Physical Analytics (previous) - overview



Physical analytics is the understanding of the inner nature of things based on measurement, modeling, insight, and deduction. The analysis is done through identification of relationships and behaviors within a physical model, context, or scenario.

Our efforts are focused on the following topics.

Distributed sensor network for energy management

High density data from wireless or wired sensor networks is processed in real time to asses the environmental conditions in buildings. Among others, some variables of interest are temperature, humidity, air flow, and pressure. This work leverages the Measurement and Management Technology (MMT) for modeling and analysis. Our latest data collection tool, the IBM Low-Power Mote Technology (LMT), is a wireless solution that allows fast and easy deployment.

Mote---full assembly Mote---partial assembly
IBM Low-Power Mote Technology


Solar analytics

Short-term forecasting technology (seconds up to hours) of renewable energy production by solar panels. We use sky imaging camera for cloud tracking and satellite imaging processing for solar radiation assessment. These methods allows forecasting the available energy and maximum utilization of intermittencies for demand control of appliances and building operations.

Solar Panels Installation


Water monitoring and modeling

High density monitoring of Keller Lake in MN was performed using a robotic boat system where the dissolved oxygen, conductivity, temperature, and pH value of water were recorded at different depths over an extended area. The data was used to model the spatial distribution of those physical parameters and to understand how pollution affect the ecosystem health.


News and Events

Deployment of IBM LMT at the Metropolitan Museum of Art in NYC