MuddSub

MuddSub is a student-run robotics team. We are currently developing an autonomous underwater vehicle (AUV) named Alfie which competes in the international RoboSub competition. In RoboSub, Alfie must navigate a challenging underwater obstacle course with full autonomy. This includes navigating through gates, bumping into targets, launching torpedoes through holes, and using acoustics to navigate to remote areas of the course.

 

This summer, we will work to continue improving Alfie’s hardware, electronics, and software. For software, projects will include developing computer vision, using reinforcement learning to learn how to complete certain tasks, and implementing SLAM (simultaneous localization and mapping) algorithms. The electronics will also be iterated to increase robustness and scalability. Mechanical projects will center on adding new actuators and sub mechanisms, while also working to make existing systems more reliable.

 

Essay prompt 

Which parts of the robot interest you most? For example, software, electronics, or mechanical design? Within that category, are there particular robotics problems you’d like to work on? Feel free to express interest in multiple projects/categories even if you have no experience!

 

Name of research group, project, or lab
MuddSub
Why join this research group or lab?

MuddSub is a student-run effort at Mudd that provides an opportunity to define and pursue subproblems that interest you. Building an AUV requires tackling problems in most modern fields of AI.

Representative publication
Logistics Information:
Project categories
Computer Science
Engineering
Circuit Design
Computer Vision
Machine Learning
Robotics
Student ranks applicable
First-year
Sophomore
Junior
Student qualifications

The skillsets valuable in MuddSub are precisely those valuable in open-ended research projects. More broadly: (1) curiosity and generosity, to sustain a team effort through many unexpected twists, (2) openness/adaptability, to develop new technical expertise as one's own, i.e., "what we do," rather than "what others do," and (3) organization and enjoyment of technical exploration for its own sake.

Time commitment
Summer - Full Time
Compensation
Paid Research
Number of openings
4
Techniques learned

Many! Some you may learn include:

 

Reinforcement learning, computer vision (both ML and classical), SLAM, electronics design, systems integration, navigation, mechanical design, prototyping, and more!

 

Contact Information:
Mentor name
Zach Dodds
Mentor email
dodds@g.hmc.edu
Mentor position
Mentor
Name of project director or principal investigator
Zach Dodds (really just the faculty conduit to the students in charge)
Email address of project director or principal investigator
dodds@g.hmc.edu
4 sp. | 38 appl.
Hours per week
Summer - Full Time
Project categories
Computer Vision (+5)
Computer ScienceEngineeringCircuit DesignComputer VisionMachine LearningRobotics