Smart Fleet Management System Using IoT, Computer Vision, Cloud Computing and Machine Learning Technologies

EasyChair Preprint 2208

Abstract

The objective of this paper was to present the effective use of niche technologies in solving most critical problems of Fleet Industry. The proposed fleet management system described in the paper used a rich set of technologies like Internet of Things(IoT), Cloud Computing, Computer Vision, Machine Learning, Deep Learning and Embedded. IBM Watson IoT and Heroku platforms hosted the application to receive data from vehicle dashboard device. Driver’s face authentication and driving pattern monitoring, Fuel Consumption prediction modules used OpenCV (Computer Vision), SVN (Machine Learning) and CNN (Deep Learning) techniques. Vehicle’s Telematics, deviation from route, unauthorized entry in the container, continuous monitoring of trucks internal environment was handled by a high end embedded device with a set of sensor box, cameras, OBD-II device and a gateway. Real data was collected to train and test the face detection and authentication models used in the system. Simulation results demonstrated that the proposed approach can achieve realistic demand of handing and manipulating humongous data coming every few seconds from several vehicles through IoT, NoSQL CloudantDB database and Cloud Computing. The paper also presented the architecture of the system and experimentation results done for various modules of the system.

Keyphrases: Bluemix, CNN, Eye Aspect Ratio, Fleet, IBM Watson, IoT, ORB-II, Open Face framework, OpenCV, computer vision, eye-blinks detection, linear SVM classifier, neural networks, traffic sign identification, traffic signal identification

@booklet{EasyChair:2208,
  author    = {Priya Singh and Milind Suryawanshi and Darshana Tak and Jayant Sant},
  title     = {Smart Fleet Management System Using IoT, Computer Vision, Cloud Computing and Machine Learning Technologies},
  howpublished = {EasyChair Preprint 2208},
  year      = {EasyChair, 2019}}