Unexpected response to rarely seen sensor inputs have caused catastrophic (driverless) car crash. Could we design an automatic testing mechanism to reduce the chances of such potentially disastrous behaviors from occurring? To start answering this question, we will explore a simple planning and learning approach to automatically and efficiently generate simulated sensor inputs that are likely to trigger undesirable behavior.
- Fluent in C++/Python (higher preference for C++) programming.
- Have a good understanding of abstract data structures.
- Have a good understanding of basic probability.
- Experience with robotics simulator (e.g., Gazebo or VREP) is a plus but not necessary.
Remuneration is available. This project may lead to internships in a self-driving car company and can be expanded into a PhD topic.
Starting date is flexible.
If you're interested in this project or would like to know more details about the project, remuneration, and internship opportunity, please email me at email@example.com .