We are looking for students for our team!
Follow us and compete in the 2017 Carolo Cup challenge!

For the preparation of the 2017 Carolo Cup competition, we are offering the following student projects:

Topic 1: Evaluation of ML-learning-based Lane-Following and Vanishing Point-based Lane Following
The lane-following capability of the miniature car is based on a peculiar feature extracted from the camera feed: the "Vanishing Point" (VP), i.e. the virtual point where the lane markings from both lanes would eventually meet.
This calculation can be done in two ways: analytically calculating the VP for each frame, or feeding the frames to an Artificial Neural Network which returns the VP coordinates.
The goal is to study and compare the two approaches gathering and illustrating significant data on how they perform singularly and against each other, and then to evaluate which one is best suited for use in a competitive setting,​ documenting appropriately the whole process.

Topic 2: Integration and Evaluation of biologically inspired OpenDLV on Scaled Cars
OpenDLV is a software suite based on OpenDaVINCI and inspired to how biological systems manage and react to the presence of stimuli coming from the surrounding world. 
It is used in the ReVeRe laboratory at Chalmers and powers up the now-autonomous Volvo FH16 truck that participated to the 2016 Grand Cooperative Driving Challenge in Helmond, Netherlands.
The goal is to integrate the OpenDLV software into the miniature car, and then gather data on how it performs on a limited computing platform in order to evaluate whether it is suited for use in a competitive setting,​ ​documenting appropriately the whole process.

Topic 3: Fast, Robust, and Dynamic Sideways Parking Algorithm
The sideways parking algorithms used in commercial systems are based on known models and equations, and are able to adapt to the distance of the vehicle from its surroundings (lane markings, obstacles, other vehicles, etc.) in order to park safely and without the need of significant adjustments after the main movement.
The miniature car needs to match these requirements too, and so the goal is to study the freely available resources on which this class of algorithms is based, to define a model for said algorithm, to implement it in the car and then evaluate and test it in order to verify that it can take into account its surroundings and use this knowledge to park with the same level of confidence of real-life systems.​ 
The students are expected to also produce the ​appropriate models and documentation regarding their envisioned solution, in order to adhere to software engineering practices and to simplify the process of analysis and correction, and the results from the systematic evaluation to be implemented in the car itself.

We are offering these topics to student groups of two persons collaborating on these projects. 
If you are interested in one of the aforementioned topics, please prepare a project proposal and project plan (cf. template) where you point out how you plan to address the suggested topics (cf. also the course syllabus for DIT 669). Please submit your proposals not later than September 04, 2016 to Christian Berger (christian.berger@gu.se).