2017 ACM SIGSPATIAL Workshop on High-Precision Maps and Intelligent Applications for Autonomous Vehicles
(AutonomousGIS 2017)

Best Paper Award: Lane Boundary Extraction from Satellite Imagery

New Keynote Speaker Confirmed: Christine McGavran, Google Engineering Lead for Geo-Automotive

November 7, 2017 (Tuesday) — Redondo Beach, California, USA
in conjunction with ACM SIGSPATIAL 2017

LEARN MORE

Workshop Description and Scope

Autonomous vehicles (aka self-driving cars) and their equipped intelligent systems are among the most exciting and innovative technologies in transportation. In order to operate safely, autonomous vehicles require high-precision geospatial datasets and maps, which contain significantly more detailed road information and the true ground (absolute accuracy) than those found in current conventional geospatial resources for driving. With the integration with artificial intelligence, machine learning, spatial data mining, image processing and other technologies, many innovative intelligent applications are fast growing. This international workshop aims to bring both academic researchers and geospatial industry engineers & research scientists together to present, discuss, and share the state-of-the-art scientific research and technology advancement on related fields for autonomous vehicles. The workshop participants will also have an opportunity to share the state-of-arts technology and brainstorm their visions of high-precision maps and development on future intelligent applications for autonomous vehicles.

Workshop Schedule (Nov. 7, 2017)

Time Presentation
8:00 am - 9:00 am Breakfast
9:00 am - 9:10 am

Welcome and announcements


9:10 am - 9:30 am

“Automatic Calibration of In-Vehicle Portable Inclinometer Using Mobile Phone for Vehicle Navigation”
Teng Zhong, The University of Hong Kong,
Anthony Gar-On Yeh, The University of Hong Kong
9:30 am - 9:50 am


“Lane Boundary Extraction from Satellite Imagery”
Andi Zang, Northwestern University,
Runsheng Xu, Northwestern University,
Zichen Li, New York University,
David Doria, HERE North America
10:00 am - 10:20 am

“An Automatic Pipeline for Mapping Roads from Aerial Images”
Yun Ye, The Ohio State University
Alper Yilmaz, The Ohio State University

10:30 am - 11:00 am

Coffee break
11:00 am - 11:40 am
Keynote: “Delivering Useful Maps at Scale to Cars”
Christine McGavran, Google Engineering Lead for Geo-Automotive
11:40 am - 12:00 pm

“Accurate Vehicle Self-Localization in High Definition Map Dataset”
Andi Zang, HERE North America,
Zichen Li, HERE North America,
David Doria, HERE North America,
Goce Trajcevski, Iowa State University

12:00 pm - 12:30 pm

Group Discussion and Wrap-up

Keynote Information (11:00 am)

Bio: Christine McGavran is currently a Principal Software Engineer at Google, leading integration of Google’s mapping services with vehicles. Prior to Google, Christine focused on mapping and navigation services for both Apple Maps and Dash Navigation, including contributions to 16 patents. She also spent much of her career on game development, including non-real-world mapping as Development Director for SimCity 3000. Christine graduated from Dartmouth College with a degree in Computer Science.
Abstract: The industry has been evolving from maps for humans to maps for vehicles. Vehicle-focused maps have existing use cases toward improving driver safety, and are aspiring toward supporting more advanced applications with higher accuracy and detail requirements, including autonomy. This talk will discuss the risk of relying on highly detailed maps, given the challenge of delivering and maintaining these maps at scale. We propose iteration toward a more intelligence-centered approach for using and improving a map, along with a need for common standards, software, and semantics for collaborative mapping.
Overview
  • Workshop Format and Submission Link

  • Full research paper: 8-10 pages
  • Short research paper or industry demo paper: 4 pages
  • Vision or statement paper: 2 pages
  • Submission Website: AutonomousGIS 2017 on EasyChiar

Topics (but not limited to)

  • Automate data pipelines for high-precision maps using deep learning
  • Computational models of the road surface and surrounding environments
  • Feature extraction (e.g., lanes) from remote sensing imageries or LIDAR point clouds
  • Automatic road sign data collection methods and systems
  • Passenger pick-up in a garage or indoor environment without accurate GPS information
  • Adoption of precise navigation system to robotics applications
  • Application of relative coordinate system in GIS for regional navigation
  • Computer visions and pattern recognition for stereo camera images
  • System integration and navigation for self-driving cars
  • Smart sensor networks and intelligent parking technologies
  • High-precision GPS for autonomous vehicles
  • Context-based routing
  • Automated delivery
  • Artificial intelligence applications for autonomous vehicles
  • Spatiotemporal data mining for autonomous vehicles

The workshop will focus on intensive discussions and knowledge sharing on this field. Thus, we will accept three types of contributions. First, full research papers have to present a specific problem or topic in the context of high-precision maps and intelligent applications for autonomous vehicles and will be selected based on peer-review results adhering to classical scientific research quality criteria. Second, we also welcome short research articles or industry demonstrations of existing or developing methods, toolkits, platforms, and best practices for creating and updating high-precision maps and developed intelligent applications for autonomous vehicles. Third, a vision for future directions or an overview statement on gaps and challenges for autonomous vehicles are also welcome.

Workshop Structure Highlights

The workshop will consist of one or two keynote presentations, 4-hours paper presentations and a following panel discussion or a round-table brainstorming session. About 5 presenters will be selected to give a 20-minutes research talk on their full or short papers, or a 10-minutes statement on their idea following by a 5-minutes Q&A. The panel discussion or the round-table brainstorming will bring all the workshop participants together to share their thoughts on key issues and geospatial challenges on self-driving cars.

Keynote Speakers

  • Keith Clarke, Professor in Cartography & GIS at the University of California-Santa Barbara.
  • Christine McGavran, Google Engineering Lead for Geo-Automotive.

Organizers

  • Song Gao, University of Wisconsin-Madison (song.gao@wisc.edu)
  • Gengen He, University of Nottingham Ningbo China (gengen.he@nottingham.edu.cn)
  • Jing Dai, Google (jddai@google.com)
  • Xin Chen, HERE Technologies (xin.5.chen@here.com)
  • Yang Xu, MIT SENSEable City Lab (yangxu@smart.mit.edu)

Program Committee

  • Shih-Lung Shaw, University of Tennessee
  • Keith Clarke, University of California-Santa Barbara
  • Paul Torrens, New York University
  • Yan Huang, University of North Texas
  • Craig Hancock, University of Nottingham Ningbo China
  • Toru Ishikawa, University of Tokyo
  • Qunying Huang, University of Wisconsin-Madison
  • Abdeltawab Hendawi, University of Virginia
  • Important dates

  • Paper Submission Due: August 31, 2017
  • Paper Submission Extended to: September 15, 2017
  • Notification of Acceptance: October 1, 2017
  • Camera Ready Paper Due: October 9, 2017
  • Workshop Date: November 07 2017