Brian O'Mullane

Gathering data is of little use if you cannot derive value from it. Increasingly, data is being generated in every part of the dairy and agri-tech sectors, including data regarding feed, livestock, yields and even logistics. As the types and volume of data grow, there is a real need to become more efficient and to recognise that data-driven decisions are the most effective way to optimise your business. We are now in the Information Age, and future growth and success will only be guaranteed to those who exploit data (both their own and others’) in the best possible ways.

Supporting positive emotional wellbeing is a critical challenge in an ageing population. Aware homes have significant potential to enhance the lives of older adults, extending the period of healthy ageing, through monitoring their health and wellbeing, detecting decline and applying interventions to arrest this decline. Our research aims to create a new method of preventative care based on early prediction of changes in emotional state and, based on this early detection, use novel "human-in-the-loop" intervention methods specific to the needs of the individual to improve the quality of life of the older person before severe problems arise. This paper discusses our research involving older adults who live in aware homes. We discuss our approach in detecting changes in emotional wellbeing, combining behavioural data from embedded sensors in the home with self-report data from the residents. We also discuss the challenges involved in interventions to support emotional wellbeing.

Introduction The 2020 US Presidential Election captivated the US public resulting in record turnout. In the months preceding the elections COVID-19, racial injustice and the economic downturn had a daily impact on the lives of voters. In this research, we analyze the sleep behavior of Americans in the lead up to the Presidential Elections. We examine specifically the nights of the Presidential and Vice-Presidential Debates and Election Night. Methods We examined sleep data from the PSG-validated SleepScore Mobile Application, which uses a non-contact sonar-based method to objectively capture sleep-related metrics and self-reported lifestyle data. The data set included 123,723 nights (5,967 users residing in the US, aged 18-85, mean age: 46.6 +/- 16.7 years, 52.3% female). Data from September 1st until November 3rd were included. This covered the nights of the Presidential Debates (Tuesday 09/29/2020 and Thursday 10/22/2020) and the Vice-Presidential Debate (Wednesday 10/07/2020). El...

Introduction Big data collected using consumer sleep technology can provide objectively measured insights on sleep behavior in the real-life environment. It has the advantage over self-report data of being less prone to bias. Here we used a non-contact bio-motion sensor to remotely capture objective sleep data. We analyzed 168432 nights of sleep data to test if differences between weekday versus weekend sleep behavior, known from self-report, would still hold using objective data in a large population. Methods Sleep data was acquired using the SleepScore Max remote sleep sensor and included 168432 nights (2730 users, mean age: 46.6 +/- 11.8 years, 33% female, all resident in the USA). Analysis was restricted to those of working age; adults between 20-65. Any sleep which ended from Monday to Friday was considered weekday sleep, and any ending on Saturday or Sunday as weekend sleep. Data records were inspected and cleaned before analyzing. Descriptive statistics and independent t-test...

Introduction Seasonal effects in sleep are often attributed to day length; however, change in obligatory daily activities might also have an impact on sleep behavior. Longitudinal measurement using consumer sleep technology enables the observation of patterns in sleep behavior in the home environment. We analyzed the impact of parenthood and gender on total sleep time (TST) over the summer break period using data collected in the home. Methods Sleep data were collected using the SleepScore mobile application from October 2018 through October 2019, with the summer break period defined as June 25th - August 5th. U.S. age and gender matched samples of parents and non-parents were selected using Mahalanobis distance from a pool of users more likely to have school-aged children. The final samples included n=345 parents (38.7 +/- 4.5 years) and n=345 non-parents (37.8 +/- 4.7 years); both groups were 46% female. Only weeknights (n=34,323) were analyzed to maximize impact of school schedul...

Introduction Bedroom temperature can influence nocturnal thermoregulation and sleep. To date, limited, small experimental studies have shown that bedroom temperatures outside the recommended range of 65 and 70°F can negatively impact sleep. However, this association has not been studied in a large-scale data set. Using over 3.75 million nights of objectively measured data, we analyzed the associations between habitual bedroom temperatures and sleep. Methods Over 3.75 million nights of sleep and bedroom temperature data were collected using S+ by ResMed technology from 34,096 Individuals (57% male, 20-90 years, mean age 48.7 +/-14.5 years, all US residents). Multilevel regression analyses were used to analyze associations between bedroom temperature and sleep. A stricter alpha level of 0.001 was used to account for the large number of observations in the dataset. Results Bedroom temperature was above 70°F for 69% of nights, with the average temperature ranging between 68.8 and 76.2°F...

Introduction The COVID-19 pandemic profoundly altered individual lifestyles, reducing commutes and restricting nocturnal in-person socialization. We examine whether the stay-at-home orders and the attendant increase in sleep scheduling autonomy, impact bedtimes and waketimes and influence circadian preference alignment. Methods We compared bedtimes and wake times during the 4 weeks before and after a March 19th, 2020 stay-at-home order announcement. Data from the PSG-validated SleepScore Mobile Application were analyzed. Users answering a circadian preference question (a five-point Likert scale ranging from “definitely a morning person” to “definitely an evening person”) who also recorded 10 or more nights of sleep both before and after the March 19th announcement were included in the analysis. The data set included 69,656 total nights of sleep from 1,487 users: 51.0% female, age range 18 to 91 years (mean = 50.3 +/- 30.3). Differences in average bedtime and wake time before and aft...

ABSTRACT With an ageing population and the constant need towards improving the quality of life for older people in our society, there comes an urgent challenge to support people where they live in an environment that adapts to their needs as they age. While much research on ubiquitous sensor systems and telehealth deviceWith an ageing population and the constant need towards improving the quality of life for older people in our society, there comes an urgent challenge to support people where they live in an environment that adapts to their needs as they age. While much research on ubiquitous sensor systems and telehealth devices focuses on this need, many of these solutions operate at less than full capacity, and with little scope at present to assess everyday aspects of wellbeing. They focus on detecting sudden critical physiological and behavioural changes and offer few mechanisms to support preventative actions. The challenge of predicting changes and prompting positive preventative intervention measures, aiding the avoidance of severe physical or mental harm, has not adequately been addressed. This paper discusses our experiences of designing, deploying and testing an integrated home-based ambient assisted living (AAL) system for older adults, consisting of ambient monitoring, behaviour recognition and feedback to support self-management of wellness, in addition to providing feedback on home security and home energy. It offers a complete system overview of an AAL solution in smart environments and discusses our lessons learned with the goal of assisting other researchers in the field in designing and deploying similar environments.

In this article we examine data from eight purpose-built aware homes over a six-month period, looking at presence in rooms to try to determine patterns among the older residents. We look for homes that have similar movement patterns using cluster analysis. We also examine how movement over days clusters within individual homes. Our analysis shows that different homes have distinct movement patterns but within individual homes residents have strong movement routines.

An audio-visual speaker identification system is described, where the audio and visual speech modalities are fused by an automatic unsupervised process that adapts to local classifier performance, by taking into account the output score based reliability estimates of both modalities. Previously reported methods do not consider that both the audio and the visual modalities can be degraded. The visual modality uses the speakers lip information. To test the robustness of the system, the audio and visual modalities are degraded to emulate various levels of train/test mismatch; employing additive white Gaussian noise for the audio and JPEG compression for the visual signals. Experiments are carried out on a large augmented data set from the XM2VTS database. The results show improved audio-visual accuracies at all tested levels of audio and visual degradation, compared to the individual audio or visual modality accuracies. For high mismatch levels, the audio, visual, and auto-adapted audio-visual accuracies are 37.1%, 48%, and 71.4% respectively.

The performance of deployed audio, face, and multi-modal person recognition systems in non-controlled scenarios, is typically lower than systems developed in highly controlled environments. With the aim to facilitate the development of robust audio, face, and multi-modal person recognition systems, the new large and realistic multi-modal (audio-visual) VALID database was acquired in a noisy “real world” office scenario with no control on illumination or acoustic noise. In this paper we describe the acquisition and content of the VALID database, consisting of five recording sessions of 106 subjects over a period of one month. Speaker identification experiments using visual speech features extracted from the mouth region are reported. The performance based on the uncontrolled VALID database is compared with that of the controlled XM2VTS database. The best VALID and XM2VTS based accuracies are 63.21% and 97.17% respectively. This highlights the degrading effect of an uncontrolled illumination environment and the importance of this database for deploying real world applications. The VALID database is available to the academic community through http://ee.ucd.ie/validdb/.

1276 shoeprints were collected at a scientific exhibition. Details regarding the age groups of the participants, style, size and manufacturer/brand of their shoes were recorded. The impressions were assigned to pattern groups showing that the most common pattern was present in only 1% of the population studied and most patterns were much less common. The impressions were digitized and a system developed for automatically sorting a database of images of outsole patterns in response to a reference image. The database images are ranked so that those from the same pattern group as the reference shoeprint are likely to be at the start of the list. A database of 486 complete shoeprint images belonging to 142 pattern groups was established with each group containing two or more examples. Tests of the system have shown that the first-ranked database image belongs to the same pattern group as the reference image 60% of the time and that a correct match appears within the first 5% of the ranked images 88% of the time. The system has translational and rotational invariance so that the spatial positioning of the reference shoeprint images does not have to correspond with the spatial positioning of the shoeprint images of the database. The performance of the system for matching partial shoeprints was also determined.

Recent research suggests that falls are the most common cause of injury and disability in older persons. Invasive systems or body worn sensors can be employed in controlled clinical and laboratory settings to determine clinical measures of gait and stability. This study by contrast aims to explore how video technology, can be employed to unobtrusively determine the same measures. Data from 63 elderly subjects, recruited through a research clinic was analyzed. The derived parameters include: the walk time, the number of steps of the TUG test and stability out of the turn. The results show that video analysis can be used to automate current clinical measures of gait and stability as well as to inform future automated interventions.

The Visual Evoked Potential (VEP) has long established itself as a useful diagnostic tool for clinical use. The VEP is typically recorded in clinical environments due to equipment and experimental set-up. An alternative is proposed in this paper, whereby a light-weight, portable neurological monitor using active dry-electrodes is proposed and VEP data compared against clinically recorded data. It has been shown that a correlation coefficient of greater than 0.8 can be achieved with this set-up, demonstrating the diagnostic benefit of the system for both remote diagnosis and neurorehabilitation.