Welcome to the Research Roundup: a collection of highlights from Husker’s latest research and creative endeavors.
Pickin’ good vibes
In the United States, nearly one-fifth of commercially reared piglets die in the approximately three weeks after birth — the time that a litter nurses from its mother before being weaned from her milk. Farmers and researchers are working together to study how sows and their litters can be better monitored to reduce this mortality rate. Although video surveillance is effective, it requires adequate lighting and digital processing and storage capabilities that make large-scale implementation difficult. Wearable sensors, on the other hand, struggle with the challenges of durability, battery life, and other issues.
Nebraska’s Tami Brown-Brandl, Raj Sharma and Asya Macon, all from the Department of Biological Systems Engineering, worked with colleagues at the University of Michigan, Stanford University and Carnegie Mellon University to explore another option. The team placed sensors on the underside of several pens to test whether vibration could help identify a sow’s posture – lying vs. sitting/kneeling vs. standing – and the feeding activity of both the sow and her piglets.
Using machine learning, the vibration-sensitive system accurately classified a sow’s posture in the vast majority of cases: 95.5% sitting/kneeling, 99.9% lying and 100% standing. It also correctly classified feeding a sow 96% of the time and correctly matched feeding piglets 91.3% of the time.
The results suggest that vibration sensors could prove to be a reliable and practical way to detect the factors – piglet size and disease, injuries caused by sows – that tend to put piglets at risk.
let the blood flow
Peripheral artery disease results from a narrowing or plaque-related blockage of blood vessels that restricts blood flow to the extremities, usually the legs. Although the resulting lack of oxygen can cause periodic leg pain, peripheral arterial disease can develop without pain, delaying treatment and possibly leading to more serious long-term health problems. And diagnosing peripheral artery disease can be especially difficult in older adults, who sometimes have health problems with symptoms similar to those of PAD.
Unfortunately, the go-to diagnostic procedure for PAD is expensive and requires special training. Nebraska Engineering’s Fadi Alsaleem, Mohammad Ali Takallou, and Ali Hazem Al Ramini contributed to a recent study aimed at developing a simpler, less expensive screening technique. Led by the University of Nebraska at Omaha, the study applied machine learning to data collected from 270 adults — 227 of them PAD43 sans – who walked a path lined with force plates while wearing reflective markers caught by high-speed cameras.
By identifying specific biomechanical and pressure-related signatures, two types of machine learning models successfully discriminated between adults with and without PAD in about 89% of the cases. Even relying only on so-called ground reaction forces – the forces that are exerted on a foot when a foot applies force to the ground – one of the model types still achieved an accuracy of 87%.
Finally, applying the study’s proof-of-concept to wearable sensors, including those built into shoes, could help alert doctors to the onset of too PAD‘ said the team.
In late 2020, the university’s Bureau of Sociological Research surveyed a random sample of 2,775 Nebraska residents. The survey arrived amid the aftershocks of two seismic events: the beginning of the COVID-19 Pandemic and nationwide protests over the murder of George Floyd by a former Minneapolis police officer.
As a researcher who has studied the intersections between public health, crime, policing and politics, Lisa Kort-Butler of the Institute of Sociology decided to analyze these survey results. She was particularly interested in whether the general anxieties and insecurities that fueled the ongoing discord might affect attitudes toward public policy and authority.
Perceive COVID-19 as a serious threat, regardless of whether that threat was viewed as an economic or health issue, corresponded with concern about crime. And Nebraskans watching COVID-19 as a health threat were also more likely to avoid areas where they felt unsafe. But those who know someone who was infected COVID-19 actually less likely to avoid these areas – a seemingly paradoxical result that nonetheless reflected national surveys.
About 20% of respondents said their communities are underspending on law enforcement. According to the survey, Nebraskans who generally believed this also have:
- perceived COVID-19 as a threat to the economy rather than public health
- Expressed confidence in political leaders but distrust of health officials on matters COVID-19
- Believed that race relations in the United States were deteriorating
Overall, Kort-Butler said, the polling data supported the idea that the uncertainty and insecurity arising from times of upheaval can be channeled into support for criminal justice policies that promise a return to normalcy.