Category Archives: Academia

It should not surprise me that birth of my third child has launched me into a deep state of introspection.  Not only does a birth shift the foundation on which your life stands, it leaves you with little time or energy to do anything but reflect.  A newborn’s life stretches out before you as you hold their tiny sleeping form, and as you wonder how their life will be and wish them well it’s only natural to reflect on your own state and think about where you want to be.

When Jonas was born in 2008, it took me two weeks to register for the GRE exam and apply to a distance program to begin earning my MS degree.  When Emily arrived in 2010, I received an email from my advisor-to-be less than an hour after her birth with the funding letter to pursue my MS and Ph.D. full-time.  Less than a week before my son’s birth I came to the final decision that I wish to abandon the theoretical path my thesis research took me on, in order to go back to applied research that builds on my consulting experience.  On the surface this may seem like a small shift in the pattern of life when compared to the actions I took when the other two were born, but I think it may have more significance than it at first appears.

I’ve spent the past two years trying to wrap my head around driver behavior during car-following events.  I’ve been trying to understand how you as a driver react to the vehicle in front of you as their relative speed and relative distance to you changes.  In theory, a better understanding of this behavior can lead to more accurate traffic simulation software algorithms, which can allow researchers and practitioners to better predict how different potential roadway conditions will affect traffic, allowing for cost/benefit analysis before construction begins.  Even models that closely match real-world behaviors may result in simulations that bear no resemblance to actual field conditions, and the calibration of these models can be more of an art-form than a science.  I have been fortunate (or not) to have access to a massive database of information, where lots of data elements (including GSP location, speed, and the radar information) for one hundred vehicles, recorded every 0.1 seconds for an entire year.  The original study sought to gather accident data, and to my knowledge I’m the first to try and pull mobility information from it.  Unfortunately (for me), the reason no-one has applied this data to mobility before now is the incredible amount of processing time and effort in order to translate the data into a usable format for analysis.  Around the four-month mark in working on this full-time, my advisor recommended that the first paper might be on the complications encountered in data reduction.  By the time I finally had a dataset to work with, I was about a month away from the deadline to submit papers to our industry’s big conference.  In that time I read the preliminary papers for the four models I was supposed to be simulating, implemented the four in excel, and wrote two papers, one dealing with getting the data and one dealing with the results I had from the simulation.  In hind sight, I should have spent four months getting the data and three months modeling, instead of six and one.  The result of all this was a rejection letter for each paper.  In the three months between submitting the papers and receiving “review comments,” I performed additional data reduction work instead of trying to build further understanding of the models I was using.  At the time I didn’t even realize that I needed to have a better understanding of the models; as far as I was concerned these were established models that have been used in research regularly, and my contribution had everything to do with the new dataset and nothing to do with the models.  It wasn’t that I didn’t want to spend time reading more about them; I was spending every available minute on data reduction trying to make my dataset as significant as possible.  This misunderstanding on my part ended up causing a great deal of extra work for me in the month leading up to my thesis defense, and in the two weeks following it!

Fortunately, I look back at my crazy research year with amusement instead of aggravation, because I did have a paper accepted at the conference, a third paper submittal I did on my own time as a hobby, entirely unrelated to my thesis research.

There is a subcategory of intersection geometries that were called unconventional intersections, and are now being marketed as alternative intersections.  Some of these designs, like the roundabout, have been widely implemented, while others exist only on paper.  In addition to the roundabout, you may have driven through or heard about the Jughandle intersection, the Median U-turn intersection, the Diverging Diamond Interchange (now referred to as the Double Crossover Diamond Interchange), or even the displaced left-turn intersection (previously called the continuous flow intersection).  From an application point of view, these intersections are very difficult for practitioners to model using the basic software applications common to all traffic engineers, and they must instead be modeled using costly and time-intensive simulation software packages generally used either by researchers, or by a specialist in a very large consulting office.  Thus they are often overlooked entirely as options.

In the conceptual planning stage of an intersection or interchange problem, all of the potential alternatives are compared for functionality, often measured in terms of average delay per vehicle in seconds.  Preliminary design is conducted for the best performing designs, including a cost estimate of each alternative.  Examining the level of operations of an alternative along with its respective cost yields a decision on which design to pursue, with final design and construction drawings produced.  Sometimes additional considerations are made, such as the ability to maintain traffic flow during reconstruction for a particular design.  If an alternative intersection design is going to be considered as one of the alternatives for design, it usually takes a direct interest from a client (municipality, local, or state government official) to get it included, in large part because of the difficulty in including it at that conceptual analysis phase.

The Federal Highway Administration is supporting the expansion of alternative intersection designs, and in 2010 published a paperoutlining a simple tool to perform comparative analysis for these intersections based on the critical lane volume (honestly, you don’t need the details on it right now).  Effectively, the tool provides a comparative analysis between lots of alternative intersection designs after about five minutes of work and no processing time.  I’ve been thinking for the last two years that this sounded too good to be true, so my pet project was to put it to the test.  I chose one alternative design, the quadrant roadway design, and compared it against a conventional intersection with a bunch of different volume combinations.  I used the simplified tool, and the base software used by all traffic engineers, and then I also did high-fidelity simulation of the two alternative designs.  I wanted to know if the results (one intersection better than the other) provided by the simplified tool were consistent with the results provided by the other two methodologies; what I found was that they were not.

The obvious next questions to ask are: 1. does the simplified method work for some of the alternative designs, but not the one I checked, and 2. what other simplified methods of comparison might work for these designs?  Here’s the catch – this kind of research could be performed by a capable traffic engineer who had some time on their hands and an inclination to spend it doing research.  The car-following theory research is far more rigorous, with the best work usually being produced by doctorates in mathematics, electronics signal processing, or fluid dynamics.  If I could buckle down, really understand the various models, make new observations about those models based on my unique dataset, and potentially develop my own model, then I might increase my chances of faculty employment four years from now.  I worry that a research record as a revved up consultant may be a disservice to me in pursuing employment.  So here’s where we come full-circle to thinking about life with the perspective of holding a newborn in your arms.

I may not be cut out to do top-notch rigorous research in highly theoretical transportation issues.  More importantly, I don’t enjoy it.  When I first started my research, and when I worked on my hobby paper, I woke up excited to go to work in the morning and see what I could accomplish.  I’m not saying I wasn’t excited to go home at the end of the day, I LOVE spending time with my family, but I’m able to enjoy that time so much more after a productive day.  Working to pull everything together for my thesis this fall was like pulling teeth; I had to force myself to stay on task and I spent more time spinning my wheels / banging my head on my desk than doing anything else.  Why would I spend three more years trying to get this to work out, in order to qualify for jobs where I’d be doing this for the next forty years?!?  I certainly don’t want to return to consulting, where the world revolves around cost-effective ways of doing things and not “right” or “correct” ways of doing things, but I also don’t want to trap myself in a part of the research world that I don’t wish to visit, much less inhabit.

I think I’ve been fighting myself on this for the past 20 years, but what I really want out of life is to teach, to inspire, to foster a collaborative atmosphere, and to mentor.  Before I started my part-time MS program I knew all of the teacher certification programs within an hour’s drive and I’d read extensively on their websites.  My sense of responsibility to my family was always preventing me from taking the pay cut to leave engineering and teach high-school, and I believe rightly so.  My return to graduate school is effectively a nuclear option to merge my desire to maintain a higher salary with my desire to teach.  Some may wonder, why all this machinating about research if you just want to teach?  Transportation engineering is taught within the larger field of Civil Engineering, and in terms of occupation accounts for around 10% of all civil majors.  In order to achieve a tenure track faculty position in transportation engineering, I will need to work in a Civil department that’s large enough to include full-time staff in transportation (instead of adjuncts), which necessitates a larger university, which generally means research intensive.  I don’t want to go the route of an adjunct faculty, because they don’t make enough money, they don’t have a say in how a department operates, and they have minimal mentoring opportunities.  So I keep driving myself in this thought loop that I need rigorous research to obtain a position that meets my needs/wants, but I’m not enjoying what I’m doing and I know I wouldn’t enjoy it ad infinitum.

I am extremely fortunate that my advisor is supportive of my pursing either of the two research topics.  I finally got the nerve up to ask him about switching the week before my son was born, and his advice to me (he’s always very direct) was that it’d be better for me to do a great job on an applied research topic, than a mediocre job on a rigorous research topic.  So with that decided I just need to figure out how I’m going to get ahold of my own classroom instead of just doing guest lectures for the next three years!

In the meantime, I’ll go back to holding my sweet six-pound little boy, and maybe take a nap.