It is indeed a clever use of electro-mechanical devices to implement state logic.
However, I'd be concerned on a few fronts:
- voltage division between the two coils, as already mentioned
- momentary impedance of inductor d2 causing corresponding voltage/current sag in d1; if d1 de-energizes (however momentarily), we're back to stage ZERO.
- I wouldn't feel quite right about a switch that requires real power to maintain state - controlling a safety-critical device.
Picture a spontaneous lite's-out as a result of a negative transient...
... induced by, let's say ... panic braking ...
Even with an electronic solution (ie. with improved transient immunity) such a default-off/auto-reset switch may raise safety concerns.