Great question. I think there are recent advances in all the sectors you mention: here are some of examples of the advances, and in all cases the research question is "does it work" or what is the "cost benefit analysis" showing:
AIR -- careful behavioral analysis and rigorous adherence to protocols that make no exceptions (El Al, Tel Aviv).
WATER -- research on employee ID (by bio metrics) working at ports.
RAIL -- metal detectors in stations. Use of smart ID and CCTV systems to spot known or suspected individuals, (see examples now in China.) Automated comparison and matching to facial recognition data bases.
FREIGHT/ROAD -- data base analysis to link the price being paid for the shipment compared to the ultimate value of the item at the destination. This might help spot anomalous shipments, but would such a system be easy to defeat?
To these topics I would add an interest in trying to sort customers by trust level. TSA has a pre-check option for those who are deemed sufficiently trustworthy to by pass some aspects of security. Has there been any error?
The pre-check option is practical, but it does not cover profiling and background check. With the current mobility of people with travel documents, it may be quite challenging to do proper profiling. Apart from Human factor, there are also a technology and infrastructure factors that may to some extend impact the security and safety provision.