facebook tracking

Master Thesis - Multi-SLAM: Probe sourcing of HD-map elements

How can affordable sensors in consumer cars be used to create High-Definition maps for use in future self-driving cars?

Background 

Standard Definition (SD) maps are the type of semi-topological maps that electronic navigators use. They describe the road network topologically, meaning how road segments are connected, and their geometric accuracy is usually only good enough to separate one road from another using GPS. Hence, they are not used in the context of autonomous driving for localization or vehicle control.

In contrast, High Definition (HD) maps contain more detailed geometric information about the driving environment such that accurate localization and safe path planning is supported. Additionally, these maps provide important information, e.g., the layout beyond a curve or how human drivers typically drive (natural driving profile) on the road that can not be easily provided in real-time by the onboard sensors. Compared to SD maps, the higher requirements on geometric detail make these maps more difficult and expensive to construct and maintain. 

However, some parts of an HD map could be efficiently crowd-sourced using sensor data from a fleet of manually driven vehicles. The basic idea is that, although it is hard to source the data in real-time from one sensor equipped vehicle, it should be possible to collect and aggregate data from a fleet of vehicles to achieve the required accuracy after combining it. For example, it should be possible to estimate the geometry of the road environment and to build statistics about typical driver behavior by aligning and aggregating sensor data from multiple vehicles travelling the same road section.

Project Description 

The goal of this thesis is to develop methods to crowd source certain elements of an HD map by collecting sensor data from a fleet of vehicles and aggregate the information from these into a single coherent view. The main focus is on describing the geometry of lane markers, content and position of a subset of traffic signs, and natural driving profiles. This information may be aggregated, on top of an existing SD map, to support with link connectivity and coarse level geometry. The scope of the thesis is limited to mapping a confined area, but the methods developed should ideally scale well when considering larger regions.

The primary challenge lies in correctly aligning the sensor data from the individual vehicles (data association) such that the data of the same element is aggregated correctly, but there are multiple secondary challenges e.g. regarding how to extract the map elements from sensor observations, that also may be dealt with.

 Qualifications 

 We are looking for 2 students, preferably with good knowledge of

  • Simultaneous localization and mapping (SLAM)
  • Data association
  • Optimization
  • Filtering and data fusion
  • Sensor models for cameras, lidars, and radars.
  • Python programming language

 

Further information

Please send in individual applications with CV, motivational letter and grade transcripts. 

Planned start: January 2022, with some flexibility.

Final application date: 15 of November 2021. Screening of candidates will happen continuously, so please submit your application early.

Duration: 30 ECTS 

For questions regarding the project, please contact: erik.stenborg@zenseact.com or karl.rundstedt@zenseact.com.

Additional information

  • Remote status

    Flexible remote

Or, know someone who would be a perfect fit? Let them know!

Gothenburg, Sweden

Lindholmspiren 2
417 56 Göteborg Directions View page

Making safe and intelligent mobility real.

At Zenseact, we lead the global movement of crafting tomorrow's mobility with the software platform of choice. Our mission is to “Make safe and intelligent mobility real, for everyone, everywhere”. This statement marks our conviction and dedication to bring autonomous driving out on the streets for real and is at the center of everything we do.

We could not dream of achieving this without our great teams of very talented people. We are on this journey together and our agile way of working is reflected throughout our entire organization; it is part of our culture and how we work, develop and grow together.

Teamtailor

Applicant tracking system by Teamtailor