Driverless-Software

Software Engineer – Perception Ref.: 701

The tracks in the driverless Formula Student competition are marked with traffic cones of different color. Our perception algorithms detect the position and color of those cones from lidar and camera data. To achieve this we use different pre-filtering techniques, feature extraction and deep learning.

Your tasks:

• Research and develop algorithms to detect cone color and position
• Evaluate different approaches to detection of cones in camera images
• Write ROS-nodes to integrate the algorithms into our system
• Optimize your algorithms for our computation unit

Your profile:

• Motivation to build a driverless race car
• Self-organized and reliable
• Communication skills and Teamwork
• Programming skills in C++
• Basic knowledge of machine learning principles
• Bonus: previous work with ML-framework, point clouds or ROS

Software Engineer – Mapping & State Estimation Ref.: 702

To be able to control a driverless race car, it is very important to know it’s exact position, velocity, heading, steering angle, etc. (its state) at any point in time. We estimate this with the aid of different sensors like an INS, wheelspeed sensors and steering angle sensor. We also use an algorithm for simultaneous localization and mapping (SLAM) to create a map of the racecourse and localize our car in it.

Your tasks:

• Research and develop SLAM algorithms
• Estimate the car’s velocity and heading using INS data

• Write ROS-nodes to integrate the algorithms into our system
• Evaluate your developed algorithms
• Optimize your algorithms for our computation unit

Dein Profil:

• Motivation to build a driverless race car
• Self-organized and reliable
• Communication skills and Teamwork
• Programming skills in C++
• Basic knowledge of stochastics
• Bonus: previous work with SLAM algorithms, sensor fusion or ROS

CI/CD Engineer  Ref.: 703

We maintain our code base on a hosted Gitlab server and use a continuous integration pipeline to ensure, that the code always works and also has a common code style. By moving to our own server, we would be able to also test our algorithms with a GPU, as well as implement a system to archive our recorded rosbag data.

Your tasks:

• Maintain our CI pipeline
• Create a concept, build and maintain a server to support software development

• Create a concept and realize it for deploying our software on the

Dein Profil:

• Motivation to build a driverless race car
• Self-organized and reliable
• Communication skills and Teamwork
• Shell skripting skills
• Experience with UNIX-servers
• Bonus: previous work with docker, gitlab-ci or ROS