Machine Learning Thesis Proposal

  • Remote Access - Zoom
  • Virtual Presentation - ET
  • Ph.D. Student
  • Machine Learning Department
  • Carnegie Mellon University
Thesis Proposals

Active Robot Perception using Programmable Light Curtains

Most 3D sensors used in robotic perception, such as LiDARs, passively scan the entire environment while being decoupled from the perception system that processes the sensor data. In contrast, active perception is an alternative paradigm for robotics where a controllable sensor adaptively focuses its sensing capacity only on the most useful regions of the environment. Programmable light curtains are a recently-invented, resource-efficient, active sensor that measure the depth of any user-specified surface ("curtain'') at a significantly higher resolution than LiDAR. The main research challenge is to design perception algorithms that decide where to place the light curtain at each timestep, tightly coupling sensing and control in a closed loop.

This thesis lays the algorithmic foundations for active robot perception using programmable light curtains. We investigate the use of light curtains for various perception tasks such as 3D object detection, depth estimation, obstacle detection and avoidance. First, we incorporate the velocity and acceleration constraints of the light curtain into a constraint graph; this allows us to compute feasible light curtains which optimize any task-specific objective. Then, we develop a suite of algorithms using a variety of tools such as Bayesian inference, deep learning, information gain and dynamic programming that intelligently control the light curtain device to accurately perceive complex and dynamic environments.

While our prior work has focused on using light curtains to estimate "where objects are", for future work, we propose using light curtains to estimate "how objects move" i.e. estimating the velocity of dynamic objects. We propose using Bayes filtering techniques based on particle filters and occupancy grids for velocity estimation and show preliminary results. Active velocity estimation paves the way for solving various perception tasks like trajectory forecasting, obstacle avoidance and SLAM using light curtains.

Thesis Committee:
David Held (Co-chair)
Srinivasa Narasimhan (Co-chair)
Katerina Fragkiadaki
Wolfram Burgard (University of Freiburg)

Zoom Participation. See announcement.

For More Information, Please Contact: