Sensor-based defect management for construction sites [summary]
This research project builds on, combines and extends the advances in generating 3D environments using laser scanners, collecting quality information about built environments using embedded sensors, and generation and utilization of semantically-rich Architecture/Engineering/ Construction (A/E/C) project models, in developing an integrated early defect detection system.

Collaborative Technology Alliance (CTA) [summary]
In collaboration with General Dynamics Robotic Systems (GDRS), we are developing advanced perceptual capabilities for robotic systems.

Automatic 3D modeling from range images [summary | details]
I am developing a system for creating 3D models of real-world objects without manual or mechanical aids.

Ground-based Multispectral Terrain Classification [summary | details]
As part of the MURI program, I helped develop a multispectral camera based on an acousto-optical tunable filter (AOTF). This device allows you to spectrally filter images in real-time. This means that you could take a look at just the "red" components of an image, for example. The wavelength or waveform of interest can be changed in real-time. I helped build a portable prototype of the system, which I used for my terrain-typing research. The AOTF was built jointly by the Carnegie Mellon Research Institute (CMRI) and CMU.

In order to achieve a higher level of understanding of the environment than previous unmanned vehicles, I looked at the problem of terrain classification. By knowing the type of terrain for a given area, the vehicle can make intelligent decisions about navigation through or around the terrain. For example, previously, it was not possible to differentiate flat ground from tall grass or mud from dirt, so the vehicle had to plan conservatively or assume certain terrain types would not occur. Terrain classification relaxes these restrictions and expands the range of terrain in which autonomous vehicles can safely operate.