Modeling and Prediction of Event Data from IoT Enabled Equipment
The objective of this project is to analyze equipment operation data and develop models to predict events of interest. The dataset is a tabulated record from multiple devices. Each device has its own period for which event data is available.
Based on this data, our aim is to develop a model for predicting the event occurrence in a device within a day’s period. The event from a device is predicted in two ways: (a) Probability of occurrence in the next hour, and (b) most-likely time of next occurrence with a margin of error. The prognostic model is developed based on the history for each device while accounting for variations.
Engine Testing Data Management and Analytics
The objective of this project is to process and analyze engine operation data collected during development testing. The dataset is relatively large. It contains the measurements of a large number of physical variables from multiple engines. Each engine (or twin engines) has its own period for which data are available.
Based on this data, our aim is to:
- Conduct data-pre-processing to group and clean the data for easy visualization and further modeling
- Conduct descriptive analysis to provide various visualization of the data and possibly expose the trend and patterns in the data
- Conduct the feasibility study of forecasting the engine degradation and predicting the failure event occurrences.
The long-term goal is to establish a close collaborative relationship with the company and build effective predictive and prescriptive models for engine operation, development, and maintenance.
IoT Analytics for Predictive Maintenance
- Leverage cutting-edge academic research
- Develop methods to address industrial analytics challenges
- Apply and test these methods using real-world IoT datasets
- Established remote asset monitoring testbed based on ice machines
- An advanced condition monitoring method using multivariate Gaussian processes is established. A research publication is submitted.
- Predictive analytics method based on mixed effects models are established.
Smart and Connected Home Project
Gain insight into how homeowners’ perspectives towards and usage of smart home devices evolve over the entire adoption lifecycle, from pre-acquisition to extended use.
A field-based study involving 30 homeowners and conducted over a period of several months.
Final survey launching. The project concluded in late 2017. Survey findings will be summarized and reported to members.
IoT Endpoint Security Project
Research and synthesize practical guidelines and best practices for IoT endpoint security that companies can use as a guide for developing secure IoT-enabled products, services, and solutions.
The project concluded on December 2017. The results of the report have been reported to IoT members.