Three ACID Challenge Problems
This is a two-part talk which describes opportunities at the intersection of machine learning, data-driven control and the use of cloud services for a new generation of Intelligent Physical Systems.
1. Energy Systems: Data Predictive Control - Bridging Machine Learning and Control Systems In December 2014, the average price of wholesale electricity in the PJM market surged from $31/MWh to $2680/MWh - an 86X increase in 5mins. Demand response (DR) is becoming increasingly important as the volatility on the grid continues to increase. Current DR approaches are predominantly manual and rule-based or involve deriving first principles based models which are extremely cost and time prohibitive to build. We have developed DR-Advisor, a Demand Response Advisor and recommendation system for energy flexibility in large buildings which involves predicting the demand response baseline, evaluating fixed rule based DR strategies and synthesizing DR control actions during price spikes. Built upon DR-Advisor is IAX, an Interactive Energy Analytics engine - think of it as a Siri for querying buildings' energy use. We are developing IAX to procedurally generate energy dashboards for open-ended questions. More info at: http://mlab.seas.upenn.edu/projectsites/dr-advisor/
2. Autonomous Systems: A Driver's License Test for Driverless Vehicles Autonomous vehicles (AVs) have driven millions of miles on public roads, but even the simplest scenarios, such as a lane change maneuver, have not been certified for safety. As there is no systematic method to bound and minimize the risk of decisions made by the vehicle's decision controller, the insurance liability of autonomous vehicles currently is entirely on the manufacturer. I will describe APEX, a tool for autonomous vehicle plan verification and execution across a variety of driving scenarios. We will see the use of synthetic environments such as computer gaming to train and evaluate machine learning and decision control algorithms in future AVs.