3 ms·
> It's not possible to engineer an autonomous system that never fails, but it is possible to engineer one in such a way that it never fails to detect that it ha
by aeorgnoieang 9y ago
> It's not possible to engineer an autonomous system that never fails, but it is possible to engineer one in such a way that it never fails to detect that it has failed.
That is an extremely surprising result. How is that possible? Are you really claiming that any control system can be engineered to detect that is has failed in any possible manner? What's an example of actual real-world system like that?
- lisper 9y ago> Are you really claiming that any control system can be engineered to detect that is has failed in any possible manner? No, of course not. But it is possible to reduce the probability of non-cognizant failure to arbitrarily low levels -- at the expense of cost and the possibility of having a system that is too conservative to do anything useful.