Review of Measuring Microenvironmental Changes at the Body–Seat Interface and the Relationship between Object Measurement and Subjective Evaluation

Being seated has increasingly pervaded both working and leisure lifestyles, with development of more comfortable seating surfaces dependent on feedback from subjective questionnaires and design aesthetics. As a consequence, research has become focused on how to objectively resolve factors that might underpin comfort and discomfort. This review summarizes objective methods of measuring the microenvironmental changes at the body–seat interface and examines the relationship between objective measurement and subjective sensation. From the perspective of physical parameters, pressure detection accounted for nearly two thirds (37/54) of the publications, followed by microclimatic information (temperature and relative humidity: 18/54): it is to be noted that one article included both microclimate and pressure measurements and was placed into both categories. In fact, accumulated temperature and relative humidity at the body–seat interface have similarly negative effects on prolonged sitting to that of unrelieved pressure. Another interesting finding was the correlation between objective measurement and subjective evaluation; however, the validity of this may be called into question because of the differences in experiment design between studies.


Introduction
Although people have been consistently advised not to sit for long durations without breaks [1][2][3][4][5], prolonged sitting has become an inevitable fact of life for some, with the potential to impact everyone (e.g., long-haul flight travel, car passengers during traffic jams and computer gamers). Furthermore, increasing portability of electronic devices (e.g., smartphones, e-readers, tablet PCs) has made it easy for individuals to prolong their working day or enjoy their leisure time in more sedentary activities. Over 20% of commuters choose to continue working while taking public transport to or from the workplace [2], resulting in an extension of their seated working time by up to one working day (7-8 h) a week. To ensure sales, seat designers need to consider the diverse background of end users and the activities they would be expected to perform while seated. The aim, indeed priority, of seat design is to provide users the experience of feeling fit after sitting for prolonged periods [5][6][7][8][9]. Though anthropometry and activities performed whilst seated are believed to have strong influence on the sitting experience, often it is not possible to create saleable seating with accurate information or to predict usage prior to the design process (e.g., classroom chairs usually have the same dimensions though students may vary greatly in terms of anthropometric characteristics) [2,4,5].
knowledge, the issue of objectively measuring microenvironmental changes at the body-seat interface and their relationship with subjective evaluations of comfort and discomfort has not yet been addressed in the peer-reviewed literature. Therefore, the aim of this study was to investigate the types of devices used to objectify sitting comfort/discomfort measurement and the physical parameters derived to represent changes at the body-seat contact surface. In addition, some recent "state-of-the-art" techniques (e.g., artificial intelligence and deep learning [9]) have also been included, as it is our opinion that these have potential for future research in this area.

Materials and Methods
A range of measurement devices and methods have been employed to model the relationship between measurably objective variables and the seated person's subjective evaluation, in attempts to quantify the extent that the body-seat microenvironment varies while sitting. The search and analysis of the literature reported here were based on objective variables used in such studies (e.g., temperature, relative humidity, movement and pressure) as well as their interaction with any subjective evaluation (e.g., comfort and discomfort). Regarding the subjective elements, several definitions and models of sitting comfort and discomfort have been presented to date [2][3][4]7,8], however, these all tend to use the terms "comfort" and "discomfort", hence the simple adoption of these two terms in our search protocol was considered adequate to reveal most articles. As shown in Figure 1, the human perceptions of sitting comfort and discomfort appear to be affected by many factors [2][3][4] including seat materials and structures, performed activities (e.g., reading, writing or consulting) and length of time being seated. To reflect objectively the influence of afore-mentioned factors, measurement tools have been employed to monitor microclimate and pressure changes at the body-seat interface. A systematic search of the literature was conducted consisting of an electronic database search and reference searching. The retrieved sources included medical (PubMed); engineering (IEEE Xplore, EI Village and ACM) and all science (Web of Science and ScienceDirect) electronic databases. Each database was searched in English only, dating from January 2000 to December 2019. The references of all selected articles were also manually reviewed to identify studies that had not been included by the online retrieval search engines. The inclusion and exclusion criteria for the review were developed a priori. To be included in the literature review, the articles needed to: 1. be published in peer-reviewed journals from January 2000 to December 2019; 2.
be written in the English language; 3.
include objectively measured parameters (using electronic devices) at the body-seat interface; 4. be either cohort studies, cross-sectional studies, case series or case reports; and 5.
include human participants (not manikins).
In the literature review, sitting comfort and discomfort are considered to be independent. The following three categorical terminologies were combined by selecting one from each group and searching within titles, key words and abstracts: group I ("chair", "seat", "cushion", "sedentary", or "sitting"), group II ("comfort" or "discomfort") and group III ("temperature", "humidity", "pressure", "posture" or "movement").
All papers found in the literature search were examined (titles, key words and abstracts) to select those considered meeting the inclusion criteria. Articles were excluded if the titles or key words indicated a sole reliance on either environmental impacts (e.g., thermal environment or moisture) or aesthetic perceptions (e.g., comfort or discomfort). If the titles and key words did not provide sufficient information to enable a clear exclusion or inclusion decision, the abstracts would be scrutinized. In addition, papers discussing only the impact of external stimuli (e.g., brightness, vibration, noise and air flow) on sitter's sensations were also excluded as those influences on sitting were not generated at the sitting interface.
A peer-review process was applied during the full-review stages in which two reviewers participated. An agreement was reached by mutual consent between the reviewers in the few cases of disagreement. Relevant references cited by the selected papers were checked meticulously by screening titles and abstracts ( Figure 2). All selected papers were classified into two categories based on the electronic devices used to collect data samples: microclimate (temperature and relative humidity) measurement and pressure measurement (including in-chair movement).  [15] includes both microclimate and pressure measurements; therefore, it was included in both categories. As a result, although the total number of included publications is N = 54, the sum of the two categories is greater (N = 37 + 18 = 55).

Results
Studies utilizing pressure measurement accounted for 37/54 of all publications included in this review, while those reporting on the microclimate (temperature or relative humidity) contributed 18/54 (NB one article [15] was reported in both categories).
When studying the publications for each category chronologically, it is interesting to find that most of the research work was published recently. For example, 13 out of 18 papers (72%) based on microclimate measurement were published between 2012 and 2019, demonstrating a slightly greater yet generally similar increment rate to pressure measurement at 24/37 (65%).

Microclimate Measurement
In recent years, due to the increasing recognition of the negative effects of unrelieved heat and moisture on human health, microclimate changes at the user-seat interface have drawn more attention [1,23,24]. As changes in this category were originally considered to be an individual's physiological sensation (e.g., "I feel hot/wet"), they used to be subjectively evaluated by questionnaires [25]. However, it is difficult to quantify the varying degrees of such perceptions as they are often individual to the subject. The rapid advancement of Micro-Electro-Mechanical System (MEMS) technology and related electronic instruments (e.g., infrared cameras) has made it possible to monitor the microclimatic changes accurately and reliably at the body-seat interface directly (Table 1). Microclimatic changes at the body-seat interface can be described by using two physical parameters: temperature and relative humidity [25].   Participants were allowed to adjust the temperature by regulating the seat air conditioning system. User's self-selected comfort was attained when there was no further request for any change between two adjustment periods (5 min each).
Thermal comfort was achieved when the seating interface temperature was lower than the body temperature Seat surface temperature was always higher than its interior temperature. Surface moisture on a seat was different from that inside the seat. Three seat cover materials showed no significant difference in subjective thermal evaluation and objective temperature measurement. Objective measurement had a positive relationship with subjective evaluation. Skin wettedness on the posterior torso was significantly different across the three cushions, while skin wettedness on the anterior torso did not. Skin wettedness played more important role in comfort evaluation than skin temperature. Nonflat surface cushions (air-filled cells and bubble-shaped surfaces) showed lower peak temperatures than a flat surface cushion. Gel-filled bubble cushions had lower maximum temperatures than foam-filled bubble cushions. Temperatures at the thighs were higher than at the ischial regions. M = male, F = female. RTD = Resistance Temperature Detector. NTC = Negative Temperature Coefficient. RH = Relative Humidity. * This article also applied pressure sensor and pressure-related contents were put into Table 2. ** In the 5th column, the comfort/discomfort scaling standards were listed. If the corresponding trial was not evaluated subjectively, the content would be indicated N/A (Not Applicable). All cited papers are presented in reverse chronological order.More than two thirds of the retrieved studies (13/18) were published in the past ten years (2010-2019), while only five articles appeared (5/18) in journals between 2000 and 2009, indicating an increasing interest in objectively measuring microclimate changes at the body-seat interface. Among these, seven papers (7/18) compared the objective results with a subjective assessment of comfort or discomfort perception using scale-rated questionnaires [15,27,37,38], self-selected thermal comfort [35] or asking questions related to subjective sensations [28,35].
Owing to the size limitation of the early generations of thermal probes, researchers were constrained in relation to how many probes could be used and where to place them without affecting the participants' comfort directly [36]. As a result, optimal locations were considered to be under ischial tuberosities and thighs [36], based on the assumption that these sites usually accumulated heat and moisture during prolonged sitting. With the rapid improvement of MEMS technology, the chip size and electronic flexibility made it possible to embed an array of sensors inside the cushions [26,27]. As result, a more complete thermal mapping of the contact surface could be constructed by employing imaging algorithms [28]. In order to explore the microclimatic characteristics at the same locations, use of integrated temperature-humidity-sensors has also been reported in several publications [11,31,[35][36][37].
Though most of the research work was conducted in simulated conditions (e.g., laboratories or research rooms), some researchers completed the microclimate measurement in real-life situations such as during on-road driving [37]. It is not possible to determine the role that additional lower limb activity, workplace stresses or ventilation may have had impacts on the outcome in such experiments, however, it could be argued that a more realistic appreciation of the seat function may be possible when measurement is made in the workplace.
To choose a cushion suitable for sedentary activities, a variety of materials have been compared including foam, fabric, straps, air cells and gel bubbles. Additionally, Cengiz and colleagues [35,37] examined the performance of different cushion and cover combinations. The conclusion from objective measurement appears consistent to that from subjective evaluation, namely that the more breathable the materials, the greater the perception of comfort the users report [16,38].

Pressure Measurement
Body-seat interface pressure has been widely explored in the study of sitting comfort and discomfort. As a result, a large number of publications have discussed the negative impacts of applying continuously unrelieved pressure on a person's health ( Table 2). As would be expected, the findings are diverse and include: the average pressure [40][41][42], the centre of force [43], the peak pressure [17,42,44] and the pressure distribution at seat pan and back rest [45][46][47]. In addition to simply quoting the pressure, this parameter has also been used to infer in-chair movement or fidgeting [48].  The pitch of the seat and interface pressure had an impact on prolonged sitting comfort.   Three subjective ratings (comfort, discomfort and overall). The comfort Scale ranged from 0 to 10, Whereas the discomfort scale ranged from 0 to −10.
Young and old subjects had different perceptions of comfort rating for the same car seats  Comfort and discomfort ratings of the whole body and six local parts. Rating scales ranged from 0 to −10, and from 0 to 10, for discomfort and comfort, respectively.
Pressure measurement was considered more suitable for short-time comfort/discomfort evaluation than long-term. Pressure ratios at buttocks were lower than for upper/lower back. 20   Diversity in the methods of expressing pressure was matched by the variety of choices in measurement point. Specific areas chosen for measurement included: the ischial tuberosity [42,67], bilateral thighs and buttocks [42,72], pressure ratios between different body parts (lumbar/total, lumbar/back, buttock/total, buttock/back) [69] and the backrest contact area [42,44].
Although pressure has been considered an important component in the generation of skin ulcers and there has been a rapid development of material science and computer technology, the number of scientific publications related to sitting related pressure measurement is not great (n = 37). However, this research field is growing rapidly, with publications doubling in the past ten years (between 2010 and 2019: 25/37) in comparison to 2000-2009 (12/37).
Sitting postures show strong associations with interface pressure distribution, which not only affects subjective sensation but also can lead to serious skin integrity problems if not adjusted properly. A number of classification algorithms have been used to distinguish between different sitting postures ( Figure 3). In the process of interface pressure measurement, the most common measurement method uses a commercially available pressure mat constructed in the form of a force sensitive resistor (FSR) matrix, such as that produced by XSensor (XSensor Technology, Alberta, Canada). There are issues with such systems, usually related to calibration and lost data due to regular distortion of the mat which breaks down the FSR matrix. In addition, these mats are subject to a phenomenon referred to as "hammocking", where the mat does not follow the contours of the seat, instead forming a cover which straddles across small deviations in the seat. Furthermore, specialized hardware platforms are needed as well as software tool kits. Publications were classified in accordance with the number of sensors being used in the pressure measurement (Figure 4). To reduce the complexity and high power consumption of electronic circuits (the inherent property of resistors), Shu et al. [59] proposed a resistance matrix approach (RMA) which identified sensor outputs by solving the resistance matrix equations. Compared with traditional approaches, RMA improved the efficiency and attenuated the complexity by eliminating the redundant components (e.g., external current sources used to prohibit crosstalk noise). A number of alternative systems have been designed and implemented in the assessment of seating pressure. One such pressure sensitive material is referred to as a "capacitive textile" which is formed of textile-based conductive electrodes placed on both sides of a compressible cushion (e.g., foam) [74]. Foam based spacers can induce hysteresis errors, leading to Meyer et al. [43] employing the Preisach model to improve the measurement accuracy. This method was shown to be capable of resolving and differentiating between different sitting postures.
An electronic textile (eTextile) cushion system was developed by combining a fiber-based yarn and piezoelectric polymer [75,76]. The use of such integrated sensors aims to produce a more comfortable and noninvasive pressure measurement system, being undetectable by the person sitting on it. A problem with this approach is the potential to generate electrical disturbance. Aiming at suppressing the electrical disturbance, a resampling calibration method was proposed by Xu et al. [64] to counter offset, scaling, crosstalk and rotation effects. In addition, the classification accuracy of different sitting postures was improved using a dynamic time-warping algorithm. To enhance the integrity and cover larger areas with a small number of sensors, Ahmad et al. [51] developed a screen-printed piezo-resistive sensor which contains 16 pressure measuring elements. Along with being thin and flexible, the customized sensors exhibited additionally useful electrical characteristics such as high repeatability and reliability (maximum deviation between different sensing elements <8%) [51]. Beyond that, the signal sampling unit was reported to be power efficient and capable of transmitting data to the computer wirelessly. Another merit of this type of piezo-resistive force sensor relates to the relative low expense associated with its manufacture: only half of the cost compared with traditional load cells and force-sensing resistors [40].

Discussion
This study reviewed and categorized studies that utilized electronic devices to determine measurable parameters at the body-seat interface associated with sitting comfort or discomfort. Publications that did not measure the microenvironmental changes of the contact surface were excluded (e.g., brightness, noise, vibration and air flow as well as aesthetic feelings and anthropometry).

Relationship between Microclimatic Factors, Comfort and Discomfort
Probes [15,29,33], sensors [11,26,28] and infrared thermography [29,30,39] have been employed to investigate microenvironmental changes between the body and the seat surface; however, each has limitations. Some initial research secured the sensor to the skin of the subject, this would create a microenvironment which could prevent the measurement from the microenvironment created between the seat material and skin [35,37]. To avoid the delicate electronic connections of sensors and probes, both are usually embedded in a seat cushion, a procedure which leads to the sensor being subject to measuring limitations caused by insulated properties of the cushion materials [15]. Furthermore, obtrusiveness of probes may directly interfere with the skin blood flow or raise the subjects' awareness to the presence of the electrodes [28]. An alternative would be to directly image the region, however, to enable cameras to acquire thermal images has required the participants to stand for short periods, which has many effects both on the subject (e.g., physiological redistribution of blood, thus limiting the pooling effect of sitting) and seat surface (potential thermal exchange between the cushion surface and the environment). In more real-life situations, applications of thermal imaging on wheelchair users could result in more problems than benefits, although it could avoid the need for direct contact while measuring.
Although several disparate cushion materials have been compared in terms of rate of temperature increase and cooling speed after prolonged sitting [10,15,29,30,39], it is challenging to give a consistent conclusion on which is the ideal choice for seat products. This indicates that it is critical to take thermal properties of different materials into consideration when manufacturing cushions. Furthermore, thermal outcomes at the contact surface can be influenced by several components including configuration (e.g., foam-fluid hybrid) [31], ventilation (e.g., strap-based) [10] and structure (e.g., honeycomb-structured) [30]. Consequently, it is crucial to study thermal management capacities of different cushions, especially for prolonged sitting usages (e.g., wheelchair cushion).
Before measuring microclimate changes at the body-seat interface, factors that have impact on the outcomes should be considered, such as the surrounding environments (temperature and humidity), on-road traffic and participants' clothing. As people tend to produce more heat and sweat more easily in hot conditions, studies usually have been conducted in controlled conditions (e.g., air-conditioned laboratories) [35,37]. Furthermore, participants were asked to wear similar or even the same clothing when attending successive trials arranged on different dates.
In addition, experiments [36] have shown that heat begins to accumulate at the body-seat interface as soon as the user sits down. At the first stage (within initial 15 min), temperature increases sharply and then the speed of thermal changes reduces, approaching to and attaining a plateau gradually [36]. Due to this finding [34,36], shorter trial intervals [13,[28][29][30]32,34,39] can be applied to evaluate the thermal properties of different cushion materials. By combining some mathematical models and computer algorithms, thermal changes can also be predicted based on the measured information [36]. Several experimental results showed the nonuniform thermal distribution at the body-seat interface [28,34,36], which highlighted the importance of employing more sensors [71,77] while investigating the thermal changes at the contact surface.
Most of the work (14/18) has employed contact sensors to measure microclimate changes at the body-seat interface, with less than a quarter of publications (4/18) [10,29,30,39] using noncontact tools (i.e., infrared cameras) to measure temperature. However, the noncontact measurements required the participants to be transferred from the seat which inevitably led to inaccuracy due to inconsistencies in the emission of heat. Moreover, thermography is not able to continuously monitor the thermal changes at the body-seat interface.
In terms of studying relative humidity characteristics at the contact surface, compared to temperature measurement there has been little. One possible reason behind this phenomenon might be that the risk of increased moisture on skin damage has been neglected [13]. In fact, accumulated moisture deteriorates the epithelial and subcutaneous tissues and eventually leads to tissue necrosis resulting in pressure ulcers [1]. However, the assessment of relative humidity is not free of complications. Unlike temperature changes at the contact surface, an abrupt relative humidity increase can be seen during the initial contact at the seat surface when a person sits down. This was found to be an artefact caused by moisture from a warmer environment interacting with a colder sensor [26]. In other words, there is a transient impulse response induced during the initial process of sitting down. Unlike the temperature showing significant difference among different materials, relative humidity had no consistent performance which indicates one should be cautious when using relative humidity as a physical factor to study at enclosed interfaces, especially where temperature changes are occurring [13,31]. The combination of relative humidity sensor and temperature sensor on the same structure might help to reduce this artifact in the future and make the relative humidity changes more consistent and reliable.
Although only seven out of 18 publications compared the measured data (temperature and relative humidity information) with questionnaire outputs, the consistency indicates that the objective measurement would be a reliable technique to detect microclimatic changes at the body-seat interface.

Relationship between Interface Pressure, Comfort and Discomfort
Pressure mapping has been widely used to objectively assess and monitor pressure changes at the body-seat interface as accumulated forces (focused or asymmetrical tissue loading) negatively impact the subcutaneous tissues and may lead to discomfort and skin diseases (such as pressure ulcer) in the worst conditions.
In terms of pressure measurement, a number of commercially available products exist, such as the Tekscan (Tekscan Co., South Boston, MA, USA) and XSensor (XSENSOR Technology Co., Calgary, Alberta, Canada) system. These pressure mapping systems have large "sensor" arrays [9,45] being based on a matrix of interconnected piezoresistors [52]. Additional conditioning circuits are necessary to coordinate such a large amount of piezoresistive sensors, resulting in issues related to both reliability and stability of the piezoresistive sensors when used in harsh environments [15,55] or repetitively. Indeed, it is our experience that frequent regular recalibration is advisable in order to ensure reliable measurements. As mentioned above, these sensors and their interlinked circuits can be affected by the environment (e.g., temperature) and the parameter they are measuring (pressure). Prolonged exposure to pressures can lead to creep in the output, furthermore, unloading or reduced loading during prolonged pressurization is associated with hysteresis [54,56].
To illustrate the characteristics of interface pressure, directly measured data were represented in various parameters, including the mean backrest pressure [44,[78][79][80], the maximal backrest pressure [44], the seat pan contact area [44], the mean seat pan pressure [44], the maximal seat pan pressure [44,46] and maximum and mean buttocks and back support pressure [72]. By converting the measured pressure values into other variables, some researchers investigated the changes at the contact interface using the pressure distribution pattern over time [80], average contact area [67] and ICM [48,73]. Combining with image processing technique, Xu et al. [64] located the boundary of pressure maps and analysed the radius between sensor elements to the pressure centre.
However, most of the pressure measurements were carried out in simulated conditions including laboratories or research rooms, the exceptions being a small number that were conducted in real working conditions [41,49,70,81]. As the performed tasks have an impact on the pressure distribution [82], it appears necessary to measure pressure in real working/living conditions [83][84][85]. Both Bontrup et al. [49] and Zemp et al. [41] studied working sitting behaviours, of call-centre employees and office workers, respectively. These two independent research groups reported that both sitting positions and body movements were associated with the activities being performed. The derived parameters included mean number of movements, mean number of positional changes per working hour, mean time period of stable sitting and percentage of transient periods during the whole working period.
During an on-road driving trial [70], interface pressure exhibited weak correlation with subjective discomfort whereas the traffic situations and environmental changes played more vital roles. The main difference between office work and vehicle driving is the activity: driving a car requires a fixed (asymmetric) body posture with hands on the steering wheel and one foot on the accelerator while working in office allows individuals to adjust the chair, desk, screen and keyboard as well as body posture. This indicates that the preferred sitting position is strongly dependent on the task being undertaken and caution should be taken when interpreting data from different sitting conditions [2][3][4].
Based on experimental data of six short-term (15-20 min) driving sessions, Kyung and Nussbaum [67] showed that the relationship between pressure measurement and sitting comfort/discomfort was associated with exposure time. Although discomfort is attributable to longer sitting durations, Fasulo et al. [50] found there were no obvious changes regarding peak pressure and average pressure values during sedentary activities. In addition, discomfort appears to be regional in so much as different body parts showed different levels of discomfort after prolonged sitting, for example Porter et al. [70] observed an increase in discomfort in the back, buttocks and thighs over time (after a 135-min drive). This suggests that subjective assessments should be considered in more detail than simply using an overall score, as increasing regional assessment might help determine where change needs to be made.
Though the relationship between the contact surface pressure and subjectively perceived comfort/discomfort has been investigated previously [2,4], there is a lack of consistency in terms of those postures considered optimal for sitting comfort. Fasulo et al. [50] pointed out that continuous body movements could release the pressure and increase lower body comfort, while Cascioli et al. [48] found that ICM [73] was a strong indicator of sitting discomfort.
There appears to be many factors that can affect sitting comfort/discomfort, such as seat pitch, legroom space and anthropometry. Passengers prefer to sit next to a vacant seat during the long-haul air travel [83]. Moreover, seats with good views and head or footrests reduce the feeling of discomfort [2,45], potentially because of distracting the person sitting. As a result, it would appear sensible to combine pressure measurement with other objective information when evaluating sitting comfort/discomfort.
Beyond the controversy of how to sit comfortably, it is a challenge to distinguish different sitting postures in real-time. Firstly, pressure mats usually contain such large amounts of sensing elements (e.g., Tekscan 32 × 32 = 1024) that ideally require higher sampling frequencies (e.g., ADC unit) to acquire data with the minimal information loss. As a result, a compromise should be considered between higher resolution (more sensing elements) and the cost of hardware. To reduce the number of sensing elements, Xu et al. [64] developed a small sensor array (N = 256) which can achieve similar accuracy to commercial products. By converting pressure measurements to fidgeting movements, Cascioli et al. [48] reduced the number of pressure sensors to four and showed it was possible to resolve an association between discomfort and ICM. Secondly, processing large sensing information is time consuming. Finally, the recognition accuracy relies heavily on efficiency of algorithms. Though machine learning and neural networks have been applied, the accurate classification rate of eight different sitting postures is only 80% [51].
By measuring blood flow, heart rate and body temperatures on legs and left tympanic membrane (core temperature), Hodges et al. [14] compared four types of wheelchair cushions and covers made of different materials and concluded that seat covers played an important role in sitting comfort. As lower-cost electronic chips with greater reliability become available, researchers have attempted to investigate the relationship between body movements and sitting comfort/discomfort by combining IMU, EMG and pressure mats [52,53,55,86,87]. Another instrument that has been used in the evaluation of sitting comfort is the motion capture system which can record body movements by attaching reflective markers on anatomical regions of interest [52]. Such multifactor measurement has been applied to many areas from office chairs [57,86] to car cushions [63] and surgery seats [65].
With the help of multiple sensors, sitting postures can be studied from lower limbs to upper bodies. In addition, the growth in available noninvasive wearable products (e.g., ActivPAL accelerometer, PAL Technologies, Glasgow) makes it possible to continuously monitor daily activities [55]. The acquired information will prove important for healthcare services and cushion purchases. As a result, such innovations should help develop strategies and self-help equipment that can help prevent pressure ulcer formation and improve the quality of life of patients. Furthermore, the fusion of information based on multiple sensors should help enhance the recognition of different sitting postures and serve as reliable feedback to seat designers. One example of this is the application of motion cameras, pressure mats and accelerometers to study an ergonomically redesigned truck seat (Force-3) during long haul driving, [52]. As a consequence of using this system, different sitting postures and body movements were found to be attributable to interface pressure changes. Comparing with the standard truck seat, the Force-3 was found to exhibit higher sitting comfort after two hours in their trials.
Accurately measuring sitting posture provides valuable and complementary information for medical professionals and physical therapists to prevent abnormal neuromuscular activities and biomechanical disorders [74,90,91]. Wearable IMUs have become an effective tool to gain reliable data related to human physical activities in daily life and to monitor sitting patterns which strongly affect comfort or discomfort for prolonged sitting [56,57].
Though measurement of multiple parameters can provide more sitting information, several disadvantages exist regarding the practical application. Firstly, using multiple devices increases the cost of the whole system, however this is not a significant barrier. Secondly, motion cameras can only be used in specific laboratory conditions and reflective markers attached to the participants interfere with "natural" body movements. In addition, existence and awareness of these markers may influence subjective perceptions such as comfort and discomfort. Thirdly, the mass, size and functional recording time of wearable electronics (e.g., IMU) rely on the embedded battery and electronic chips.

Reliability and Validity of Sensors
Although sensors in practical applications have been calibrated by manufacturers, it is necessary to verify their validity and reliability before conducting any measurement [28]. For example, McCarthy et al. [36] and Liu et al. [13,34] utilized traceably calibrated environmental chambers to study the accuracy and linearity of the sensors along with the associated repeatability and hysteresis. The reason for employing such caution is that the measuring range of commercially available sensors is usually broader than the practical microclimate measurement encountered in such experiments (e.g., temperature values at the body-seat interface usually vary between 20 • C and 35 • C, if users sit in a room with normal conditions [36]). Additionally, datasheets of sensors provide a general description of the product instead of one which is specific to the chip used in the experiment.
Regarding pressure sensors, it is critical to calibrate the whole system prior to carrying out any measurement, a suggestion endorsed by both the product manufacturers (in their manuals) and researchers [45,52]. Due to the sensitivity of materials forming pressure sensors, any bending or external force will cause deformation to the components. Consequently, the performance of the whole pressure measurement system must be calibrated before (and preferably after) conducting any trial [58].

Future Work
To avoid bias inherent in the subjective assessment methods, questionnaire-based sitting comfort or discomfort evaluation usually requires the study of a large population, which increases the cost and time consumption. In addition, environmental interaction [88], sitting time [84,92,93] and the task performed [94,95] were found to have influence on individual's judgement and perceptions. As a result, objective measurement can not only serve as a useful tool but provide reliable feedback irrespective of external interference.
With the rapid development of modern technology, noncontact measurements have been applied to seat comfort/discomfort evaluation such as radiography [96], ultrasonography [97,98] and MRI [99][100][101]. After installing specific apps, smart phones could be turned into portable sitting posture recorders, which could be used to study motion as traditional products once did [102].
As the majority of the retrieved research works had recruited young healthy individuals, it will be necessary to include various population groups (e.g., old generation or wheelchair users) to derive more general conclusions. In addition, different seat structures and functions should be taken into account when measuring sitting comfort or discomfort. For example, wheelchair users are usually vulnerable to losing balance or being unable to adjust postures [10,15,54]. As a result, along with sitting comfort, support and protection are important for wheelchair design. As field tests showed that rough roads have negative influence on sitting comfort of drivers [35,37,70], vibration reduction has had to be considered when designing vehicle seats. Regarding comfort measurement of office chairs [57,86,96], adjustability of seat pan and back rest is usually critical as office workers usually have different anthropometry.
Though sitting comfort/discomfort study belongs to the field of ergonomics, it is also associated with many scientific areas such as electrical and electronic engineering, philology, psychology and biomedical engineering as well as medical and healthcare science. It may be better to diversify the composition of research teams by involving experts with different knowledge backgrounds.

Conclusions
The aim of this review was to study typical methods used to objectively measure the microenvironment changes at the body-seat interface. In addition, we also investigated the relationship between objective measurement and subjective evaluation on sitting comfort/discomfort. Though body-seat interface pressure was reported to correlate with perceived comfort/discomfort [2][3][4], other factors (e.g., temperature and relative humidity) should be taken into account when assessing the comfort/discomfort of prolonged sitting. Additionally, there was no consistency in terms of the relationship between various measured variables and subjective perception, probably due to the variations in experiment designs. So, it still appears necessary to determine the most influential factors capable of accurately reflecting the subjective feelings of comfort and discomfort after engaging in sedentary activities.
As a note of caution, although objective measurements can effectively monitor metabolic changes of the human body at the sitting interface and even aid in the development of more comfortable seating, sedentary activity is still going to be harmful to an individual's health. Additionally, prolonged sitting can lead to endothelial dysfunction which is a biomarker of cardiovascular diseases [1,90]. So, it seems more important for end-users to avoid sedentary activities than to choose comfortable seats.