4 ms·
It seems stupid to use these unreliable "signals" at all, but even if they have to use them for "safety" they don't have to use them for operator evaluation. It
by jessaustin 9y ago
It seems stupid to use these unreliable "signals" at all, but even if they have to use them for "safety" they don't have to use them for operator evaluation. It would be real easy for every operator to carry a device that could prove e.g. she drove an appropriate speed even though her train tripped a poorly-maintained signal. I'm not foolish enough to imagine this would lead to the signals getting fixed, and from a political perspective it might be important to stipulate that it wouldn't, but that seems secondary to the horrible schedule problems that tentative driving has caused.
- gamblor956 9y agoWhat sort of magical device would accomplish this feat? It couldn't be GPS-based, as GPS signals wouldn't penetrate to the subway level. It couldn't be momentum-based, because that's not sufficiently accurate. It couldn't be connected to the train's computers because that would require major upgrades to every train in the system, which would take decades.
- nisa 9y agoEvery truck at least in the EU has such a mandated device. Basically it's a signed timestamp with current speed. They have a shedule, they know the driver. If they fear manipulation just include the device in the train and read the data every other week. Trains in the EU also have this.
- gamblor956 9y agoWe're talking about subways, which do not have access to GPS, and which is what EU trains use. Most of the trains in the MTA system predate the EU and a fair number of the countries that are members of the EU! The London system uses a different positional system based on location beacons built into the lines to locate trains (essentially, land-based GPS).
- jessaustin 9y agoI don't have an amazon link for you, but it doesn't seem that magical to me? Submarines have had inertial nav for decades. This is a simpler problem, because travel is restricted to rails and drift can be corrected as often as necessary. Every airbag in every automobile is equipped with a very accurate fairly inexpensive MEMS acceleration sensor, which would the only specialized component required for such a system. Hobbyist components like the ADXL345 are also cheap. It wouldn't surprise me if such a system already exists for e.g. tracking underground mining equipment, but if not it wouldn't be an impossible task for an organization with the budget of MTA. It's not as though the system would need to be incredibly precise; it's replacing one that at its best can only measure average speed over some huge distance.
- gamblor956 9y agoSubmarines also need to surface periodically to verify their navigational position as dead reckoning is frequently off by a margin. Airbags are equipped with acceleration sensors that have a very simple test: is the acceleration force greater than X? If so, deploy! The point of the system is to have a system that is accurate enough to replace the current system, which neither of your solutions achieves.
- jessaustin 9y agoYeah that's what's meant by "drift". Submarines correct every few thousand nautical miles. The hypothetical device could do so at every station. The current system is not accurate at all, and even if it were is quite imprecise by design. But you've convinced me. A much better solution would just be a camera pointed at the speedometer.
- jessaustin 9y agoNew idea: mount a camera so that the train's speedometer is in view.
- bluGill 9y agoIt isn't magical. A bicycle computer style senor to measure speed, record that once per second to a raspberry pi ish computer. When the e-brake triggers (simple momentum sensor is enough) log that too. From the above I can be close enough to how fast you were going. If you should be going 15mph and are doing 16 and slowing down - well that is in the margin of error and you are fine. If you should be going 15mph and are going 25 I'll know that. The key to the above is margin of error. I don't need centimeter level accuracy, just enough that I can reconstruct if you were close to right or not.
- gamblor956 9y agoA bicycle computer uses the radius of the wheel, multiplied by rotations, to determine distance traveled and uses that to back out to speed. More expensive bicycle computers use GPS to verify their measurements. This method is accurate enough for its use; it's not accurate enough to be the primary safety mechanism for trains carrying hundreds of people. For example, the difference between 15mph and 16mph on a bike is irrelevant because it doesn't trigger a response. But for an emergency breaking system, going above the threshold triggers a physical response (i.e., the emergency brakes). On top of that, you're asking the MTA to buy hundreds of new devices and maintain those on top of the existing devices they don't have the capacity to maintain...
- bluGill 9y agoThere are (or have been - with GPS so cheap these days they might not make that style anymore?) some that use some other senor. I'm not sure if it was ultrasonic, radar, or something else, but they have existed. You misunderstand what I'm proposing. I'm not suggesting this system replace anything. I'm suggested a much cheaper system that just collects data to monitor the performance of the existing systems. The comments are full of suggestions that the existing e-brake system is causing drivers to go too slow because if the e-brake triggers the driver is in trouble even if the driver was going a safe speed and the e-brake was wrong. (It is a good thing that the existing e-brake fails in this way!) My proposal was a data source so we could know if the driver was really in the wrong, or the automated system. The proper response for the driver being in the wrong is very different from the e-brake being in the wrong, and that data seems to be missing. I'm asking MTA to buy and maintain these devices on top of everything else because this is data they need anyway: a second source of data for reconstruction after the fact. Airplanes have a "black box" data recorder for much the same reason: nobody can predict everything that might go wrong. However if we have the right data after the fact we can at least make sure that particular failure never happens again. Airplanes a extremely safe because that data is analysed and acted on. MTA doesn't seem to have the data (though I will admit it is entirely possible they have it and are unwilling to look at it for whatever reason)