Learning when to stop

Over the last nine chapters, we’ve worked on designing a specific robot to perform a specific task. We designed Albert the robot to pick up toys in an unstructured environment, namely a family home. To do this, it needed to be able to recognize toys with a camera, pick up toys with a robot arm and hand, navigate inside a house, and deposit toys in a toy box. We also added interaction, teaching the robot to listen and react to commands. Finally, it received an artificial personality and simulated emotions.

So, the next question is, are we finished with designing, building, and testing our robot? Sometimes, the most difficult part of designing, building, and testing a robot is determining when it is finished. Quite often, I see that some little thing might be improved, or some detail added, or some feature enhanced. Oh, the robot needs a spotlight. It would be nice if it could remember what toys are in the toy box. What if it had two arms? And so on. You...