Below is a selection of our current projects. A project is open for funding and IP negotiation unless otherwise noted. Click the icon to view a more detailed description of the project's background and goals.

Current Projects

Software Design for Device and Network Reliability Models

We have a novel approach to network and device reliability models under highly general conditions, which include random workload. This work covers many scenarios in device and network reliability, and this project will implement these methods and build a software package for use in design and monitoring of any network across industries.

Pollutant and Particulate Clumping

When counting particulates is needed, an accurate count is essential. Clumping is a phenomenon that can bias the count by forcing a sensor to count a group of particulates as one, potentially misidentifying or miscounting. We apply novel probabilistic research in dependency theory to study the probability of finding clumps of various sizes in samples. This research is applicable to industrial hygiene, medicine, aerospace, and manufacturing.

Better Anomaly Detection through Fuzzy Numbers

Anomaly detection is on everyone's mind, especially in time series. Real-time decisions are essential in the world of IoT and high-granularity data. Traditional machine learning models require far too much training time, computational power, and have lost all interpretability in the "black box" they generate, making true understanding impossible. This project returns to the roots of time series analysis and proposes a brand new software library based on continuing current research in the area of fuzzy ARIMA modeling for time series anomaly detection and classification.

Structural Autocorrelation

Events that occur in a sequence, such as failures, purchases, or arrivals are rarely completely independent. The notion of autocorrelation is meant to measure the amount of statistical correlation between a current event and some "lag" or previous data point. However, this measure is insufficient to answer questions about the structure of the effects the past has on our future. This research builds from our dependency theory and seeks to uncover statistical methodologies that aid us in understanding the effects of past events on future ones.