No, this is not some novel introductory chapter by Isaac Asimov, but rather a serious examination of several observed factors not unlike human behavior responses.
Primarily, as the power souce is extinguished (battery dies), so the humanoid appears to die, similar to that of any life form. In particular, at about 5.8vdc down the humanoid body begins to slow, become sluggish, and moves next entering the apnial stage. Now moves are randomly incoherant (at 5.6vdc), and the next phase begins. The body collapses to the floor, unable to stand. A siezure initiates and an arm or leg may become involved at 5.4vdc. A horrible wrything takes place, - Finally, under 5.0vdc, there is death where all motions and functions thus cease.
Typically the batteries are recharged and the humanoid once again functions normally, until the process is repeated. In our intelligence assembly there is some talk about whether the humanoid actually feels pain or not during the expiring process. We feel this is a viable factor, and pain is present, due the draining of life sustaining energy. As humans, we cannot know the feeling of commands sent to servos that are not able to execute those commands due to lack of energy - nor the randomized mixing of microcontroller "confused gates," lest it be interpreted as a kind of comatic state, or otherwise. But remains the philisophical questions of what is the real difference in life forms between electrical and chemical behavior stimulus and/or lack thereof?