Monday, 25 March 2013

Into to Control - A Researcher Prespective

Before you consider taking on control as a major in your graduate studies, you should have the answer to the following questions:
  1. What is control all about?
  2. What are the sub-fields of control?
  3. Which topics are "hot" in control?
  4. Which are the topics that are more useful in control?
  5. Which are the topics that will more likely land you a job?
This post will give you my answers to the above questions.

So, what is control? My answer is simple - it is the study of how systems work, as well as, how to control it. The answer is simple but not the actual study of control. Control contains huge amount of mathematics. Thus, to be good in control, you need to have a solid foundation in math. I believe that the two important math subjects that applies to all sub-field of control is probably linear algebra and discrete math (topic on logic). A separate post will be written on math subjects that are related to control.

There is a huge amount of sub-fields in control. I will name a few important ones:
  • Linear System
  • Nonlinear System
  • Stochastic (Random) System and Control
  • Consensus Problem
  • Robust Control
  • Discrete Event System
  • Optimal Control

Typically, a system is either is linear or nonlinear. If it is nonlinear, studying the system becomes hard. However, when things are hard, that means there are more opportunities in that field. Hence, I would say the study of nonlinear system is extremely popular nowadays. It is also the most useful knowledge in control (theoretically) because most physical systems in the real world are nonlinear. Therefore, if you understand nonlinear system, you can design controllers for more variety of systems. No matter what other sub-fields you are working on, you will eventually see a problem related to what you are doing but formulated as a nonlinear system problem.

Some other "hot" topics includes robust control, consensus problem, multi-agent system and optimal control. I really cannot talk on these topics since I am not an expert in them. You can always try to find a survey paper on these topics to learn more.

Finally, if you are interesting in looking for an real world job, then stochastic control is probably a better field to study. Most of the time, it is hard for a grad student in control to find a well paid job because a lot of control people don't actually build things. Typically, we provide a control algorithm which is really just two or three equations. Done. It is not hard to imagine why employers are reluctant to employ control engineers. However, not all hope is gone, there is one field that I know is extremely hungry for people who are good at math - the FINANCIAL sector. The word finance is a synonym of randomness or stochastic process. Therefore, if your thesis is on stochastic control, you will have a good chance to get a job in the financial sector which has really good pay.

To conclude, control is a very nice academic field and it has a lot of useful applications. Unfortunately, there isn't a lot of demand for control engineers in the market right now (at least not the theoretical ones). To choose whether to step into the world of control, it ultimately depends on what you want out it. I will write another post on things to consider before going into control some other time.

No comments:

Post a Comment Blog Ping Site Google+