Comparison with Experimental Measurements for Closed and Open Grille Realistic Car Models

Simerics Inc.-Sujan Dhar, Yu Jiang
  • Technical Paper
  • 2019-01-0655
To be published on 2019-04-02 by SAE International in United States
Over the past few years, the DrivAer car geometry has driven development of simulation methodology for vehicle aerodynamics. Due to the realistic geometry of the DrivAer car model the results obtained from simulations can be easily extended to the aerodynamic design process of production vehicles. This paper details a comprehensive CFD study of all three variants of the DrivAer car geometries: Fastback, Notchback and Estate configurations. The most realistic geometry was chosen for each of the variants; with detailed underbody, wheels and mirrors. In addition to the closed-grille standard DrivAer models, the open–grille variant has also been simulated. Simulations are performed both with and without ground simulation. Mesh generation was performed without surface alterations (e.g. wrapping) using a novel Binary-Tree automatic unstructured mesher. All simulations were performed using RANS turbulence model. A consistent modeling methodology was developed that was rigorously applied to all variants of the DrivAer model and the simulations are shown to have consistently good agreement with experimental measurements. This demonstrates the potential of using such a methodology for several different types of…
 

Analyzing and Preventing Data Privacy Leakage in Connected Vehicle Services

Ford Motor Co., Ltd.-Yu Seung Kim, Pramita Mitra
University of Michigan-Huaxin Li, Di Ma, Brahim Medjahed
  • Technical Paper
  • 2019-01-0478
To be published on 2019-04-02 by SAE International in United States
The rapid development of connected and automated vehicle technologies together with cloud-based mobility services are revolutionizing the transportation industry. As a result, huge amounts of data are being generated, collected, and utilized, hence providing tremendous business opportunities. However, this big data poses serious challenges mainly in terms of data privacy. The risks of privacy leakage are amplified by the information sharing nature of emerging mobility services and the recent advances in data analytics. In this paper, we first provide an overview of the connected vehicle landscape and point out potential privacy threats. We demonstrate two of the risks, namely additional individual information inference and user de-anonymization, through concrete attack designs. Our experiments on real-world datasets show that the individual information inference and user de-anonymization attacks are feasible in corresponding scenarios. We also propose corresponding countermeasures to defend against such privacy attacks and consider maintaining data usability at the same time. We evaluate the feasibility of our defense strategies using real-world vehicular data.
 

Surface Mapping in 3D Coordinate Metrology for Automotive Industry

Ain Shams University-Monir Koura, Ahmed EL-Akkad
National Institute for Standards-Salah H. R. Ali, Ahmed Sobhy
  • Technical Paper
  • 2019-01-1270
To be published on 2019-04-02 by SAE International in United States
Surface mapping for freeform using CNC-CMM plays an important role in advanced coordinate industries. CNC-CMM machine can assess the parts specification, which has a significant impact on design, manufacturing, inspection and quality control processes. This article, presented four important issues to give a comprehensive view of the progress in this area. Eventually, the aim of this work is to give an overview of 3D plot mapping using freeform artifact utilizing computer aided inspection technique in the area of industrial dimensional metrology. This work happens as a new guide to cover and develop the lack of standard artifact for evaluation geometrical surface features in 3D coordinate metrology to optimize the precision, accuracy and verification time especially for automotive industry.