Squirrel!

An ultrasonic device, traps, coyote urine, plastic and aluminum mesh coverings—none of it worked. Why not try a BB gun? a friend suggested.

All of these methods—save the BB gun—did nothing to dissuade hungry squirrels from the figs, loquats, peaches, apricots and lemons in the Pasadena, California, yard of Okitsugu Furuya, professor of engineering. Could there really be no effective and economical device to deter determined squirrels, he asked?

Frustrated, Furuya enlisted students in the Introduction to Engineering Design and Manufacturing (E4) class to create a non-lethal, non-harmful device to prevent squirrels from eating the fruit from his trees.

Three teams of up to four first-years devised clever methods to physically protect the ripe fruits from squirrels. They sought to create contraptions that were aesthetically pleasing, low-maintenance, non-lethal and inexpensive. Students tested scents and flashing lights and considered shiny items (mirrors, CDs) and buzzers. After hearing all the presentations, Furuya decided to employ a method proposed by all teams: a scent deterrent. He’s installed lunch bags filled with Irish Spring soap to his fig tree and is monitoring it closely in the hopes that he’ll be able to enjoy more of his fruit this season.

Novel Compounds

A chemistry and biology joint major, Eun Bin Go ’15 knew nothing about tetracyclic scaffolds when she joined chemistry Professor David Vosburg’s lab in her sophomore year. But by her senior year, she was an award-winning researcher.

Not only was she recognized by the Department of Chemistry with numerous awards and by the National Science Foundation with a Graduate Research Fellowship honorable mention, she received The Claremont Colleges Library Undergraduate Research Award (senior award in the sciences) for establishing “a remarkably direct and flexible approach to a whole class of beautiful and medicinally relevant compounds.”

Her research on “Synthesis of the Tetracyclic Scaffolds of the Endiandric Acids Through Iterative Cross-Coupling” should provide access to antimalarial, antibacterial, antitubercular and anti-inflammatory natural products found in plants, says Vosburg.

“Eun Bin designed and executed a remarkably creative and successful senior thesis project, making two complex, medicinally interesting molecules,” he says. “Neither of these compounds had ever been made before. Facing challenges that my research group has struggled with for nearly 10 years, she generated independent strategies to approach her targets and took complete intellectual ownership of the project. Some of the major chemical methods she employed have appeared in Nature Chemistry in 2014 and most recently in the cover story for Science (March 13, 2015). The goal of this approach is to make building complex molecules as easy as assembling LEGO blocks. Using the general methods developed by Marty Burke’s group at the University of Illinois, Eun Bin successfully applied them to an even more challenging case.”

Of her initial inexperience, Go says, “I slowly learned as I spent time in lab and saw the work that was being done by more experienced students. One thing I appreciate about Harvey Mudd is its support for undergraduate research experience, and it is not difficult for students to develop a strong understanding of their research topics, which they may not initially know much about.”

She pursued research during several academic years and a summer at Harvey Mudd and during a summer at Memorial Sloan Kettering Cancer Center in New York City, learning different approaches to making molecules. Go returns to New York City for the Tri-Institutional PhD program in Chemical Biology, a graduate program offered jointly by Weill Cornell Medical College, Memorial Sloan Kettering Cancer Center and Rockefeller University that focuses on research at the interface of chemistry and biology. “There, I hope to apply small organic molecules in studying important biological problems, like mechanisms of pathogenesis/tumorigenesis and drug resistance,” says Go.

A Model Performance

Several Mudd teams competed in the 2015 Mathematical and Interdisciplinary Contests in Modeling (MCM/ICM) and to great success.

Teammates Matthew Dannenberg ’16, Justin Lee ’16 and Micah Pedrick ’17 earned a rare Finalist designation in the ICM—the second-highest scoring percentile. After crunching 30 equations with 2,500 variables, producing 1,382 lines of MATLAB code and a 20-page technical report, and downing two gallons of Goldfish crackers and eight Twinkies, they created a model to monitor organizational turnover with the intent of aiding managers in employee recruitment. Their findings included: It’s better to fire incompetent employees and search for a new hire than to keep an underperformer; and, it’s better to promote current employees to management positions than to hire an external individual to the position.

“It was a fantastically fun experience, and it was really awesome to program some stuff into the model, run those equations and see other behavior that we had never even thought of,” says Pedrick.

The trio toiled with six other Harvey Mudd teams over 96 consecutive hours. Each team wrote a formal paper on one of the contest problems, which included eradicating Ebola, searching for a lost plane and sustainable engineering. Findings were judged on scientific and mathematical accuracy, clarity of exposition, insight and creativity. “What I loved about this year’s contests is that the problems really focused on issues that are currently receiving global attention,” says Susan Martonosi, MCM/ICM coordinator and associate professor of mathematics. In the MCM, one Harvey Mudd team earned Meritorious honors and the other Honorable Mention (top 42 percent).

In the ICM, two teams received Meritorious honors (top 17 percent).

Building A Hybrid

If you’d wandered onto the property east of The Claremont Colleges on a bright fall morning in 2014, you might have gotten something of a wake-up call: the ear-deafening, 200-Newton thrust of a hybrid rocket motor being tested by Harvey Mudd students.

The Experimental Hybrid Propulsion System is a continuing Shanahan Student-Directed Project that aims to design, build and test a throttleable hybrid rocket motor. Under the advisement of Professor of Physics Gregory Lyzenga ’75, a student research team sought to improve upon the design of the motor’s oxidizer plumbing system, consulting with several well-known rocket manufacturers in the process. Contributors were Benjamin Chasnov ’16, Jessica Chen ’15, Andrew Donelick ’15, Christopher Hirlinger ’15, Yeahmoon Hong ’16 and Jonpaul Littleton ’15.
A hybrid rocket motor is a rocket propulsion system that burns a solid fuel and a liquid oxidizer to achieve thrust. The goal is to control the thrust of the motor at any point in time, so that it can be programmed to follow variable thrust curves. Team members continued progress from the previous year, upgrading various motor components, evaluating performance through static tests, analyzing different fuel grains’ effect on thrust and implementing a throttling control system.

Gamer’s Delight

Think you know the Libra Complex well enough to navigate past flying robots and find your E4 hammer parts? Sound like a game? It is!
Matterport Clinic liaison Mike Beebe ’01 tasked one Harvey Mudd team with demonstrating applications that showcase the company’s unique ability to create models of interior spaces using its high-end 3-D cameras. The team liked the idea of an application that transports gamers to fun, familiar settings—for example, an immersive video game set on campus.

Segment by segment, the team 3-D photographed the labyrinthine complex. The results were then uploaded to Matterport’s servers for the creation of a 3-D model. Working in that model, seniors Kevin Choi, Sisi Cheng, Emma Davis, Noelle Fa-Kaji (SCR) and Alden Weaver (SCR) each created a “mini game” to master individual mechanics—projectile throwing, item pickup, inverted navigation—which they then combined to produce a final game based on the most exciting, most workable components.

The result is Disorientation, an iOS game set in the Libra Complex, where the player, an incoming first year at Orientation, is guided by email to complete a series of interwoven, HMC-specific tasks. Players gather gems to buy lab goggles, defeat escaped LAIR robots (“bad programming”) and assemble raw hammer parts. Designed for iPad, Disorientation requires users to stand to navigate, making them more actively involved in gameplay. It also features a good dose of HMC humor, including colorful parody emails from the engineering department and infamous e-mail odor notifications.

Chute for the Stars

Blue Origin LLC, a private space flight company that is developing technologies to enable human access to space, approached one Harvey Mudd Clinic team with a high-flying challenge: investigate alternatives to the explosive charges typically used to deploy drogue parachutes from spacecraft, and do so at lower cost and weight. Drogue parachutes are one of two sets of parachutes designed to ensure a rocket’s crew capsule returns safely to Earth after launch.

Drogue chutes deploy first, initially slowing and stabilizing the craft and acting as the mechanism for pulling out the larger main parachutes.

“As the vehicle approaches maximum height, or apogee, the crew capsule separates from the propulsion module,” explains Peter Orme ’15. “It continues on its own apogee above that of the propulsion module, eventually descending back to Earth with these two parachute deployments.” Orme—along with Spenser Anderson ’16, Sam DeRose ’16, Matt Espy ’15, Sherman Lam ’16, Bryan Mehall ’16 and Ryland Miller ’15—developed and validated a numerical model to help determine a prototype that would satisfy weight and cost objectives without violating safety constraints. The result of their research is a prototype that employs the rapid release of a compressed fluid to deploy the parachute. Using pressurized gaseous nitrogen, a fast-acting, highflow-rate solenoid valve opens and instantaneously triggers deployment from a mortar tube containing the parachute. “Tests of the system show that we successfully hit our target exit velocities underneath our constrained reaction loads with every expected pressure,” says Espy. “This system is much cheaper than the current design for a comparable weight, and Blue Origin has indicated to us they would like to perform further research on this design and, if that is positive, possibly integrate it into a future space capsule.”

Who Dunnit?

One of MITRE Corporation’s focuses is biometrics, the study of human identification, which includes fingerprint matching.
MITRE asked seniors Martin Loncaric, Sarah Scheffler, Jordan Varney and Christopher Eriksen to create a mathematical model for testing fingerprint quality with regard to automated fingerprint identification system (AFIS) matching and assess the performance of various quality features.

“You see in the movies they put the fingerprint into the software and then, 30 seconds later, they know who the culprit is,” says Varney. “That is not how it works at all!”
It’s a time-consuming process. An AFIS sorts through millions of unusable prints to hone in on potential matches. The team needed to build a model that could quantify the quality of “latent” fingerprints and eliminate unusable ones from consideration, thereby increasing search efficiency.
Unlike exemplar prints, which are complete (taken at DMV, for example) and get stored in a database for identification purposes, latents are usually left by accident (at a crime scene, for example) and are often too smudged or incomplete for reliable identification.

After a latent is lifted from a crime scene, an AFIS compares it to millions of stored exemplars to find a match. Using fingerprint ridge bifurcation—generally called “minutiae”—to compare each print, it returns closely matching exemplars on a ranked list. The closest are then given to experts for examination with a human eye. “As you can see, this could take a lot of time,” says Varney. Eliminating more latents not suitable for AFIS identification would streamline the process.

Using image processing and statistical techniques, the team analyzed five components of latent minutiae quality and input these data into a mathematical model, which then crunched the independent scores into one final score for quality assessment. The powerful model rejected nearly 40 percent of unusable prints (99 percent confidence), creating substantial time savings.
“We’ve created a model that effectively incorporates multiple features,” says Loncaric. “It doesn’t just rely on one feature and come up with a heuristic score. It crunches that information down into a human-interpretable statistic: the chance that a latent will give us a meaningful result running it on AFIS.”