Skip to main content

You are not logged in. Your edit will be placed in a queue until it is peer reviewed.

We welcome edits that make the post easier to understand and more valuable for readers. Because community members review edits, please try to make the post substantially better than how you found it, for example, by fixing grammar or adding additional resources and hyperlinks.

12
  • $\begingroup$ "we consume power even when we are sitting down" <- robots do this to. Yes, a robot can be turned off without dying which is nice, but a robot that is "awake" and idle consumes a lot more power than a human that is awake and idle. $\endgroup$
    – Nosajimiki
    Commented Jan 25 at 22:24
  • $\begingroup$ @Nosajimiki - I gotta disagree - a computer that is on Standby mode uses 15 watts, a Human sleeping around 100 watts. A Human sleeping is about the same as an average laptop in use. $\endgroup$ Commented Jan 25 at 22:52
  • $\begingroup$ A laptop uses 15 watts, yes, but not a robot. A robot in standby mode is still actively running an AI to process sensory input and make decisions as to if it should stop being on standby and start doing something. This uses a LOT more processing power than an idle laptop because of the nature of neural-networking software. It also takes power to hold a robot in a standing position (assuming humanoid features); so, just standing still so that you don't fall over consumes power. Boston Dynamics robots for example idle at about 1kw to just stand and observe thier environment. $\endgroup$
    – Nosajimiki
    Commented Jan 26 at 16:37
  • $\begingroup$ @Nosajimiki the boston dynamics robot is not in standby when using 1kW, it is active, but unmoving. Real standby would get the humanoid robot into some metastable position, ideally lying down, but mayhaps just locking knees and setting the head onto the ground, and then power down the servos. Then the processing hardware goes into standby, meaning keeping the RAM powered, and possibly a watchdog routine on some subprocessor going online every 100ms for 1ms to check for some wakeup signal. 15W is actually rather wasteful for that. Still 24h*1000(error) is not 'decades', it's three years. $\endgroup$
    – bukwyrm
    Commented Feb 1 at 21:39
  • 1
    $\begingroup$ @Nosajimiki this comment thread is already getting a bit longish. Again, we might have differing definitions of standby. But i, also, have robotics experience. And in the state that i call standby (no posture/position is held, robot will come back online without boot on button press) the power draw for multi-kw robots is in the single watt digits. Which makes sense, because nothing apart from some memory keep-alive and some temperature watchdogs is actually getting power. $\endgroup$
    – bukwyrm
    Commented Feb 7 at 20:19