Digital and Physical Replication of Railcar Lid Configurations for Machine Learning in an Automated Loading System

RAYHAWK has an autonomous solution for opening and closing railcar lids are actively working to expand their capabilities. Railcars are not standardized and all Railcar Top Objects (RTObjects) include lids, lid strappings, latches, latch seals, and walkways which make for expansive railcar configurations that need be supported in future solutions. The key challenge to expanding the solution is to rapidly support a multitude of additional railcar types needed to address more markets. This project will augment physical replicas of railcar tops with digital replicas and behavioral models, or digital twins, that can be used to create a Virtual Training Platform, VTP. The VTP can be used to more efficiently and comprehensively train the system, and the physical replicas will still be required for real world performance validation. This project lays the foundation for future autonomous loading solutions requiring expanded railcar configurations, in both digital and physical form, for machine learning, training, and performance validation.

Faculty Supervisor:

Leon Lipoth

Student:

Partner:

RAYHAWK

Discipline:

Engineering

Sector:

Manufacturing; Professional, scientific and technical services

University:

Saskatchewan Polytechnic

Program: