Using the firm’s Railcar Inspection Portal (RIP), customers can use the new AI detection models to target and detect specific defects and anomalies on railcars.
Most of the AI detection models offered by the company are used for the identification of railcar defects associated with inspection points that are needed by the Federal Rail Administration (FRA ) and other regulatory agencies.
The new models are suitable for the monitoring of brake beam bent, ladder stile condition, retainer valve handle position, side handhold condition, and ladder tread condition.
These models were successfully tested in customer-owned environments for several weeks besides undergoing lab testing and the firm’s strict quality assurance processes.
Duos CEO Chuck Ferry said: “Our detection models are core to extracting the most value from our RIPs for our customers, and these expanded applications will be useful in both our traditional business as well as our new efforts towards subscription offerings for a broader market opportunity.
“We’re looking forward to seeing the efficiency and performance of these algorithms, as well as the others that are in development, continue to improve as they are exposed to datasets in the field.”
The company is also planning to launch a further six AI detection models by the end of this year.
These models are expected to be used for the monitoring of brake systems and running gear.