What could be simpler than making a train go down a track? No steering required, just a throttle and brakes. Right?
Ever see the control booth of a locomotive? One look and you realize driving a train isn’t so simple after all. Old steam train cabs look like the work of a madman plumber with too many supplies. One false move and you’d either stop the train dead, or blow it up. Today at the Union Pacific Railroad, you must undergo 14 weeks of training before they let you touch even a track switch. New Jersey Transit puts would-be engineers through 20 months of training.
How much more complex it is to steer a Navy ship? It takes a lot longer than 20 months before they let you work in the bridge. Yet in last year’s fatal South China Sea crashes, the Navy blamed poor seamanship. Investigators cited “fundamental failures in the crew’s ability to operate their ships effectively.”
Now the Navy is also looking at the design and integration of its ships’ bridges. As Jared Serbu reports, Navy officials are talking about how the bridges have become “Frankensteins.” So much new gear has been crammed in, the crews can lose track of what’s going on. The average bridge looks like a data center.
It takes several sailors to properly man a bridge. Each person has to be trained on more than just the non-trivial basics of seamanship. They’ve got to know the particulars of that ship. No two Navy ships are identical, like a pair of Bayliners.
On the fateful morning aboard McCain, the commanding officer noticed something. The helmsman was struggling to control both the rudder and the throttle. But the helmsman and the second guy brought in to handle the throttles weren’t all that familiar with McCain’s systems.
It gets worse.
What nobody realized was that control had somehow shifted to a second station altogether. So the sailors fiddled with the controls in vain. The ship seemed to have a mind of its own. McCain steamed on for 3 minutes, out of control in heavily congested waters.
Eventually, the crew realized what was going on. Too late, though. The ramming killed 10 sailors in a berth below the water line. Their 15-foot-wide space was smushed to 5-feet wide. One who escaped used his “blindfolded egress” training, gulping for air and feeling his way around chaos as the water reached the ceiling where the hatch was.
If it moves, it’s gotten more sophisticated and complex over the years. It also has more automation. But the human interaction systems don’t seem to have caught up, at least not in the Navy’s case. I think an important question is this: What could the Navy do to abstract the differences in its ships’ control systems? Then build interfaces such that sailors would have an easier time moving from one to the next. That would reduce training requirements and the likelihood of more crashes.
Faulty leadership, seamanship and training ultimately caused these accidents. You can’t get around that. But better technology integration might have helped avoid disaster. And it would help in the future. The Navy generated a list of recommendations for itself to fix the bridges. Now it’s getting around to that list.
Navy investigators stated: “Technology intended to simplify navigation, improve situational awareness for smaller watchteams, and increase flexibility for ship control is complex, and in some cases, more difficult to operate.” This for 9,000 ton ships that cost almost $2 billion apiece. In the age of artificial intelligence and expert systems, that’s nuts.