Cyber-Physical Systems:

Computational Materials

We are investigating new gestural paradigms for interaction with pervasive computing environments.  We envision that future ubiquitous computing environments will be filled with distributed embedded computing systems.  In particular, we envision that distributed sensor-actuator networks will be embedded within building structures and walls, allowing windows to open, air to circulate, and shades to be drawn or opened.  To control such computational materials, we are investigating distributed pattern recognition algorithms that can recognize gestures using a network of distributed simple sensors and actuators - forming a Smart Wall.

We have thus far built a multicast protocol that self-organizes islands of sensor/actuators who recognize the same gesture. 
We have simulated and implemented this algorithm in an actual smart wall.  More details can be found at the Correll Lab's Web site.

For more on computational materials, see the recent Workshop on Robotic Materials held at CU Boulder on March 2017.

Our work has been published as follows:

The "Holodeck" Project in Cyber-Physical Systems:

Our idea is to actuate a physical array or grid of moving pixels, which we call moxels, using Graphics Processing Units, or GPUs.  We find that this GPU-based solution is cost-effective and scalable, such that large grids could be controlled to render physical surfaces in a Holodeck-like environment.  Moreover, our GPU-based approach allows us to leverage existing graphics rendering software technology to assist in the physical rendering.  However, non-trivial modifications need to be made to the graphics rendering software.

An early version of this work was published at the 1st Workshop on Cyber-Physical Systems (WCPS), "
Towards Cyber-Physical Holodeck Systems Via Physically Rendered Environments (PRE’s)".

The Beijing Olympics Opening Ceremony featured a version of our moxel technology on the grand scale (right top picture), but we admit that the moxels were human actuated instead of GPU actuated (right bottom picture).

                                        boxes 1
Olympic boxes 2

home page