Learning based SLAM & Spatial AI

Currently open to new students?

Yes
No

Description

We aim to build a generalizable, robust, and accurate learning-based SLAM system that lays the foundation for spatial intelligence.
Our previous work,  MAC-VO , won the ICRA 2025 Best Paper Award & Best Perception Paper Award. For more information, see here:  https://theairlab.org/highlight-macvo-bestpaper/ 


Student Learning Objectives

  • Publish a paper at the top conference (IROS, ICRA, CoRL, etc) and top journal (T-RO, Science Robotics)


Skills Required

Experience in at least 2 of below:
  • PyTorch
  • State estimation/SLAM/IMU kinematic
  • Deep Learning Model Acceleration
  • ROS and C++
  • Really Good Programmer

Classes Accepted into Project

Senior (if you're really good)
Graduate Student

Compensation

Units 9
Units 12
Pay

Contact

We currently only accept in-person collaboration. Please do not contact us if you're mostly remote