Master Thesis - Optimal Tuning
Master Thesis Opportunities – Gothenburg, Sweden
Research on autonomous vehicles is a topic of significant interest worldwide. The goal is to let the driver choose when to drive actively, and when to let the car drive itself. The outcome is two-fold: First, this could make traffic safer, since inattentive drivers are no longer a traffic hazard; and second, it can relieve the driver of unstimulating driving experiences, such as stop-and-go traffic. Important steps towards fully automated vehicles consist of a large number of active safety systems designed to warn or support the driver by, for example, adapting the speed to the vehicle ahead or breaking/steering to avoid a collision.
For some applications of autonomous vehicles it is considered critical to localize accurate in a sensor map. State of the art algorithms for performing such localization usually contains a great deal of tunable parameters which will determine the performance of the algorithm. Typically the parameters have very complex, non-linear relationships making them incredible
hard and time consuming to tune manually.
We are looking for 1-2 Master Students with a passion for the industry and high aspirations to contribute to the field. You will study the possibility to automatically tune a localization algorithm using different optimization techniques and they will be expected to:
• Do a literature survey on available classic and stochastic optimization techniques
• Implement optimization algorithms in python
• Analyze results and draw conclusions regarding the different techniques
We are looking for students with some background in:
• Optimization, preferably both classic and stochastic
• Python and Matlab
• Working with computer clusters
Further information and contacts
Final application date: November 1st 2017. Please send in individual applications. If you wish to partner with someone, simply note that in your application.
Planned start: From now until February 2018, with some flexibility.
Duration: 30 ECTS
For questions regarding the project, please Contact Louise Bichler, firstname.lastname@example.org.