I took AI4R in the Fall of 2019, and thoroughly enjoyed the course format and the lectures. The lecture format is designed for an intensive 6-week course, much shorter than the 16 weeks we get in the OMSCS program. However, don’t let that fool you – you’ll still need to stay on top of the projects and coursework. I did think this was one of the more fun courses to take, and felt I could have taken a second course (along with working full-time).
The homework assignments are relatively straightforward,if you put effort into the lecture quizzes. So note – do those quizzes and understand what’s going on there! The instructors also provide solutions with some explanations, but I HIGHLY suggested actually taking a stab at them yourself, and use these solutions only for learning and confirming your homework solutions are correct.
The projects, while very fun and interesting, increased with difficulty over the course of the … course. The first few are fine (workload wise) and the allotted time (I think 2 weeks for each_ they give is plenty. I got a little big in the head for the last one, and definitely did not start this one with enough time to get the value I got out of the other projects.
The best part of the course was the opportunity (for extra credit, too) to get some hardware, build something that had software/hardware interface, and implement something we learned in the course. I choose the Elegoo Robot Car kit, and I implemented PID control of the speed of the car. I hope to add on to this, and implement PID for controlling turning movement of the car, and eventually implementing SLAM algorithms.
Topics in the course included:
- Kalman filters and particle filters for localization of a moving object (with and without noise).
- Using A* search algorithms for path planning
- PID Control of a robot’s movement
- SLAM algorithm for 2D and 3D movement