Question: how does the robots memory work. like does it create a memory of the instructions you gave it if the instruction were to last for a month or does it just keep doing the same motion (depending on the instructions) for moths on end.

Giuseppe Cotugno answered on 20 Jan 2020:

Good question. At present, most of the robots are using a desktop computer to work. As such, all the information (both the algorithm and the data) is stored on an hard drive and it is used pretty much as any other computer program.

There is a part of it which contains the algorithms which is run like a program on Linux (another operative system like Windows) and does the same thing. Then there is the data. Part of it is data which is always refreshed at every restart (i.e. how many objects the robot sees at the moment). Another part, instead, used to be volatile (e.g. changed all the time) when the robot was developed but when the robot is used this data never changes or changes only once at set up. For example, the way a robot shall move its arm to collect an object might need to be fine tuned if the robotic arm is placed in a corner, while if it is placed in open space, it can be used straight away.

A minority or robots do not require a computer, but have ad-hoc electronics designed to help the robot doing something, pretty much as it happens for the car electronics. In those robots the part that changes for every execution / restart is less.

In general, you rarely consider the interval of months, but rather the last time the robot was restarted.