Tuesday, April 20, 2010

For those who track 3D

Folks in the field of animal locomotion would know how kinematics data are usually obtained. But allow me to summary the general procedure in a few sentences. To track anything in 3D, at least two camera views have to be available at all time. After space calibration, one can calculate the 3D configurations of the objects in the analysis software of his/her choice. Ideally, video tracking can use any inherent features of the subject. However, to facilitate automatic tracking, high contrast makers are often attached to the subject. Infrared markers offer a way to highlight the features of interest without compromising the lighting for the normal video acquisition. After I created two families of soft-bodied robots, I was challenged by the need of quantitative data. These kinematics data are critical for any mechanical analysis on the robot locomotion.I set up our VICON 3D system to track my robot kinematics at Tufts Advanced Technology Laboratory. VICON is a company that makes 3D tracking systems for research in locomotion and animation industry. It employed several near infrared high speed cameras which would detect the IR signals coming off the retro-reflective markers attached to the subject. Unfortunately, retro-reflective marking is really not the way to track small animals such as insects or robots of the same scale. After going through many types of IR florescent chemicals, I finally decided to go with semi-conductor IR emitters (or infrared LEDs).
This works out really great for my application because I do not need to worry about IR light flooding or bad camera focus. These surface mount IR emitters produce point-source lighting smaller than 1mm. The IR cameras pick them up like many distant stars. In fact, a little out of focus actually increase the pixel numbers from which the centroid positions are derived.
Data are coming alright, but my data crunching techniques are still too slow for the rate by which these high speed cameras acquire data. I better work on that!

Friday, April 2, 2010

Gait transition and embeded AI

A couple of years back, I was deeply impressed by probably the most well-known bio-inspired robot which demonstrated the effects of central pattern generator on gait transitions. This is the EPFL salamander robot with coupled non-linear oscillators. In this research amphibious robot, smooth gait transitions were accomplished by tuning the gain of oscillators coupling.

As I started working on soft-bodied robots, I discovered that many non-linear characteristics of the soft materials and actuators can be exploited to engineer behaviors. So I took a completely different approach to robot control. Instead of programming complex behaviors on a micro-processors, I "tuned" the body and actuators so they create desirable behaviors when I switch on a behavioral circuit. Amazingly, when the motor variations and body properties reach a certain domain, the robot was able to achieve gait transitions with a simple scaling of motor-pattern. This is a very intriguing demonstration because it provokes a rather radical inquiry: how much logic/intelligence can we embed in a piece of material? To what extent can we use morphing morphologies to perform computation (or thinking if you will)?

Thursday, April 1, 2010

Multi-threading...

A lot has happened in the past few weeks. Besides my secondary injury during my recovery of my bone fracture, everything else seems to progress in a positive direction.

First of all, my paper on caterpillar ground reaction forces was finally printed. It's been really over-due for a year now. Most data were collected by Christmas 2008, and I actually presented the major finding at the SICB 2009 January. I felt pretty bad about this delay but the robotics project last year really took my life from March through October. To summarize the findings in a few sentences: large caterpillars such as Manduca sexta load their bodies in constant tension when they are attached to a substrate. Locomotion was achieved by progressing the body tension/deformation forward. Biomechanically speaking, these critters use the substrate as their external skeletons. We call this strategy: environmental skeleton. For more details on this radical view of soft-bodied animal body control, check out the April 1st issue of the Journal of Experimental Biology. If you would like a PDF copy of my paper, simply e-mail me at huai-ti.lin@tufts.edu and I will gladly send you one.

Besides my old new paper, I've been planning a field trip to Costa Rica for this May and June. Last spring at the SICB conference, I bought a few books about caterpillars. Among them, I was really impressed by a couple of books regarding tropical caterpillar diversity. So I contacted the authors Dr. Daniel Janzen et al and was struck by the idea of visiting the home of caterpillars in the wild. Lab animals are always somewhat unnatural. This idea was incubated in the back of my mind for many months until I finally formulated it into a more concrete field study project. My mentor Dr. Barry Trimmer was very supportive of the idea and quickly decided to make it happen. In any case, we have now arranged a 17 days field work at a conservation in Santa Rosa, collaborating with Dr. Janzen's team from UPenn.

Finally, to continue the imaging theme from last time, let me share a few images from our histology for Manduca caterpillars. Working together with my great undergraduate lab-mate Dan, we've been able to produce very clean cross-sections of caterpillar abdomens.Through some imaging techniques, we can enhance the cuticular folds.Or we can also highlight the muscles! So awesome... the biology I mean (but we're not bad either)