Table 2.
Summary of research works that use passive infrared motion sensor technology.
| Research Authors (Year) | Target | Research Techniques | Results |
|---|---|---|---|
| Alwan et al. [32] (2005) | Recognition of activities of daily living | The work used the following approaches: Rule-based recognition of activities (e.g., eating and showering); Fifteen on/off switches in different places, such as the microwave oven and different doors; Binary features (on/off) were used for rule-based recognition of activities of daily living; More than five weeks of activity monitoring; Subjects were provided portable personal digital assistant (PDA) devices for recording ground truth data. | 91% sensitivity; 100% specificity. |
| Austin et al. [33] (2011) | Gait analysis | The work used Gaussian mixture modeling on motion sensor data for three years of residence monitoring of different people. | 95% accuracy. |
| Austin et al. [34] (2011) | Gait analysis | The authors applied Gaussian-kernel-based probability density functions for three years of monitoring of two elderly subjects. | The approach detects abrupt changes in gait function and slower variations of gait velocity over time. |
| Barger et al. [35] (2005) | Recognition of activities of daily living | The work described probabilistic mixture model raw motion sensor data for recognition of different activities. Subjects were monitored for 65 days. Then, results were accumulated. The project utilized of a set of low-cost motion sensors. Two types of evaluations were performed: work and off-days. | The motion sensor data were grouped into 139 clusters. The experimental results showed that there were some frequent clusters that occurred consistently over time with low classification uncertainty. |
| Celler [36] (1995) | Recognition of activities of daily living | It was a pilot project with five months of monitoring the functional health status of the elderly at home. Parameters that are sensitive to changes in health were continuously recorded. | The project explained the technical functionality for monitoring the functional health status of the elderly in the smart home. |
| Cook & Schmitter-Edgecombe [37] (2009) | Recognition of activities of daily living | The work adopted Markov models for modeling daily activities. | 98% accuracy. |
| Dalal et al. [38] (2005) | Recognition of activities of daily living | The work adopted rule-based recognition based on correlation algorithms. Each elderly person was monitored for 37 days. | 91% sensitivity; 100% specificity. |
| Demongeot et al. [39] (2002) | Recognition of activities of daily living | The authors applied mostly threshold features for rule-based recognition. | Only analytical studies were performed, rather than reporting accuracies of proposed approaches. |
| Fernandez-Llatas et al. [40] (2010) | Recognition of activities of daily living | Simple rules were applied to an ongoing project to focus on various daily activities. | The work was only an analysis of an ongoing project, which was carried out to test different approaches without reporting any specific results. |
| Franco et al. [41] (2010) | Recognition of activities of daily living | The work used circular Hamming distance based on temporal shift, which was applied to monitor elderly persons for 49 days. | Different days were considered to explain the functionality. |
| Glascock & Kutzik [42] (2006) | Recognition of activities of daily living | The work applied Gaussian mixtures to model human activities. The study was performed on two field sites, where elderly monitoring was carried out for a half a year and a full year. | 98% reliability. |
| Glascock & Kutzik [43] (2000) | Recognition of activities of daily living | Multiple activities were annotated based on specific software to monitor behavior. Elderly monitoring was performed for 12 days. | The functionality of the behavior monitoring system was elaborated for different days. It can be used in eldercare centers to obtain temporal information based on behavioral variations. |
| Hagler et al. [44] (2010) | Gait recognition | A simulation study was performed on gait analysis in a predefined laboratory setting. | 98.9% accuracy. |
| Hayes et al. [45] (2004) | Recognition of activities of daily living | A Gaussian-kernel-based approach was described that was based on probability density functions for describing walking in-home. Eight weeks of monitoring of walking was carried out. | 98.1% accuracy. |
| Kaye et al. [48] (2010) | Recognition of activities of daily living and gait | For an average of 33 months, different types of sensors were installed in the homes of 265 elderly people. Different metrics were assessed, such as total daily activity, time out of the home, and walking speed. Participants were also assessed yearly with questionnaires, physical examinations, and neuropsychological tests. | Elderly people left their homes twice a day on average for approximately 208 min per day. Average in-home walking speed was 61.0 cm/s. They spent 43% of days on the computer for an average of 76 min per day. |
| Lee et al. [49] (2007) | Recognition of activities of daily living | A behavioral monitoring system was developed for elderly people who are living alone. The PIR-sensor-based in-house sensing system could detect the motion of an elder and send the data to a database. In addition, a web-based monitoring system was developed for remote monitoring of the elderly by caregivers. The system was installed in nine elderly homes for three months. | 86.6% accuracy. |
| Noury & Haddidi [50] (2012) | Recognition of activities of daily living | A simulator was proposed that focuses on human activities based on presence sensors in the smart home for elderly healthcare. Previously recorded real activity data were used to build a mathematical model that was based on HMMs for producing simulated data series for various scenarios. In addition, similarity measurements were obtained between real and simulated data. | 99.91% accuracy. |
| Shin et al. [51] (2011) | Recognition of activities of daily living | Several sensors were installed in different places in a smart home to monitor abnormal activity patterns. Observations were made for 51 and 157 days. | 90.5% accuracy. |
| Tomita et al. [52] (2007) | Recognition of activities of daily living | A case study was performed for two years of elderly monitoring in smart homes. | 91% recommendation. |
| Virone [53] (2009) | Recognition of activities of daily living | It was a simulated case study in which a pattern recognition model for daily activity monitoring was tested. Activity deviation was also considered during activity monitoring. | 98% accuracy. |
| Wang et al. [54] (2012) | Recognition of activities of daily living | Activity pattern deviations were considered for early detection of health changes. Dissimilarities among different activity density maps were computed to automatically determine changes in activity patterns. Elderly subjects were monitored for one, four, and three months. | Dissimilarities among activity density maps were in the range of 0.30–0.52. |
| Willems et al. [55] (2011) | Recognition of activities of daily living | A pilot study was performed to examine potential effects of activity monitoring on users and formal and informal caregivers. The study was performed based on the observations from two years of monitoring. Various questionnaires were used to assess quality of life and health status. | The functionality of the system was illustrated in detail. After the assessment, no significant variations were found based on the client questionnaires. |