Abstract
In this chapter, we present a background on the state of the art regarding Ambient/Active Assisted Living (AAL) related topics (i.e., Ambient Sensing, Medical Technology, and Geriatrics and Sociology). Later in the book we present the basis for current, new, and common technologies on the market (presented in Chapter 4), and ongoing research in AAL (examples presented in Chapter 5). The presented interdisciplinary content consists of social, engineering, medical, electrical, and mechatronics science, and together, they encompass the field of AAL, as well as eHealth.
In this chapter, we present a background on the state of the art regarding Ambient/Active Assisted Living (AAL) related topics (i.e., Ambient Sensing, Medical Technology, and Geriatrics and Sociology). Later in the book we present the basis for current, new, and common technologies on the market (presented in Chapter 4), and ongoing research in AAL (examples presented in Chapter 5). The presented interdisciplinary content consists of social, engineering, medical, electrical, and mechatronics science, and together, they encompass the field of AAL, as well as eHealth.
In recent years, AAL and eHealth have become important topics because of increased life expectancy. Even though this topic is age related, young people have already started to have a high interest in this field due to the potential for the prevention of age-related diseases.
Therefore, this chapter introduces the technological feasibility for Ambient Sensing and will be linked with the possibilities of medical technology and the necessary geriatrics background. At the end, the potential of technology of robotics in this field is presented.
2.1 Basic Knowledge in the Field of Ambient Sensing
When people enter a stage of physical and cognitive decline, typically associated with the natural aging process, the independent exercise of Activities of Daily Living (ADLs) becomes increasingly difficult [35]. If sensing and intervention technologies are not developed, this decline may progress prematurely and unnecessarily to the point when affected groups will eventually not live independently at home and become a burden on family members and institutionalized nursing-care systems. These considerations are particularly important since every emerging industrial nation is experiencing demographic change problems [36]. In this aspect, Ambient Sensing solutions could undoubtedly contribute to addressing these emerging problems.
Ambient Sensing refers to environments populated with sensors responsive and sensitive to the presence of people. An environment populated with electronic elements and microsystems can undoubtedly contribute to enhancing the independence of the elderly by introducing a degree of ambient assistance. Monitoring people’s movements in complex environments, analyzing the resulting motion patterns, and understanding people’s gestures corresponds to a high level of visual competence that can most appropriately be identified as Ambient Intelligence (AmI) [37]. Thus, Ambient Sensing – belonging to the AmI research umbrella – builds upon advances in sensors and sensor networks, pervasive computing, and artificial intelligence. Because these contributing fields have experienced tremendous growth in the last few years, Ambient Sensing has strengthened and expanded, revolutionizing daily human life by making people’s surroundings flexible and adaptive. Technologies are deployed to make computers disappear in the background, while the human moves into the foreground in complete control of the augmented environment. Ambient Sensing systems are a user-centric paradigm, supporting a variety of artificial intelligence methods and works pervasively, nonintrusively, and transparently to aid the user. They support and promote interdisciplinary research encompassing the technological, scientific, and artistic fields, thus creating a virtual support for embedded and distributed intelligence. They will eventually become invisible, embedded in our natural surroundings, present whenever we need them, enabled by simple and effortless interactions, attuned to all our senses, adaptive to users, context-sensitive, and autonomous.
The basic idea consists of a distributed layered architecture enabling omnipresent communication, and an advanced human–machine communication protocol. The Ambient Sensing paradigm sets the principles to design a pervasive and transparent infrastructure capable of observing people without interfering with their lives but at the same time adapting to the needs of the user. It must be noted that populating a home environment with sensors must be performed following a space-efficient utilization scheme. Elderly people, and especially the ones using assistive devices such as wheelchairs and rollators, require increased barrier-free space for mobility purposes.
In the following sections, an overview about the rule of vision systems as well as radio-frequency identification (RFID) technology in this context is provided in order to explain the different underlying technologies used in the realization of such systems, and their corresponding application areas.
2.1.1 Vision Systems
Vision is arguably the strongest of the senses in humans and in many other creatures. It allows us to fully understand the surrounding environment by providing spatial information of objects around us. With this amazing ability, we can determine the position, identity, and status of the various objects in the environment, so that we can interact with and react to various unexpected events. It is therefore reasonable that we attempt to give a sense of vision to the machines in order to turn them into even more useful and efficient tools. Many vision-based sensors have therefore been developed throughout the years. Three-dimensional (3D) vision systems base their operation on the collection of stereoscopic image pairs and on decoding the depth of information by examining the relative displacements of objects within a pair of images, relative to each other. This process is called stereo-photogrammetry. The observation, using our eyes, allows us to perceive the relative distance (depth) of objects that enter our field of vision. However, the human brain is the mechanism that is responsible for successfully decoding the depth information, i.e., the stereoscopic image pair, giving us the ability of depth perception. Conversely, in stereoscopic vision systems, an algorithm that can analyze the digital images taken by a stereo camera pair and recover the important depth information by sampling the areas which are illustrated in the optical scene must be devised. Depth estimation in a scene using image pairs acquired by a stereo camera setup is one of the important tasks of stereo vision systems.
Disparity map extraction of an image is a computationally demanding task; practical real-time hardware-based algorithms require high device utilization recourse usage, and depending on their disparity level operational ranges, this may lead to significant power consumption. Apart from digital camera sensors, other technologies which base their “visual” sensing performance on laser, such as Light Detection and Ranging (LiDAR), or infrared such as Depth Sensors have also been developed.
2.1.2 RFID Technology
In recent years, RFID technology has moved from insignificant into conventional applications that aid in simplifying the handling of items and objects. RFID enables identification from a distance, and unlike the earlier barcodes technology, it functions without requiring a line of sight or a specific visual pattern to be detected, recorded, and processed [38]. RFID tags (see Figure 2.1) support a larger set of unique IDs than barcodes and can incorporate additional data such as manufacturer, product type, and even measure environmental factors such as temperature. Furthermore, RFID systems can discern many different tags located in the same general area without human assistance. In contrast, consider a supermarket checkout counter, where the personnel must orient each barcoded item toward a laser scanner reader in order to identify it. If all items had an RFID tag attached on them, the checkout process on the counter could have been fully automated without explicitly requiring human assistance.
Figure 2.1 Exemplary RFID Tag.
Many types of RFIDs exist, but at the highest level we can divide RFID devices into two broad classes [39]: active and passive tags. Active tags require a power source, i.e., they are either connected to a powered infrastructure or use energy stored in an integrated battery. One example of an active tag is the transponder attached to an aircraft that identifies its national origin. However, batteries make the cost, size, and lifetime of active tags impractical for most small-scale applications. Passive RFID tags are thus preferred because they don’t require an external supply source. The tags also have huge operational lifetimes and are tiny enough to fit into a practical, adhesive label.
A passive tag consists of three parts: an antenna, a processing unit attached to the antenna, and some form of encapsulation. A tag reader is responsible for powering and communicating with a tag, which is attached either to a personal computer or to a digital communication network. The tag antenna captures energy and transfers the tag’s ID. The tag’s processing unit is responsible for coordinating the communication and transmission process. The encapsulation maintains the tag integrity and protects the antenna and processing unit from environmental conditions or damage.
In a home environment, this technology can be utilized to assist elderly people by providing them with a real-time inventory of their high priority items. Experiments conducted during the proposed study, revealed the efficiency of this “invisible” technology, and the variety of potential applications to which this technology can contribute. By combining computerized databases and inventory controls linked through digital communication networks spread across the home environment and across a global set of locations, RFID technology can efficiently pinpoint individual items as they move between locations, warehouses, vehicles, and stores.
2.2 Basic Knowledge in the Field of Medical Technology
Life quality and health are very closely related to each other. It does not matter how rich or poor someone is, as soon as their health is gone, they suffer. Therapy is necessary to successfully cure a person. The type of therapy is dependent on the disease, and therefore it is a major task for a physician to identify the cause of a disease and then decide together with the patient how to recover or at least treat the disease.
Depending on the kind and stage of a disease, it may happen that the affected becomes unconscious or comatose suddenly and without any warning symptoms. This happens with diseases that have silent symptoms and break out suddenly. For example, heart attacks can be a result of a permanently high blood pressure. To treat a person, the cause of the disease must first be identified. An anamnesis (if possible) in combination with first measurements (e.g., blood pressure, pulse, glucose, and ECG) allow the physician to find the proper treatment, and sometimes to identify additional health risks. For example, type II diabetes is a typical disease, which normally in a physical examination gets accidentally diagnosed [40] by a physician because of other reasons. However, this means regular physical examinations increase the chances to recognize the beginning of a disease at an early stage, where the chances of a cure are high. The field of medical technology therefore focuses a lot on identifying diseases.
More people, old and young, are aware of these facts. Therefore, wearables, which measure physiological parameters, have successfully entered the consumer market. Wearables, however, tend to be forgotten to wear, and many people feel that wearing them all day is inconvenient. In order to improve health screening and to increase the security of the user, a new strategy has started in this field of research: the unobtrusive implementation of health sensors in the environment.
Sensors must be noninvasive for both physicians and patients and designed to work in the background when implementing health sensors into the user environment. In the following sections, devices used for noninvasive diagnostics are introduced, which are highly interesting because they can be unobtrusively implemented into the user’s environment. The possibilities are too large to present every type of measurement; however, we will give an overview of the most important measurements with regards to AAL and eHealth.
2.2.1 Pulse Oximetry
Pulse oximetry is used to measure pulse and oxygen saturation at the same time. To do this, a sensor has to be attached on the earlobe or finger [41], or at the heel for newborn and premature babies. The sensor consists of a light sensor and an emitting source. There are two possibilities of how the sensor measures and receives pulse and oxygen blood saturation data: either by the transmission of the tissue from the emitted light, or by remission (see Figure 2.2).
Figure 2.2 Left: Exemplary sketch of pulse oximetry on the earlobe with light transmission.
Right: Exemplary sketch of pulse oximetry on the finger with light remission.
The pulse is then counted by the arrival of the absorbed light on the photodiode. The emitted light gets absorbed by a blood wave, which is passing by the measurement spot. Usually for this technology, an LED-emitting light to a photodiode is used [42]. To make the signal usable, a power amplifier is normally used for an analog low-pass filter, which amplifies the signal [43].
To measure the oxygen saturation of the blood, two LEDs are necessary: one should emit light at 660 nm (red light), and the other LED at 950 nm (infrared). The reason for this is due to the extinction coefficient of blood (shown in Figure 2.3, data source [44]). When saturated with oxygen, the erythrocytes absorb more infrared light and absorb less light at 660 nm. The 950 nm LED is used to get a relative value for the reference measurement Hb, whereas 660 nm is used for the HbO2. Using Eq. (2.1) allows for calculating the precentral saturation in relation to the overall hemoglobin.
(2.1)However, this sensor has two weak points; heavy movements can easily disturb the measurement and, at the end, the device can only detect how much of the blood is saturated. Normally the oxygen is binding with the hemoglobin, which the erythrocytes carry. However, other gases like carbon dioxide bind much more strongly than oxygen to the blood. This leads to a perfect measurement result, although the blood does not carry oxygen. This dangerous life situation, which, for example, can occur in response to a fire gas intoxication, cannot be detected using this technology.
Figure 2.3 The extinction coefficient of unsaturated hemoglobin (Hb), and hemoglobin saturated with oxygen (HbO2).
Also, smoking can lead to wrong results. A person who smokes during a measurement period will have 100 percent saturation. However, the true saturation will be visible after some time and will drop to around 95 percent. In [41], the following thresholds for pulse oximetry are defined: A healthy person normally has 95–100 percent saturation. People who have 94 percent saturation or lower normally suffer from hypoxia and need treatment. People with a saturation of less than 90 percent are a medical emergency case.
2.2.2 Blood Pressure Meter
Those who have used a blood pressure meter know that the device gives three values: the systolic blood pressure, the diastolic blood pressure, and the pulse. In the past, physicians and caretakers measured the blood pressure by palpation (according to the Riva–Rocci method) with a cuff. Therefore, e.g., the caretaker, or physician, must find the pulse, e.g., on the wrist of the patient. Once the pulse has been found, the cuff, placed on the upper arm, gets pumped (using a small hand pump) until the pressure is large enough to close the brachial artery [45]. This stops the blood flow to the hand, which is not very convenient if the pressure is high for too long.
Of course, now there is no noticeable pulse. Slowly opening the valve on the cuff leads to a slow reduction of the pressure, which the cuff uses to close the artery. The nurse, or physician, has to wait until the first pulse wave is noticeable on the wrist: this is the blood pressure that is strong enough to open the artery for a very short moment. This value is called the systolic blood pressure. This value is the highest pressure and marks the pressure occurring during a heart contraction.
However, there is also a constant low blood pressure while the heart is not contracting. This value is known as the diastolic blood pressure. To also measure the diastolic value, physicians and caretakers use a stethoscope to hear the Korotkoff-sound (auscultatory measurement).
These Korotkoff-sounds, which sound similar to a heartbeat but are of distinct origins, are the result of turbulences, which occur when the cuff is narrowing the artery (see Figure 2.4) [46]. As soon as the vessel is able to send blood through the occlusion (caused by the pumped cuff), this sound occurs and marks the systolic value (as it is with the palpatory measurement). The Korotkoff-sound stops as soon as the turbulence stops and the blood continues with a laminar flow, which is achieved as soon as the pressure of the cuff is weaker than the lower blood pressure (the diastolic value). This means that the last Korotkoff-sound the physician or caretaker hears, while reducing the cuff pressure by the valve, marks the diastolic blood pressure.
Figure 2.4 Left: No blood flow, because the cuff pressure is higher than the systolic blood pressure. Center: Turbulence blood flow, which causes the Korotkoff-sounds, because the cuff pressure is between systolic and diastolic blood pressure. Right: Laminar blood flow, because the cuff pressure is lower than the diastolic blood pressure.
However, both methods (palpation and auscultation) are very subjective measurement methods dependent on the sensibility and hearing abilities of the person taking the measurement. Herein, this measurement method has been automated. The device in principle consists of an automated pump and a pressure sensor (e.g., a capacitive pressure sensor) measures the pulse. The device pumps the cuff to close the artery and then stepwise reduces the pressure (as depicted in Figure 2.5). Once the cuff has a larger pressure than the cuff, the pulse oscillation increases, which the blood pressure meter marks as a systolic value. Once the oscillation strength returns to its normal level, the device marks the cuff pressure as a diastolic value.
Figure 2.5 Exemplary sketch describing the working principle of a blood pressure meter.
However, the most precise method to measure a blood pressure is the invasive blood pressure measurement. Here, the pressure sensor is not attached to a cuff; it is directly attached via a catheter or cannula to the arterial blood. Using this method allows care staff or a physician to differentiate the blood pressure in the region and system. The blood pressure meters, as described in this section, measures the high-pressure system. However, the blood pressure of the veins belonging to the low blood pressure system is not measured by the commercial devices [47].
Normally the blood pressure should give a result of 100 mmHg up to 140 mmHg (optimal 120 mmHg) for the systolic, and less than 100 mmHg for the diastolic blood pressure value. However, the blood pressure is also dependent on the body region, as well as of the kind of blood pressure system.
2.2.3 Temperature Measurement
The body temperature is a physiological parameter, which can change according to the environment or the health condition. An increased temperature, caused by diseases like influenza, is called a fever. Often, a fever is part of the defensive response of the immune system, with the objective to kill intruding microbiological organisms like bacteria or viruses [48]. Although this means that a fever is supporting the curing process, a fever can be dangerous if the temperature is above 43°C [49]. Therefore, a fever is a good measure of someone’s health. For the elderly, it is specifically important since high fever can harm the cardiovascular circulation, and this often needs to be suppressed.
To identify a fever, touching the forehead may be enough. However, this can be a very imprecise method. Fortunately, fever measurement devices are very cheap, and nearly every drugstore sells such devices. Mercury-in-glass thermometers (see Figure 2.6), which measure the temperature related expansion of the liquid, have now been replaced by the most common digital version (see Figure 2.6) of this measurement device (which measures by a thermistor). The disadvantage of these devices is the fact that they need up to 5 minutes to receive a result. Measurement areas are axillar, oral, and rectal, whereas axillar is not considered to be the most precise measurements area [50], but the axillar measurement is one of the most convenient measurement areas. Of course, there are also other measurement areas, which are quite inconvenient or even painful, e.g., rectal [51].
However, user impatience (especially in small children) often leads to false measurement results. Therefore, devices which measure the fever very fast and reliably by infrared on the eardrum ([52], [53], [54]) now exist. However, there is a high risk with this device related to wrong measurements, e.g., by wrongly holding the tympanic fever measurement device, which is then measuring the meatus, instead of the ear drum. An additional disadvantage is that all the mentioned body temperature measurement methods or devices need direct body contact, which leads to hygiene problems. For example, [55] points out the importance of sterilizing these devices, if they are to be used for more than one person.
Therefore, thermometers which measure fever on the forehead without direct contact by infrared also exist. Depending on the environmental temperature, the accuracy of this kind of measurement may be compromised [56], [57]. On the other side, these kinds of devices allow a fast and hygienic measurement of the body temperature. In the last years, this aspect of fast and hygienic measurements became very important not just for the health care, eHealth, and AAL sectors, but also with regards to travel of tourists and business people through different climatic areas. Traveling brings together people of different immune system strengths, which in return supports the development of epidemics and pandemics (e.g., the influenza pandemic in 2009 [58]). Since influenza is a health and even life risk for the elderly [59], thermal cameras have been used at airports for quick mass screening [60]. Therefore, it is not surprising that there are several studies (e.g., [61], [62], [63]), which have investigated the usage and reliability of thermal cameras for fever measurement.