1. Simulation of a large (11ft
× 35 ft × 9 ft) conference room includes a variety of repositionable color sensors, 12 wirelessly
(Fig. 1a). This conference room is currently under physical controlled multispectral LED lights each with a circular diffuser
construction and is being designed to enable highly advanced
sensing, lighting, and control systems. The simulation accurately in a cone-shaped can, and a 2.3ft×4ft panel light in the middle
of the ceiling that can simulate skylight. The room contains no
reflects the geometry of walls, windows, and furniture in the windows to the outside or daylight.
space under construction, and is roughly accurate in terms of
The objective of this experiment is to demonstrate that with
the varying material properties (e.g., color, reflectivity and trans careful modeling of the actual light sources and color sensors
parency) of surfaces in the environment.
based on measured data (e.g., cone angle, spectral sensitivity, The conference room is simulated with 60 2ft×2ft light panel color channel crosstalk), the simulation of a candidate control
fixtures distributed across the entire ceiling, and a downward algorithm can be made to match physical reality both qualita-
pointed color sensor at the center of each panel. The objective tively and quantitatively, supporting the use of the simulation
of this experiment is to show how design iterations using the tool for making accurate predictions of lighting system behavior.
simulation framework allow the parameters in a candidate con
trol algorithm to be easily tuned to achieve a lighting designer’s
vision for fidelity to a given setpoint, color consistency, bright We note that while we instantiate our simulations with par-
ness consistency, and energy usage. ticular choices of environments, sources, sensors, and control