Multimodal Visibility Project

Transport Canada’s Transportation and Economic Analysis (TEA) Directorate is implementing a digital data strategy in support of the multi-year Trade and Transportation Information System (TTIS) and the recently announced Supply Chain Digitalization initiative. One of the key objectives of these initiatives is to address data gaps and provide support for the routine monitoring of transport network performance through the innovative use of non-traditional sources of transport information. The widespread use of digital sensors and improved methods for fusing different data sets together promises to revolutionize the way in which Transport Canada and other agencies collect and analyze information on multi-modal activity and performance. TC wishes to access additional sources of data on freight activity, especially as it relates to multi-mode interactions, as a complement to its traditional data holdings.

One of the current unfilled data gaps is coverage of port-rail interactions between smaller ports and short line railways through the fusion of sensor data and available shipping documents. Some of the broad functional needs that would profit from access to more timely and complete multimode freight flow data are expected to be the following: • Monitoring and measuring railway network performance in and out of marine ports, especially smaller ports, and by short-line railways through the installation of in situ non-intrusive detection technologies. Collection of port-rail network performance data including dwell and transit times by different types of freight supports the department’s supply chain digital strategy identifying potential bottlenecks and other vulnerabilities and permits greater optimisation of scarce network capacity. • Better knowledge of the interaction between marine vessel activity and short-line rail transport through the fusion of available shipping documents collected by supply chain partners. More complete multimode freight flow tracing from shipper to final destination improves knowledge of freight operations by mode and by commodity, improving visibility and optimising system-wide performance. The requirement includes the following elements: A. Train capture platform 1) Using proprietary technology, install two (2) permanent train detection camera stations on port property at the main railway approaches to the port; 2) Collect train consist images continuously (24 hours a day) for a period of 12 months; 3) Using proprietary machine learning technology, develop a Machine Learning (ML) model for converting the imagery into a database of train counts and railcar characteristics. 4) The train database will cover the following elements: i. Count of trains in each direction passing by the site ii. Date/Time of observation (Local time) iii. Number of locomotives and cars in the train consist iv. The type of train (at minimum Manifest, Intermodal, Coal Unit, Grain Unit, Potash Unit, Petroleum Unit, Automotive, Other) v. The type of cars in the consist (at minimum the ML model should identify Box, Tank, Hopper, Gondola, Flat, Intermodal Stack, Centre beam, Auto) vi. Number and size of shipping containers in the consist (minimum 20, 40, 53 foot) vii. Where visible, the serial number of the cars and locomotives (e.g. GATX 165789) viii. Dangerous goods placard information (Placard class and UN number) ix. Train speed at site 5) Feed the derived train activity information and related performance metrics into a custom operational dashboard. 6) Minimum update frequency of train data: Daily

B. Architecture for fusing shipping documents and related sources Collect examples of available shipping documents and related supply chain information from existing railway and port records: • CBSA - A6/A6A shipping declarations and manifests • Railway-based Waybill documents • Associated train consist/manifest data collected by the connecting railway • Available GPS sources collected by the port or railway • RailState Camera platform data on sites in proximity to the port [TC-supplied] • Available shipping/manifest documents • AIS traces of marine vessels departing and arriving – historical data

C. Develop an architecture for merging/fusing these documents into a comprehensive multi-mode picture of port-rail activity in and out of the Port. This should cover the following: i. Transport Modal activity (Vessel/Train counts and characteristics) ii. Time-based modal performance metrics (Dwell and Transit times of train and vessel flows) iii. Commodity flows carried by Rail and Marine modes (Commodity type and Origin-Destination) iv. Develop a workflow for feeding fused supply chain data into the operational dashboard.