I really like how you emphasize the "judicious introduction" of time delays. I have seen some folks use time delays as a quick fix without carefully evaluating their effect on the entire system, which can either be a great solution or it can introduce other errors in subtle ways.
I have been involved in developing, improving and consulting anything related with control system for the last 15 years in electronics component testing and automation system. Time delay is really important and at some cases it is the only mean to solve the problems.
When the control loop is some derivative of a PID loop, maintaining phase is very good advice.
When there are unavoidable delays present in the system, your suggestion of some kind of "wait and act" strategy is probably good advice. When you think about it, adding a delay in the control loop is just another way to "conserve phase".
Are they robots or androids? We're not exactly sure. Each talking, gesturing Geminoid looks exactly like a real individual, starting with their creator, professor Hiroshi Ishiguro of Osaka University in Japan.
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.