LewsTheri's

I have recently updated my PhD website. I guess it is about time to do that as well. So the first thing I did was to change my photo. It used to be this:



Ugly, isn't it? It's now been changed to the following photo which I took myself:



And the equipment beside me, that's my precious experimental test bed. Took me in total about 8 months last year to complete it from scratch. The whole thing was custom designed, commissioned, built, debugged, tested and finished in the research school. The whole project was undertaken by myself, with advice from my supervisor and as well as help from the Engineering Department workshop manager for some of the parts. So i'm quite proud of the achievement.

An explanation of the various components of the test bed will certainly bore you guys so I won't go into that. The important point is that we now have the ability to conduct experiments that were never before possible or had been done by others in the research of Shape Memory Alloy actuators. Most of the papers in the literature are usually about position control and not many have delved into SMA force control systems. With our test bed, we can investigate constrained force control and even combined position and force control.

So far, our results are very convincing. We have achieved very accurate force control and fast response as well. A paper for ICRA next year has been submitted based on our simulation and experimental force control results so I am looking forward to this conference and trip to Rome, Italy.

Our final research milestone is to implement combined position and force control using an antagonistic pair of SMA actuators. The two SMA actuators are attached to a rotating pulley and the rotation of the pulley can be measured using an optical encoder/sensor. The idea is to have a position outer loop, to track position commands (rotation of the pulley), and a faster force inner loop. The reason is because we already have a very accurate SMA force control, so we can treat the actuators as ideal force actuators. This means that whatever force command we ask for, the control system is capable of getting the actuator to produce the required force.

Although a single position feedback loop is always easy to implement, and can track the position commands, which was demonstrated countless times in the literature, the introduction of an external load or disturbance to the plant will cause oscillations and result in tracking inaccuracy. The reason is that the load introduces extra dynamics which are difficult to compensate with a simple position control. With the combined position and force control, we hope to solve the oscillation problem. In fact, when I recently implemented such a control scheme, it worked. Oscillations were not observed. What we plan to do in the next few months is to tune up our control systems to produce faster and more accurate response and to obtain a model of the SMA as a design and simulation aid.

Hmm... I wasn't planning to explain what I did and why or how it was done that way but I guess I ended up getting worked out over it. Perhaps for the engineers reading this, I hope you may find my work interesting. This certainly is the exciting period in my PhD. After this, I will start writing up my thesis next year At the same time, I will start planning my career and searching for jobs.

I wonder if they'd offer scholarships if I were to do another PhD :)