Sponsors and partners

Problem descriptions

The Black Arcs (TBA) builds agent-based microsimulations to explore how changes impact a community. These microsimulations use a synthetic population layered on top of real geospatial data and simulate travel activities using an implementation of Miller and Roorda’s Travel Activity Scheduler for Household Agents (TASHA). One emergent property that can come from these microsimulations is a synthetic contact matrix generated from the activity of the agents over a period of time.  Contact matrices have proven to be a valuable tool for predicting the spread of infectious diseases, with COVID-19 being a recent example.  Specifically, these contact matrices can directly inform parameters in ordinary differential or network based models of disease spread. This challenge looks to investigate how this type of disease modelling can be applied to synthetic contact networks generated in TBA’s microsimulations in order to explore a variety of infection scenarios.

The City of Fredericton operates a parking garage that can be accessed on a pay-as-you-go (hourly) basis or using monthly/daily parking permits.  The City wants to ensure there is always some space available for hourly parking to support downtown businesses while keeping the garage as full as possible with monthly permit holders to ensure efficient use; i.e., if too many parking permits are sold, then there will not be enough spaces for pay-as-you-go users.  In order to find the optimal number of permits sold, students will have access to a dataset generated by automated cameras operated by Hotspot Parking.  These have recorded a large sample of the license plates of cars using the parking garage over a 6+ month period, which can be cross-referenced with a database of permit holders and hence used to model the permit holder demand for parking spaces as a function of the time of day.

CFM (a JDI subsidiary) uses drones to inspect critical infrastructure in various industry sectors. Multiple flights of the same structures are flown over time to determine whether changes (corrosion, structural damage, etc.) have taken place. Currently, the process is manually-driven and does not offer automated change detection for the large and often complex structures we inspect. This process is cumbersome and offers little value to the end-user. An automated means of detecting change using background information from CFM’s inspection group and drone imagery may offer significant cost savings and value to our many customers.  The goal of this problem is to create an automated workflow where drone imagery from multiple campaigns is overlaid and compared for any visible change using relevant inspection criteria and analytics.  The preferable outcome is a user-friendly visual tool that displays any areas where change has been automatically detected between campaigns.  Visualization tools are currently available in the market place, however the change detection algorithm needs to be developed.

The New Brunswick Department of Natural Resources and Energy Development requires the development of a deer population estimation model. The Fish and Wildlife Branch currently uses a population dynamics model which accounts for annual population change (births and deaths) that provides information on trends but not abundance. A model is required that will produce reliable abundance estimates at a Wildlife Management Zone level and predict future populations under various harvest and climatic scenarios. Existing digital data that may be used for statistical analysis includes harvest data (sex/age), vehicle collision data (age/sex, female productivity), and summer/winter survival rates.