Your Selections

Northeastern University
Show Only

Collections

File Formats

Content Types

Dates

Sectors

Topics

Authors

Publishers

Affiliations

Events

   This content is not included in your SAE MOBILUS subscription, or you are not logged in.

A Multi-Objective Recognition Algorithm with Time and Space Fusion

Northeastern University-Huan Wang, Jianning Chi, Chengdong Wu, Xiaosheng Yu, Qian Hu
Published 2019-04-02 by SAE International in United States
Multi-target recognition technology plays an important role in the field of intelligent driving. In this paper, we propose a novel multi-target recognition algorithm with high accuracy and efficiency. We design a time series based recurrent neural network that integrates historical appearance information on the timeline, which can effectively improve the recognition accuracy. The target appearance characteristics extracted from the feature fusion network are then sent to the recursive neural network with the function of long-term and short-term memory for prediction, extending the learning and analysis of the neural network to the space-time domain. After the LSTM interprets advanced visual features, time series based regression is used as an appearance model to regress features to a particular visual element position through preliminary position inference. We evaluate our proposed algorithm on the KITTI data set and a large number of real scene experiments. Compared to other multi-target recognition methods, our algorithm provides much better accuracy. The experimental results show that the accuracy of road target algorithm can be effectively improved by learning historical visual semantics and target…
This content contains downloadable datasets
Annotation ability available
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.

An Improved Multi-Pedestrian Tracking System Based on Deep Neural Network

Northeastern University-Yuan Gong, Jianning Chi, Xiaosheng Yu, Chengdong Wu, Yifei Zhang, Na Gao
Published 2019-04-02 by SAE International in United States
The intelligent vehicle driverless technology has become a very hot topic in the past two decades. To solve the road safety problem, which is one of the most important factors inhibiting the development of intelligent vehicle technology, the multi-target tracking system has attracted more and more attention in recent study since it can detect and track multiple objects in traffic scene so as to help the whole driving system plan the safe route. In this work, a novel multi-pedestrian tracking system based on deep neural network is proposed to improve the tracking efficiency while providing high recognition accuracy. The proposed tracking system consists of two parts: 1) pedestrian detector, and 2) pedestrian tracker. For the detector part, we first transform the image convolution operation in spatial-temporal domain to the coefficient product operation in complex frequency domain, and then replace the maximum pooling and mean pooling operation in the traditional SSD detection network by the proposed product operation, and the modified SSD model is used as the pedestrian detector. For the tracker part, we train a…
This content contains downloadable datasets
Annotation ability available
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.

Increasing Development Assurance for System and Software Development with Validation and Verification Using ASSERT™

Northeastern University-Panagiotis Manolios
GE Aviation Systems LLC-Craig McMillan, Mark Stephens, Daniel Russell
Published 2019-03-19 by SAE International in United States
System design continues to trend toward increasing complexity as more functionality is added to aviation systems and the level of automation is increased. Since exhaustive validation and verification of this functionality becomes increasingly difficult, reliance on development assurance is needed to provide confidence that errors in requirements, design and implementation have been identified and corrected. To address this need for increased development assurance, GE is introducing a tool called ASSERT™ (Analysis of Semantic Specifications and Efficient generation of Requirements-based Tests). The system developer uses this tool to capture requirements in an unambiguous way with built-in semantic error checking. The requirements analysis engine is then used to assist in requirements validation to identify common problems which may include requirements that conflict with one another, requirements that do not fully specify the behavior of a function, requirements that are not independent of one another, and requirements that are either always true or false. Having unambiguous and complete requirements also enables the tool to consistently generate a complete set of requirements-based test cases and procedures to ensure the…
Annotation ability available
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.

Quantifying Uncertainty in Predictions of Kinetically Modulated Combustion: Application to HCCI Using a Detailed Transportation Fuel Model

Northeastern University-Richard West
Argonne National Laboratory-S. Scott Goldsborough, Aleksandr fridlyand
Published 2018-04-03 by SAE International in United States
Simulation of chemical kinetic processes in combustion engine environments has become ubiquitous towards the understanding of combustion phenomenology, the evaluation of controlling parameters, and the design of configurations and/or control strategies. Such calculations are not free from error however, and the interpretation of simulation results must be considered within the context of uncertainties in the chemical kinetic model. Uncertainties arise due to structural issues (e.g., included/missing reaction pathways), as well as inaccurate descriptions of kinetic rate parameters and thermochemistry. In fundamental apparatuses like rapid compression machines and shock tubes, computed constant-volume ignition delay times for simple, single-component fuels can have variations on the order of factors of 2-4. This work investigates how kinetic rate parameter uncertainties manifest themselves in terms of combustion phasing under variable-volume, homogeneous charge compression ignition combustion. Iso-octane is used under lean fuel loadings (equivalence ratio = 0.35) at naturally aspirated conditions, where a range of combustion phasings is achieved via changes to the initial, bottom dead center (BDC) temperature, covering phasings from near top dead center (TDC) to 18 crank angle…
This content contains downloadable datasets
Annotation ability available
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.

Edge Enhanced Traffic Scene Segmentation Algorithm with Deep Neural Network

Northeastern University-Wei Liu, Jun Hu, Huai Yuan
Neusoft-Huan Tian, Shuai Cheng
Published 2017-09-23 by SAE International in United States
Image segmentation is critical in autonomous driving field. It can reveal essential clues such as objects’ shape or boundary information. The information, moreover, can be leveraged as input information of other tasks: vehicle detection, for example, or vehicle trajectory prediction. SegNet, one deep learning based segmentation model proposed by Cambridge, has been a public baseline for scene perception tasks. It, however, suffers an accuracy deficiency in objects marginal area. Segmentation of this area is very challenging with current models. To alleviate the problem, in this paper, we propose one edge enhanced deep learning based model. Specifically, we first introduced one simple, yet effective Artificial Interfering Mechanism (AIM) which feeds segmentation model manual extracted key features. We argue this mechanism possesses the ability to enhance essential features extraction and hence, ameliorate the model performance. Other modifications of model structures were also designed for further improving model’s feature extraction ability. Besides, one Pixel Alignment Unit (PAU) was presented for pixel level alignment. The unit is designed based on Bidirectional Long Short Term Memory (Bi-LSTM) unit and, according…
This content contains downloadable datasets
Annotation ability available
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.

Real-time Pedestrian Detection using Convolutional Neural Network on Embedded Platform

SAE International Journal of Passenger Cars - Electronic and Electrical Systems

Northeastern University-Jun Hu, Wei Liu, Shuai Cheng, Huan Tian, Huai Yuan, Hong Zhao
  • Journal Article
  • 2016-01-1877
Published 2016-09-14 by SAE International in United States
The convolutional neural network (CNN) has achieved extraordinary performance in image classification. However, the implementation of such architecture on embedded platforms is a big challenge task due to the computing resource constraint issue. This paper concentrates on optimization of CNN on embedded platforms with a case study of pedestrian detection in ADAS. The main contribution of this proposed CNN is its ability to run pedestrian classification task in real time with high accuracy based on a platform with ARM embedded. The CNN model has been trained with GPU locally and then transformed into an efficient implementation on embedded platforms. The efficient implementation uses dramatically small network scale and a lightweight CNN is obtained. Specifically, parameters of the network are compressed by adopting integer weights to reduce computational complexity. Meanwhile, other optimizations have also been proposed to adapt the general ARM processor architecture. Finally, a robust and efficient CNN architecture is executed to run pedestrian detection algorithm in real time. The embedded platform employs a 1GHz Cortex-A53 ARMv8 based CPU. The network performs nearly 200 regions…
Annotation ability available
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.

Recent Developments in the Design of a Ceramic Automotive Oil Filter

Northeastern University-Yiannis Levendis
Notox A/S-Lars T. Johannesen
Published 2011-04-12 by SAE International in United States
This manuscript describes two different design configurations for a novel environmentally-friendly automotive oil filter. In both cases the filter seamlessly retrofits existing engine applications. The filter element is housed in an easy-to-dismantle casing. The filter element may be replaced at every oil change, but not the casing which may last the lifetime of the engine. The filter element is made of ceramic honeycomb material, which typically possesses high-filtration efficiency characteristics and large contaminant accumulation capacity. When a filter element is replaced with a new unit, the old unit is sent out for treatment so that it can be reused. These cartridges are practical, durable, cost effective, user-friendly and environment-friendly. In this paper, emphasis is given to the engineering design process that took place in the development of this product, i.e., formulation of design problem statements and specifications, comparison of the two alternative solutions, feasibility considerations and detailed system descriptions.
Annotation ability available
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.

An Experimental Study Using EEG to Detect Driver Drowsiness

Northeastern University-Hongjie Leng, Yingzi Lin, Ronald R. Mourant
Published 2008-10-07 by SAE International in United States
Drowsy driving is a serious issue in modern transportation. Our review of current ways to detect driver drowsiness revealed that electroencephalography (EEG) has the potential to achieve excellent performance in driver drowsiness recognition. This paper presents an experimental study for finding EEG features that are sensitive to a driver’s drowsiness, yet robust to noise signals. We recorded participants’ EEGs in an alert state and a drowsy state for a driving task performed in a driving simulator. Twelve EEG features were analyzed for their power of distinguishing the alert state from the drowsy state. As a conclusion, specific EEG features can be used to detect drivers’ drowsiness. The classification power of the EEG features depends on the scalp locations where the EEG is measured.
Annotation ability available
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.

Effect of Fuel Chemical Structure and Properties on Diesel Engine Performance and Pollutant Emissions: Review of the Results of Four European Research Programs

SAE International Journal of Fuels and Lubricants

Northeastern University-Y. A. Levendis
Hellenic Air Force Academy-R. G. Papagiannakis
  • Journal Article
  • 2008-01-0838
Published 2008-04-14 by SAE International in United States
During recent years, the deterioration of greenhouse phenomenon, in conjunction with the continuous increase of worldwide fleet of vehicles and crude oil prices, raised heightened concerns over both the improvement of vehicle mileage and the reduction of pollutant emissions. Diesel engines have the highest fuel economy and thus, highest CO2 reduction potential among all other thermal propulsion engines due to their superior thermal efficiency. However, particulate matter (PM) and nitrogen oxides (NOx) emissions from diesel engines are comparatively higher than those emitted from modern gasoline engines. Therefore, reduction of diesel emitted pollutants and especially, PM and NOx without increase of specific fuel consumption or let alone improvement of diesel fuel economy is a difficult problem, which requires immediate and drastic actions to be taken. A direct means for reducing diesel engine emitted pollutants, while preserving fuel economy, is the reformulation of conventional diesel fuel. A very promising way to improve diesel fuel behavior is the addition of synthetic or biologically renewable oxygenates. However, as the characteristics of the “parental” conventional diesel oil affect directly the…
Annotation ability available
   This content is not included in your SAE MOBILUS subscription, or you are not logged in.

An Experiment to Non-Intrusively Collect Physiological Parameters towards Driver State Detection

Northeastern University-H. Cai, Y. Lin
Published 2007-04-16 by SAE International in United States
Acquiring drivers’ physiological parameters is helpful for driver assistance systems to reliably estimate drivers’ comprehensive mental states, but traditional methods of measurements have to intrusively attach sensors or fasten electrodes to the body. Intrusive methods of measurement are acceptable in experimental settings but unacceptable as routine operation in daily driving. In this paper, the Smart Wheel System is introduced. The Smart Wheel System is a prototype system that collects the following cognition/emotion physiological signals: blood volume pulse, skin conductance, skin temperature, and respiration wave. It uses embedded sensors and electrodes in steering wheel and seat belt with near-zero interference. Due to the fact that hand and body movements cause temporary gap in the real-time data stream, gripping force and gripping position are collected as indicative data to compensate breaks in the physiological data stream. In the current experiments, FlexComp biofeedback system is used as a standard test instrument to verify the performance of the Smart Wheel System. The correlations between the Smart Wheel System and the FlexComp measurements range from 0.85 to 0.96. The average…
Annotation ability available