Contact me

Carnegie Mellon University
5000 Forbes Ave, Pittsburgh, PA 15213
Smith Hall EDSH 223
kitchenette [at] gmail [dot] com

Katherine Chen

Speedily a tale is spun, with less speed a tale is done...

Research Projects

Some projects I've worked on or been involved with during college and graduate school.


Modelling Wireless Coverage with Gaussian Processes

9.2006 - present
CMU Robotics Institute, SELECT Lab Advisor:Carlos Guestrin

This is on-going research for my masters degree. Wireless systems are a growing presence in our tech-ed out society. Businesses are installing building-wide networks and even cities are installing city-wide networks. One concern is how to provide good quality of service (QoS) where people need it the most. Some things we'd like to focus on include:

  • how to monitor the signal quality at every location in the space
  • how to accurately predict locations of bad signal quality
  • how to find locations of good signal quality (hot spot detection)
The current research in this topic focuses on using signal pathloss models which require a lot of EE-specific knowledge about how radio signals propagate through different media, around corners and in hallways. We'd like to employ a probabilistic approach mixed with these domain-specific pathloss models to create more accurate models of signal quality.


Ultrasonic Sensors for Obstacle Avoidance on K10 Robot

6.2005 - 12.2005
NASA Ames | Intelligent Robotics Group (IRG)

K10 is the name given to a type of lunar robot designed to work on construction projects in conjunction with human beings (EVAs) and other robots. It is a rather fast moving robot (~1 meter/second top speed), can move in any direction, and has lots of sharp corners. So to make it safer to work around, we installed a ring of ultrasonic sensors to allow it to detect humans and large rocks surrounding it. This was a summer internship project under NASA's educational associates program.


Correlation Estimation from SAR Images

12.2003 - 8.2004
Radar Interferometry Group
Project Poster

Synthetic Aperture Radar (SAR) is a technique that uses satellites to obtain images of the ground (in this case ice sheets in Antarctica) by measuring correlation of radar echoes. These images can be used to infer how much ice accumulation has occured. Sometimes, phase distortion (so-called phase fringes) can throw off correlation measurements. Presented is a technique for removing phase fringes and results showing more accurate correlation measurements resulting from using this technique. This was a project I did with Prof. Howard Zebker as an undergrad and presented at the REU poster session (REU is a summer research program for undergrad EEs at Stanford).