Machine Learning for Intelligent Transportation Systems

Similar documents
Deployment and Testing of Optimized Autonomous and Connected Vehicle Trajectories at a Closed- Course Signalized Intersection

Virtual Worlds for the Perception and Control of Self-Driving Vehicles

Domain Adaptation & Transfer: All You Need to Use Simulation for Real

VSI Labs The Build Up of Automated Driving

23270: AUGMENTED REALITY FOR NAVIGATION AND INFORMATIONAL ADAS. Sergii Bykov Technical Lead Machine Learning 12 Oct 2017

Transer Learning : Super Intelligence

BIG DATA EUROPE TRANSPORT PILOT: INTRODUCING THESSALONIKI. Josep Maria Salanova Grau CERTH-HIT

SIS63-Building the Future-Advanced Integrated Safety Applications: interactive Perception platform and fusion modules results

Neural Networks The New Moore s Law

How Innovation & Automation Will Change The Real Estate Industry

Constructing a Traffic Control Process Diagram

DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition. ECE 289G: Paper Presentation #3 Philipp Gysel

interactive IP: Perception platform and modules

The Virtues of Virtual Reality Artur Filipowicz and Nayan Bhat Princeton University May 18th, 2017

arxiv: v1 [cs.lg] 2 Jan 2018

Intelligent Technology for More Advanced Autonomous Driving

A Winning Combination

KÜNSTLICHE INTELLIGENZ JOBKILLER VON MORGEN?

A Spatiotemporal Approach for Social Situation Recognition

Stanford Center for AI Safety

Robotics and Autonomous Systems

Interaction in Urban Traffic Insights into an Observation of Pedestrian-Vehicle Encounters

Choosing the Optimum Mix of Sensors for Driver Assistance and Autonomous Vehicles

Reinforcement Learning for CPS Safety Engineering. Sam Green, Çetin Kaya Koç, Jieliang Luo University of California, Santa Barbara

Autonomous driving made safe

Invited talk IET-Renault Workshop Autonomous Vehicles: From theory to full scale applications Novotel Paris Les Halles, June 18 th 2015

Intelligent Driving Agents

Speed Enforcement Systems Based on Vision and Radar Fusion: An Implementation and Evaluation 1

Evaluation of Connected Vehicle Technology for Concept Proposal Using V2X Testbed

FORESIGHT AUTONOMOUS HOLDINGS NASDAQ/TASE: FRSX. Investor Conference. December 2018

Addressing the Uncertainties in Autonomous Driving

Traffic Management for Smart Cities TNK115 SMART CITIES

Big data in Thessaloniki

P1.4. Light has to go where it is needed: Future Light Based Driver Assistance Systems

MOBILITY RESEARCH NEEDS FROM THE GOVERNMENT PERSPECTIVE

An Introduction to Machine Learning for Social Scientists

What will the robot do during the final demonstration?

MATLAB DIGITAL IMAGE/SIGNAL PROCESSING TITLES

GNSS in Autonomous Vehicles MM Vision

Following Dirt Roads at Night-Time

DEEP LEARNING A NEW COMPUTING MODEL. Sundara R Nagalingam Head Deep Learning Practice

Vision & Industry 4.0: Towards smarter sensors. Dr. Amina Chebira Vision Embedded Systems, CSEM SA October 4 th, 2016

EE631 Cooperating Autonomous Mobile Robots. Lecture 1: Introduction. Prof. Yi Guo ECE Department

Semantic Localization of Indoor Places. Lukas Kuster

BATTELLE AND THE SMART CITY. Turning vision into reality for tomorrow s urban environments.

Perception platform and fusion modules results. Angelos Amditis - ICCS and Lali Ghosh - DEL interactive final event

The GATEway Project London s Autonomous Push

Bayesian Filter to accurately track airport moving objects

Tsuyoshi Sato PIONEER CORPORATION July 6, 2017

The next level of intelligence: Artificial Intelligence. Innovation Day USA 2017 Princeton, March 27, 2017 Michael May, Siemens Corporate Technology

Law, Economics, Political Science, and Public Policy. Associate Professor F. Scott Kieff School of Law

COGNITIVE MODEL OF MOBILE ROBOT WORKSPACE

Autonomous Mobile Robot Design. Dr. Kostas Alexis (CSE)

Computer vision, wearable computing and the future of transportation

List of Figures List of Tables. Chapter 1: Introduction 1

FLASH LiDAR KEY BENEFITS

Image Processing Based Vehicle Detection And Tracking System

CS6700: The Emergence of Intelligent Machines. Prof. Carla Gomes Prof. Bart Selman Cornell University

CS343 Introduction to Artificial Intelligence Spring 2010

Semantic Segmentation on Resource Constrained Devices

Cómo estructurar un buen proyecto de Machine Learning? Anna Bosch Rue VP Data Launchmetrics

CSC384 Intro to Artificial Intelligence* *The following slides are based on Fahiem Bacchus course lecture notes.

ARGUING THE SAFETY OF MACHINE LEARNING FOR HIGHLY AUTOMATED DRIVING USING ASSURANCE CASES LYDIA GAUERHOF BOSCH CORPORATE RESEARCH

Deep Learning for Infrastructure Assessment in Africa using Remote Sensing Data

Scene Perception based on Boosting over Multimodal Channel Features

Learning Pixel-Distribution Prior with Wider Convolution for Image Denoising

PERCEIVING MOVEMENT. Ways to create movement

Autonomous Vehicle Simulation (MDAS.ai)

A NEW NEUROMORPHIC STRATEGY FOR THE FUTURE OF VISION FOR MACHINES June Xavier Lagorce Head of Computer Vision & Systems

Application of AI Technology to Industrial Revolution

MATLAB 및 Simulink 를이용한운전자지원시스템개발

Driving Using End-to-End Deep Learning

Innovative mobility data collection tools for sustainable planning

March 10, Greenbelt Road, Suite 400, Greenbelt, MD Tel: (301) Fax: (301)

CS343 Introduction to Artificial Intelligence Spring 2012

Automatic understanding of the visual world

Data collection and modeling for APTS and ATIS under Indian conditions - Challenges and Solutions

Demystifying Machine Learning

GESTURE RECOGNITION FOR ROBOTIC CONTROL USING DEEP LEARNING

Content-Based Multimedia Analytics: Rethinking the Speed and Accuracy of Information Retrieval for Threat Detection

Current Technologies in Vehicular Communications

Telling What-Is-What in Video. Gerard Medioni

Overview Agents, environments, typical components

Real-time Cooperative Multi-target Tracking by Dense Communication among Active Vision Agents

Powerful But Limited: A DARPA Perspective on AI. Arati Prabhakar Director, DARPA

Israel Railways No Fault Liability Renewal The Implementation of New Technological Safety Devices at Level Crossings. Amos Gellert, Nataly Kats

WHO. 6 staff people. Tel: / Fax: Website: vision.unipv.it

Trust in Automated Vehicles

Presented by: Hesham Rakha, Ph.D., P. Eng.

Video Object Segmentation with Re-identification

Robust Positioning for Urban Traffic

Responsible AI & National AI Strategies

Roadside Range Sensors for Intersection Decision Support

Wi-Fi Fingerprinting through Active Learning using Smartphones

Spring 2018 CS543 / ECE549 Computer Vision. Course webpage URL:

On Emerging Technologies

Advanced Techniques for Mobile Robotics Location-Based Activity Recognition

ADAS Development using Advanced Real-Time All-in-the-Loop Simulators. Roberto De Vecchi VI-grade Enrico Busto - AddFor

Introduction to Vision & Robotics

A.I in Automotive? Why and When.

Transcription:

Machine Learning for Intelligent Transportation Systems Patrick Emami (CISE), Anand Rangarajan (CISE), Sanjay Ranka (CISE), Lily Elefteriadou (CE) MALT Lab, UFTI September 6, 2018

ITS - A Broad Perspective What is ITS? Working definition Utilizing cutting-edge, synergistic technologies to develop and improve transportation systems of all kinds

What is ITS? ITS - A More Narrow Perspective ITS for improved urban mobility Source: https://www.arch2o.com/future-urban-mobility/

What is ITS? ITS for Urban Mobility - Autonomous Vehicles Source: http://www.vtpi.org/avip.pdf

Intelligent Transportation Systems What is ITS? ITS for Urban Mobility - Traffic Surveillance

What is ITS? ITS for Urban Mobility -

Machine Learning

Machine Learning Working definition Extracting patterns and abstractions from datasets to make intelligent decisions on previously unseen data

Other Intelligent Tools Machine learning is rarely used in isolation, and often overlaps with the following fields: 1 Discrete and continuous optimization 2 Signal processing 3 Distributed systems 4 Control theory 5 And more...!

Machine Learning for ITS Deep neural networks trained on massive datasets are at the cutting-edge in terms of performance. The theory is lagging behind! Source: http://yann.lecun.com/exdb/lenet/

Source: Andrew Ng: https://www.slideshare.net/extractconf

ML Computer Vision A primary use of ML in ITS is for intelligent perception Some key tasks 1 Object detection 2 Multi-object tracking 3 Activity recognition

Autonomous Vehicles Source: https://www.wired.com/story/waymo-launches-selfdriving-minivans-fiat-chrysler/, http://sitn.hms.harvard.edu/flash/2017/self-driving-carstechnology-risks-possibilities/

Autonomous Vehicles Source: https://www.wired.com/story/waymo-launches-selfdriving-minivans-fiat-chrysler/, http://sitn.hms.harvard.edu/flash/2017/self-driving-carstechnology-risks-possibilities/

Intelligent Transportation Systems Traffic Surveillance Use Computer Vision to try to answer these questions: Are pedestrians crossing? How many vehicles? Any driving the wrong way?

Intelligent Transportation Systems Object detection It can explicitly/implicitly answer the following questions 1 Where are the interesting objects within my field of view? 2 What are the object classes (pedestrian, bicyclist, sedan,...)? 3 How many objects are there?

Intelligent Transportation Systems Object detection It can explicitly/implicitly answer the following questions 1 Where are the interesting objects within my field of view? 2 What are the object classes (pedestrian, bicyclist, sedan,...)? 3 How many objects are there? For simplicity, we re lumping localization (where in the image are the objects) and classification (what class) into detection.

Object Detection with Real world challenges The current best way to handle variations in lighting, orientation, and scale when deploying is data augmentation. Source: http://cs231n.github.io/convolutional-networks/

Multi-object Tracking Goal is to estimate the trajectories of all objects in a dynamic scene MOT from a stationary traffic cam ource: Luo, et. al. Fast and Furious: Real Time End-to-End D Detection, Tracking and Motion Forecasting With a Single onvolutional Net. CVPR 2018. MOT using LiDAR from an AV

Obstacles to solving MOT 1 Object detectors don t handle partial/full occlusion or drastic variations in lighting, color, orientation very well 2 Stitching detections together over time into tracks is a hard discrete optimization (or inference) problem 3 Sensors are unreliable/noisy 4 MOT systems are typically overly-complex and contain lots of hand-tuned problem-specific parameters ource: Emami, Patrick, et al. Machine Learning Methods for olving Assignment Problems in Multi-Target Tracking. arxiv reprint arxiv:1802.06897 (2018).

Obstacles to solving MOT 1 Object detectors don t handle partial/full occlusion or drastic variations in lighting, color, orientation very well 2 Stitching detections together over time into tracks is a hard discrete optimization (or inference) problem 3 Sensors are unreliable/noisy 4 MOT systems are typically overly-complex and contain lots of hand-tuned problem-specific parameters Interesting research question keeping me up at night Is there a principled way to learn the concept of object permanence within an MOT system? ource: Emami, Patrick, et al. Machine Learning Methods for olving Assignment Problems in Multi-Target Tracking. arxiv reprint arxiv:1802.06897 (2018).

Activity Recognition Using object detections and trajectories, can we then extract patterns at the level of behaviors? 1 Pedestrian safety; ID ing whether a person is walking/about to walk into the street 2 Vehicle collision prediction 3 Multi-agent modeling at traffic intersections and merging zones for AVs

Collision Prediction Source: Xiaohui Huang, Sanjay Ranka and Anand Rangarajan. Real-time Multi-Object Tracking and Road Traffic Safety Measurement. In preparation.

Traffic Flow Prediction Traffic Intersections Guiding question Using sensors and edge computing, can we maximize the efficiency of traffic flow through a road network in real-time?

Traffic Sensors Traffic Flow Prediction Traffic Intersections

Short-term Traffic Flow Prediction Traffic Flow Prediction Traffic Intersections Accurate forecasting of congestion levels enables real-time traffic planning Train a model (e.g., deep network or Random Forest) to predict next 15-30 minutes of traffic flow. Source: Polson, Nicholas G., and Vadim O. Sokolov. Deep learning for short-term traffic flow prediction. Transportation Research Part C: Emerging Technologies 79 (2017): 1-17.

Traffic Intersection Optimization Traffic Flow Prediction Traffic Intersections Source: Pourmehrab, M., Elefteriadou, L., Ranka, S., & Martin-Gasulla, M. Optimizing Signalized Intersections Performance under Conventional and Automated Vehicles Traffic. arxiv:1707.01748 (2017)

Conclusion Intelligent Transportation Systems Plenty of challenges when applying ML to ITS 1 Collecting, cleaning, and labeling large-scale datasets 2 Law-makers and policy has to keep up with the tech 3 Brittle models that break when applied to new domains 4 Security and privacy

Conclusion Intelligent Transportation Systems Plenty of challenges when applying ML to ITS 1 Collecting, cleaning, and labeling large-scale datasets 2 Law-makers and policy has to keep up with the tech 3 Brittle models that break when applied to new domains 4 Security and privacy But we ve made great progress!

Thank you! Intelligent Transportation Systems Questions? Twitter: @patrickomid, email: pemami@ufl.edu Slides available at: https://pemami4911.github.io