16.S688 NEET Autonomous Machines, Junior Seminar
Fall 2018
Instructors: Sertac Karaman, Gregory Long, Ross E. Allen, Mark M. Mazumder
Lecture:
Announcements
RACECAR IAP
Hi NEET students, happy New Year!Please consider taking the 6.S184/16.S685 RACECAR IAP if you are interested in more complex and exciting autonomous navigation and racing challenges! Description and link below:
http://student.mit.edu/iap/nc1.html
RACECAR (Rapid Autonomous Complex-Environment Competing
Ackermann-drive Robotics)
Michael Boulet, Andrew Fishberg, Mark Mazumder, Nathan Hughes,
Jason Nezvadovitz, Sertac Karaman
Modern robots tend to operate at slow speeds in complex environments, limiting their utility in high-tempo applications. In the RACECAR course, you will be tasked with pushing the boundaries of unmanned vehicle speed. Participants will work in teams of 4-5 to develop dynamic autonomy software to race a converted RC car equipped with LIDAR, a stereo camera, an inertial measurement unit, and embedded processing around a large-scale, "real-world" course. Working from a baseline autonomy stack, teams will modify the software to increase platform velocity to the limits of stability. The course culminates with a timed competition to navigate a racecourse. The 2019 course will emphasize machine learning approaches, such as a pixel-to-actuator deep neural network control architecture. Classes will provide lecture overviews of relevant algorithms and lab time with instructor-assisted development. Participants should plan on 4-10 hours per week of self-directed development. Students must have experience with software development. Past exposure to robotics algorithms and/or embedded programming will be useful. See also: https://www.youtube.com/watch?v=p18879Dji4c .
To subscribe, please send an e-mail to racecar-iap-course-subscribe@mit.edu with a brief description of your programming/robotics experience.
If you would like to register for credit, please also register through WebSIS in either 6.S184 or 16.S685.
Announced on 04 January 2019 9:40 a.m. by Mark M. Mazumder
Post-thanksgiving seminar schedule
Hi everyone, great job on completing the line-following race this week! A quick note on the plan for the rest of the semester:On Wednesday 11/21 there will be no class.
After Thanksgiving break, we will switch to a different format, where you are welcome to work on projects of your own choosing.
Some ideas: optimize the line-follower for speed or precision, add obstacle avoidance through the front camera, perform more complex sky-writing, or use deep learning to visually servo the drones via object detection or human pose estimation. I will be around during our usual class times to help and advise on any of these topics, or if you have another topic in mind you would like to explore. You are all welcome to change your teams or work individually, per your own interests.
You all have successfully accomplished everything we expected of you for the seminar! Again, great job, and please reach out with any questions. Best,
Mark
Announced on 15 November 2018 6:39 p.m. by Mark M. Mazumder
Additional tutorials
Hi all,If this is your first time using a shell, git, or ssh, here are a few tutorials to help acquaint yourself with the basics. There's also a full lecture on ROS at the bottom if you want to get a head-start on this week's material.
https://github.com/mmaz/neet/blob/master/resources.md
Cheers,
Mark
Announced on 11 September 2018 9:54 a.m. by Mark M. Mazumder
Week 2 Class Prep
Last week we dove straight in by getting our hands on the drones and performing some remote controlled, i.e. "manual", flight tests. Hopefully this conveyed some sense of the difficulty of piloting a non-autonomous quadrotor and motivates why we want to automated certain processes in the flight controller (e.g. position keeping). In order to start implementing such autonomy, we need to first understand the architecture of the drone as a system of processors, sensors, actuators, flight software, and communication protocols as well as develop a common vocabulary for each of these components. In preparation for this weeks class, please watch the following video series that will introduce some of these concepts:
https://bwsi-uav.github.io/website/intel_rtf_architecture.html
Announced on 10 September 2018 5:12 p.m. by Ross E. Allen
Thursday session location
Hi all, for students in the Thursday session we will meet at 9am in 31-120. See you tomorrow!Mark
Announced on 05 September 2018 10:36 p.m. by Mark M. Mazumder
