Lab Course: Build Your Own Demonstration Platform for Autonomous Cars

Organizer: M.Sc. Sina Shafaei, Dr. Stefan Kugele

Contact:

shafaei (@) in.tum.de (Organization and AI),

stefan.kugele (@) tum.de (Software and SOA),

vadim.cebotari (@) tum.de (SOA)

hafeez.osman (@) tum.de (Safety)

 

Modul: IN2106

Registration: Via Matching System and Submitted Application

Type: Master Lab Course

Semester: WS 2018/19

Time & Location: 03.07.011 / No weekly sessions

                                                                            


News (Registration)

  • The first official session of the lab course will be on "19.10.2018" (9:00-11:00) at room "03.07.011".
    There won't be weekly sessions for this lab course, but instead students will be in close coordination with each one of the organizers through the moodle and irregular meetings in order to deliver the objectives successfully.
  • There is one available seat in this lab course, if you're interested to be registered for the WS18/19, please send an email to "shafaei(@)in.tum.de" including your CV

  • The preliminary session was on Thursday 28th of June at 03.07.011, 13:00-14:00
  • Interested applicants must send an email to "shafaei (@) in.tum.de" (set the subject of your email to: ADPlatform) with a copy of their CV, transcripts of results and a brief explanation of their motivation, no later than 1th of July 2018. All of the applications will be examined carefully with the organizers and the selected candidates will be notified on 2th of July.

    If you get a confirmation email on 2th of July by the organizer, please place this lab course as your first priority at Matching System.

Content

In this lab course you're going to develop a demonstration platform for autonomous cars!

There are four domains which you need to get yourself familiar with beforehand:

0- Build your own car

First of all you need to have a model car which you can make it autonomous, right? We will give you the necessary hardware, motors, chassis, cables, etc. You got the weapons, time to fight!

1- Make it Autonomous

You will develop a lane following and obstacle avoidance system including an object detection appication from scratch by yourself which drives your car autonomously in an specific track, But how will you ensure that the passengers of your car will survive and you won't face any crash? (hint: check the second point)

2- Make it Safe

You need to develop some safety assurance mechanisms (some call it safety limiters) to ensure that your car is performing in a safe manner and the passengers feel safe in it and eventually, integrate them into your intelligent driving components, how will you do this integration? (hint: check the last point)

3- Service Oriented Architecture (SOA)

Your car is built of a set of embedded devices like raspberry pis, a distributed architecture, there exist many data sources, ultrasonic sensors, cameras etc. You may also want to have some cloud-based services integrated into your car, who knows? We propose SOA to you, to set up the publish subscriber structure of it in ROS.

Wanna know more about all of these? Come to our preliminary session!


What do we expect from you at the end of the lab course?

A software platform which can be used for autonomous cars in order to deploy the AI-based components in a safe manner and SOA-based architecture.You will use the demo car for testing and evaluating your developments.

 

What will be the final evaluation?

By the end of your work, you may be given a fully professional demo car (with real sensors and LIDAR) and you may be asked to demonstrate the capabilities and features of your platform on a predefined autonomous driving scenario. If the demonstration was successful then the job is done!


Workflow

Will be discussed during the preliminary session


Material