Camera + Various CV algorithms … Affordable, easy to use

Slides:



Advertisements
Similar presentations
Clustering Crowdsourced Videos by Line-of-Sight FOCUS: Clustering Crowdsourced Videos by Line-of-Sight Puneet Jain, Justin Manweiler, Arup Acharya, and.
Advertisements

Hybrid Context Inconsistency Resolution for Context-aware Services
Ray tracing. New Concepts The recursive ray tracing algorithm Generating eye rays Non Real-time rendering.
A Novel Approach of Assisting the Visually Impaired to Navigate Path and Avoiding Obstacle-Collisions.
VTrack: Accurate, Energy-Aware Road Traffic Delay Estimation Using Mobile Phones Arvind Thiagarajan, Lenin Ravindranath, Katrina LaCurts, Sivan Toledo,
Detection and Measurement of Pavement Cracking Bagas Prama Ananta.
“Green Light Wave” Traffic Control System Liron Netzer - CITI Yossi Gabay - Marvell Shay Avivi - Motorola Solutions.
G4 Apps The Impact of Connected Vehicles on Traffic Operations ISMA Traffic Expo October 1, 2014.
Perception and Communications for Vulnerable Road Users safety Pierre Merdrignac Supervisors: Fawzi Nashashibi, Evangeline Pollard, Oyunchimeg Shagdar.
Mining Motion Sensor Data from Smartphones for Estimating Vehicle Motion Tamer Nadeem, PhD Department of Computer Science NSF Workshop on Large-Scale Traffic.
IntelliDrive Safety Workshop July 20, 2010 Alrik L. Svenson US Department of Transportation National Highway Traffic Safety Administration IntelliDrive.
Department of Electrical and Computer Engineering Development of a Portable Work Zone Traffic Safety Information System using DSRC Based V2I and V2V Communication.
Visual Odometry for Ground Vehicle Applications David Nister, Oleg Naroditsky, James Bergen Sarnoff Corporation, CN5300 Princeton, NJ CPSC 643, Presentation.
TEAM 8 Barry Li Victor Chen Elyas Sepasi Aron McKinnon ENSC 305/440 April 15 th,
Adaptive Traffic Light Control with Wireless Sensor Networks Presented by Khaled Mohammed Ali Hassan.
Rutgers: Gayathri Chandrasekaran, Tam Vu, Marco Gruteser, Rich Martin,
Occupant Classification System for Automotive Airbag Suppression A.Jaffer Sharief EEL
Automatic Collection of Fuel Prices from a Network of Mobile Cameras
Sensys 2009 Speaker:Lawrence.  Introduction  Overview & Challenges  Algorithm  Travel Time Estimation  Evaluation  Conclusion.
Human tracking and counting using the KINECT range sensor based on Adaboost and Kalman Filter ISVC 2013.
Abstract Design Considerations and Future Plans In this project we focus on integrating sensors into a small electrical vehicle to enable it to navigate.
Computers In Farming  Damien Gallagher  2 nd BSc Information Technology 
Co-operative Systems for Road Safety “Smart Vehicles on Smart Roads”
Concept Design Review THE DUKES OF HAZARD CAMILLE LEGAULT, NEIL KENNEDY, OMAR ROJAS, FERNANDO QUIJANO, AND JIMMY BUFFI April 24, 2008.
Geopositioning and Applications in Transportation GMAT9205 Students: Ellis Leung ( ) Terry Nham ( )
Analysis of the SHRP 2 Naturalistic Driving Study Data [S08(B)] Evaluation of Offset Left-Turn Lanes Jessica M. Hutton Presentation to MCTRS August 15,
University of Maryland Department of Civil & Environmental Engineering By G.L. Chang, M.L. Franz, Y. Liu, Y. Lu & R. Tao BACKGROUND SYSTEM DESIGN DATA.
Sybot: An Adaptive and Mobile Spectrum Survey System for WiFi Networks Kyu-Han Kim Deutsche Telekom R&D Lab USA Alexander W. Min and Kang G. Shin Real-Time.
Navigation System for the Visually Impaired Based on an Information Server Concept Ari Virtanen, Sami Koskinen.
Dynamic 3D Scene Analysis from a Moving Vehicle Young Ki Baik (CV Lab.) (Wed)
What do you do when…. when making a right turn on a red light after a stop? you stop where there is a stop sign or flashing red signal? making a left turn.
STREET VIEW Wenyang Liu GMAT9205: Fundamentals of Geo-Positioning Session 1, 2010.
Timothy E. Barnett, P.E., PTOE State Safety Operations Engineer Alabama Department of Transportation.
Intelligent Transportation System Oum Saokosal Cambodian Graduate Student April 2009.
Applications of Remote Sensing in Transportation.
Evaluating Handheld GPS Receivers for Recording Archaeological Survey Data GEO444 – Practical GPS William Eichmann and Jake Reidel.
Road Inventory Data Collection Re-engineering Collected Data Items (more than 50 items): –Street Names. –Pavement width, number of lanes, etc. –Bike path,
WEST VIRGINIA UNIVERSITY Lane Department of Computer Science and Electrical Engineering CROWDSOURCED TRAFFIC MAP Team Members: Faculty Mentor: David Williams.
CARLOC: Precisely Tracking Automobile Position
The Cyber-Physical Bike A Step Toward Safer Green Transportation.
Turning a Mobile Device into a Mouse in the Air
Visual Odometry for Ground Vehicle Applications David Nistér, Oleg Naroditsky, and James Bergen Sarnoff Corporation CN5300 Princeton, New Jersey
GPS & GSM BASED CAR SECURITY SYSTEM. Student’s Name with USN No.
Auto-Park for Social Robots By Team Daedalus. Requirements for FVE Functional Receive commands from user via smartphone app Share data with other cars.
CS 2310 Final Project - Driving Behavior Monitor Haifeng Xu Dec. 5, 2013.
Mapping of Traffic Conditions at Downtown Thessaloniki with the Help of GPS Technology P. D. Savvaidis and K. Lakakis Aristotle University of Thessaloniki,
NOTE: To change the image on this slide, select the picture and delete it. Then click the Pictures icon in the placeholder to insert your own image. “Smartphones,
Intelligent and Non-Intelligent Transportation Systems 32 Foundations of Technology Standard 18 Students will develop an understanding of and be able to.
Technology Solutions for Tolling and Traffic Management N Video Detection Technology and Marketplace Michael Wieck Business Development Manager, Roadway.
Urban Traffic Management System Development of an Intelligent Urban Traffic Management System Riza Atiq bin O.K. Rahmat.
SixthSense Technology Visit to download
ParkNet: Drive-by Sensing of Road-Side Parking Statistics Irfan Ullah Department of Information and Communication Engineering Myongji university, Yongin,
Zhaoxia Fu, Yan Han Measurement Volume 45, Issue 4, May 2012, Pages 650–655 Reporter: Jing-Siang, Chen.
Signals, pavement markings, and proper turns QUIZ
Traffic Signal Operations: An Evaluation of Vehicle Detection Systems
Intelligent Transportation System
APPs FOR GPS Chad Sperry, GIS Center
Camera is useful in many applications
Signal and Image Processing Lab
Road Safety Behaviour Symposium: New technology, new connectivity
The Importance of ADAS Technology
Embedding Technology in Transportation Courses
Motion Planning for Multiple Autonomous Vehicles
Xiangyu Xu*, Jiadi Yu*, Yingying Chen†, Yanmin Zhu*, Minglu Li*
Vehicle Segmentation and Tracking in the Presence of Occlusions
Sensor Fusion Localization and Navigation for Visually Impaired People
Airport Parking Space Navigation
KEYLESS ENTRY PUSH TO START BACK UP CAMERA.
Evaluate the integral {image}
Presentation transcript:

Invisible Sensing of Vehicle Steering with Smartphones Dongyao Chen Kyong-Tak Cho, Sihui Han, Zhizhuo Jin, Kang G. Shin May 19, 2015

Camera + Various CV algorithms … Affordable, easy to use Camera helps detect lane markers for Detecting steering maneuvers http://www.phonearena.com/news/Nexus-5-likely-coming-with-an-OIS-camera_id47409 Affordable, easy to use + Various CV algorithms … https://www.youtube.com/watch?v=NJ8iXYZfUmA Autonomous vehicles use cameras to detect Lane markers Objects on road … Camera determines the functionalities of Performance is closely relate to the functionaility http://www.theguardian.com/technology/2013/jul/09/google-android-reader-why 1

CarSafe[MobiSys’ 13] app How’s Camera Doing? iOnRoad app https://play.google.com/store/apps/details?id=com.picitup.iOnRoad Citroën DS5 http://www.mycitroen.be/nl/citroen-ds5/innovatie/ CarSafe[MobiSys’ 13] app Performance of camera depends on the environment http://wallpaperswide.com/road_to_death_valley-wallpapers.html Smartphones with camera Built-in devices also use … Research works also mentioned … Our observation is that: the camera-based approaches 2

CarSafe[MobiSys’ 13] app How’s Camera Doing? iOnRoad app https://play.google.com/store/apps/details?id=com.picitup.iOnRoad Citroën DS5 http://www.mycitroen.be/nl/citroen-ds5/innovatie/ CarSafe[MobiSys’ 13] app Performance of camera depends on the environment Lighting Weather Pavement http://www.annarbor.com/news/residents-frustrated-with-lack-of-solutions-for-pothole-ridden-north-territorial-road/ Placement http://smartphones.wonderhowto.com/how-to/keep-your-smartphone-charged-during-power-outage-0140088/ However, this slides will show most of existing driving assistant systems still overly-rely on the usage of camera. Camera face down, if it cannot see, it definitely cannot work. 2

CarSafe[MobiSys’ 13] app How’s Camera Doing? iOnRoad app https://play.google.com/store/apps/details?id=com.picitup.iOnRoad Citroën DS5 LDW System http://www.mycitroen.be/nl/citroen-ds5/innovatie/ CarSafe[MobiSys’ 13] app How well a camera can see Bottleneck of application’s performance Real road condition Lighting Weather Pavement http://www.annarbor.com/news/residents-frustrated-with-lack-of-solutions-for-pothole-ridden-north-territorial-road/ Placement http://smartphones.wonderhowto.com/how-to/keep-your-smartphone-charged-during-power-outage-0140088/ Performance of driving assistant systems is undermined by a fundamental limitation of camera, which is. 2

Key Idea: Camera-free Sensing of Vehicle Steering Reliability -Resistant to changing environments Accuracy -Visual similarity will not effect the detection results Lightweight -Lower computational consumption The challenge is to find the right signature Which can directly gives us a boost of reliability, accuracy, and lightweight. “the right signature”: what we detect, how we detect 3

Differentiates maneuvers V-Sense Utilizes gyroscope Differentiates maneuvers Base on this idea, we propose V-Sense. V-Sense utilizes… To increase the accuracy… In the end… Works as a middleware 4

Sensing Steering with Gyroscope In left turns In lane changes Gyroscope in the smartphone Here is a logic jump 5

Understanding Vehicle Steering In left & right turns Detect bumps in gyroscope reading How to detect bumps? How to relate bump detection to different steering maneuvers? Lane changes Lane changes Gyroscope in the smartphone Here is a logic jump 5

Bump Detection Three-state algorithm: No bump → One bump→ Waiting for bump To determine if the following bump is source from the same maneuver, we detect the delay of next bump. T_bump = 1.5s; T_next_delay=3; Delta_s = 0.05; Delta_h = 0.07 ----- 会议笔记(4/1/15 16:11) ----- Peak detection. Spend more time on the analytical things. Show that the result is analytical reliable. ----- 会议笔记(4/28/15 11:34) ----- Dont need hyphend 6

Increase Accuracy: Steering & Curvy Road Lane change S-shape road Turn L-shape road We also need to increase the accuracy, because, the driving trajectory of steering and on curvy road could be similar. Horizontal displacement: WCURVY >WLANE , WCURVY >WTURN Integrate horizontal displacement at each time step 7

Calculate Horizontal Displacement W4 W3 W2 W1 Ts 2Ts 3Ts 4Ts Velocity v1 v2 v3 v4 θ1 θ2 θ3 θ4 Time Ts 2Ts 3Ts 4Ts Velocity v1 v2 v3 v4 θ1 θ2 θ3 θ4 Time Ts 2Ts 3Ts 4Ts Velocity v1 v2 v3 v4 Time The whole lane change process can be separated into multiple time slots; At each time slot, combine the changing angle with the previous result; We Figure: Horizontal displacement of lane change 8

Detect Horizontal Displacement on Curvy Road Let’s see the performance of this algorithm in the real world 9 Route image adapted from Google map: https://www.google.com/maps/place/2345+Lancashire+Dr,+Ann+Arbor,+MI+48105/@42.3099558,-83.7171271,17z/data=!3m1!4b1!4m2!3m1!1s0x883cac2d8254dfbb:0x2df93ff4aa13179b

Horizontal Displacement on Curvy Road Show the driving direction 10

Horizontal Displacement on Curvy Road Show the driving direction 10

Evaluation Setups Tested V-Sense on Samsung Galaxy S4 and S3 5 volunteers, 2 cars 40+h test 11

Reliability and Accuracy Resilient to different placements: Success Rate ----- 会议笔记(4/1/15 16:11) ----- 1. what could happen if misclassification happens. ----- 会议笔记(4/28/15 10:12) ----- success ratio for correctly detect the maneuvers 12

Computational Costs 13 ----- 会议笔记(4/1/15 15:57) ----- show bars. S4 is even better than S5. 13

Comparison with Other Applications ~1,000,000 installs ~50,000 installs ~50,000 installs 110 ratings Success Rate ----- 会议笔记(4/1/15 15:57) ----- Divde with ground truth. ----- 会议笔记(4/1/15 16:00) ----- parameters are different from previous slides ----- 会议笔记(4/1/15 16:11) ----- mention the evaluation condition. Specify the evaluation scenario. Specify highway, and local road. If you only drive on highway, the performance could be Success rate Caption: compare of lane change detection Figure: Success rate of detection of lane change in urban area 14

When Steering your Car... 16 Before we show the first application. Image: http://www.wakeforesttoday.com/voices-letters/note-to-self/4569-turn-signals-people.html#.VUPwONpViko

Application I: Detection of Careless Steering ----- 会议笔记(4/1/15 16:11) ----- 1. Specify the detail; 2. Change to careless. 17 Car icon: http://publicdomainvectors.org/en/free-clipart/Top-view-car-vector/2260.html

Turn Signal Detection Matched filter results > ground truth? Explain SAME Size!!! ----- 会议笔记(4/28/15 11:29) ----- Should have captions for all figures Use real image Add CAPTIONS! 18

Application I: Detection of Careless Steering ! ----- 会议笔记(4/1/15 16:11) ----- 1. Specify the detail; 2. Change to careless. 19

Application I: Detection of Careless Steering ----- 会议笔记(4/1/15 16:11) ----- 1. Specify the detail; 2. Change to careless. 19

Application II: Fine-grained Lane Guidance https://idroid.us/google-maps-app-for-ios-updated-with-optimized-for-ios-8.html GPS is unstable and its accuracy varies with environment GPS Accuracy http://365stagandhenweekends.co.uk/blog/ 20

Application II: Fine-grained Lane Guidance Lane tracking at intersection: Left to Right Right to Left ----- 会议笔记(4/28/15 11:50) ----- ADAPTIVE TO DIFFEREWNT CASES 21

Conclusion V Two proof-of-concept apps A reliable, accurate and light weighted middleware for detecting steering maneuvers Two proof-of-concept apps Careless steering warning Fine-grained lane guidance ----- 会议笔记(4/1/15 15:57) ----- ssss! 22

Q&A Thank you! 23 ----- 会议笔记(4/1/15 15:57) ----- include email. chendy@umich.edu 23