Hacker News new | past | comments | ask | show | jobs | submit login

How is it possible for thrusters to fire without a) human oversight and b) nobody knowing?



It's an autopilot with no communication connection. What would be the alternative, drifting uncontrolled whenever it's outside of communication range?


It's not out of 'communication range' with personnel aboard the station.

An ultra basic integration of alarms and communications would be the minimum requirement to allow personnel to monitor and modify mission critical systems.


It was prior to docking. It was a separate entity which had to fly itself without continuous communication to the station in order to be integrated. The entire problem was that the module did not recognize that it was docked with the station and was trying to maneuver itself. Why would you put in a "tell the station to which you are attached that you are firing your thrusters which should never be fired while attached to the station" signal?


It's not a crazy suggestion to have ability for an emergency stop or docking success signal sent from the station to the module.


The problem was that the "docking success signal" was not properly recieved.

The very presence of an emergency stop system should be a signal that the thrusters should not activate, if you are ever in a situation where you would desire to send an emergency stop signal, it is because that emergency stop system has already failed.


"The problem was that the "docking success signal" was not properly recieved."

Then it's not docked and docking procedures should be observed by all personnel as a serious ongoing operational concern, right?

So, if there is integration, which there hopefully would be, that there wasn't proper integration would imply something bad.

Otherwise, while the crew obviously believed it to be docked, why wasn't the auto-correcting features etc. on the Russian module disabled while it's docked?

Again, as part of the docking procedure?


But they thought it was properly received.

This was a computer bug. Function updateDockingState() executed and returned exactly what it was supposed to, but it never properly updated the value of variable DOCKED, so function checkIfDocked() later returned False when it should have returned true. The exact nature of the bug and how it went unnoticed is yet to be determined, but the issue is not with the theory of operation. No integration is going to fix that. Are you proposing that a manual review of every line of code should be part of the docking procedure?

It is amazingly presumptuous to assume you know how to make a safe spacecraft better than a literal space agency.


Docking brackets can carry electrical contacts and SPI with redundant conductors would more then cover this usecase.


That system exists, the problem is that's the thing that failed here. The module didn't recognize it was attached to the station explicitly because there was an issue with those connections.


"Computers never lie, kid!"


I tell you how it is possible: it is both a bug, and a design issue.

To say anything deeper about the whys one would need that investigation.


Simple: the module was out of comm reach at the moment.


That's not 'simple'. The module should be integrated with the other modules such that any control terminal on the station should highlight that information.

Moreover, one would imagine that applying thrust should generally require some oversight and that it should not be automated unless there's an emergency in which case there would be appropriate alarms.


The module will be integrated over the next few months. It's not as simple as just docking the module and expecting all the wiring and software config to kick in automatically.


It could be thou.


a) software

b) outer space




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: