Wearable and Smartphone-Based Sensors in Support of Human Comfort-Driven Structural Analysis of Building Components

: The continuous progress and advancement of innovation in technology and development of digital tools makes modern structural engineers and technicians of the building and construction sector increasingly able to solve a multitude of design issues. There, in most of cases, they can take advantage and support from often low-cost and event portable sensors characterized by generally high accuracy and commercial availability. In this paper, the attention is focused on the analysis of recent investigations which have been carried out with the scope of human comfort-driven structural analysis and design of building components. More precisely, the use of wearable and smartphone-based sensors for the experimental derivation of mechanical parameters of utmost importance and technical interest for design of pedestrian systems is explored. On one side, as shown, the elaborated setup makes easy and rapid the acquisition of body motion parameters for pedestrians moving on different substructures. At the same time, relevant feedback could be possibly derived on the side of customers and corresponding comfort.


Introduction
Worldwide, it is generally recognized that human comfort in buildings and constructions is a target for a multitude of reasons and multidisciplinary aspects [1,2]. In the same way, however, there is a clear view of uncertainties and complexities which are intrinsically involved in comfort analysis and optimization. As a matter of fact, the definition of human comfort itself is rather wide [3][4][5], and thus necessitates specific design assumptions and performance indicators, such as measures to optimize thermal, acoustic, lightening, but even vibration serviceability aspects and many others.
Overall, a major advantage in the construction sector has been offered, especially in the last years, by a multitude of sensors and devices which are commercially accessible, often low-cost, and aimed at supporting specific activities and consequent decisions. Wearable and smartphone-based sensors can be found in daily activities, and can be optimized as health-monitoring tools but also to improve human well-being against potential risks. Safety, in this sense, is a primary target for those operations in which humans can be potentially subjected to danger and risk of damage ( Figure 1a). Typical examples can take the form of (i) smart watches (for health and activity monitoring, fall detection, and safe communication); (ii) smart boots (able to detect pressure from shocks and falls, and inclusive of location sensing); (iii) smart helmets (where sensors can be used to monitor fatigue, prevent microsleeps, detect collisions); (iv) augmented reality glasses (for the identification of hazardous materials, and visualization of safety protocols); (v) smart body wears (to track body core biometric parameters); etc. Similar devices can be also intended to be used as sophisticated sensors and instruments in support of engineering issues and problem solving. In this manner, human and biometric parameters are tracked for health monitoring and risk prevention scopes, but can be further exploited as key input performance indicators for structural design, structural health monitoring, functionality and safety maintenance, structural optimization. As such, it was for example shown in [6] that human behaviours on constructed facilities reflect both mechanical and structural conditions and phenomena, but also nervous states and emotional reactions which are mutually affected by comfort and structural responses (Figure 1b).
In this context, is it thus possible to use wearable and smartphone-based sensors for coupled well-being optimization and structural design improvement, based on comfortdriven considerations? The question is rather challenging and certainly necessitates of wide experimental validation. Besides, the present research study tries to partly answer this question by taking into account some experimental evidences and typical structural issues of vibration serviceability assessment. How body motion is affected and modified by the built environment? And how the structural features of a given load-bearing system can modify the behaviour of customers?

Sensors for Human Comfort-Driven Structural Design
The current investigation starts from the basic consideration that there is a reciprocal and mutual interference and interaction of human behaviours, and thus comfort levels, and the structural features of a given load-bearing system [7,8]. This interaction may result from multiple aspects, such as for example the aesthetic impact of a construction (and thus emotions [9][10][11]), or the sensitivity to human motion (like in terms of perceived vibrations [12,13]), and the subjective reaction of humans to structural responses. This is particularly relevant when "emotional architectures" are the context of human activities [9][10][11], and in those configurations, glass material has a primary emotional role on humans, for many reasons [12][13][14][15]. According to Figure 1, wearable sensors able to capture specific biometric parameters of customers can thus have a key role in the quantitative measure of human reactions [14,15].
The open question is thus not only how to optimize comfort of customers against a given external action / conditions, but how can we take advantage of quantitative measure of nervous reactions, emotions, and body motion features to efficiently support the design of those architectures and constructions, and thus integrate traditional and consolidated mathematical models (which are typical of structural / building design) with human parameters.

Present Elaboration
The overall experimental strategy is based on the assumption that human-structure interaction (HSI) phenomena are intrinsically involved in the design of any kind of pedestrian structure [16,17]. Furthermore, additional basic considerations summarized in Figure 2 are taken into account, namely: • flexible pedestrian systems involve magnified HSI phenomena on pedestrians [6], and thus their body motion is reciprocally affected by the structural response but also by possible motional states (Figure 2a and [14,15]), and • wearable sensors can be efficiently integrated to classical instruments for structural health monitoring purposes (Figure 2b and [6,18,19]); • finally, glass material in buildings and constructions is a critical component to design, in terms of structural vulnerability against mechanical loads [20], intrinsic transparency and its emotional effects on customers [14,15], intrinsic flexibility and sensitivity to vibrations [13,21,22]. Based on the above aspects, this experimental application aims thus at demonstrating that there is a modification of human behaviours on glass floors, and different mechanical reactions are transferred among them during motion, thus both human comfort and structural design are mutually affected by each other.

Setup
Most of experimental records during tests were collected for the author equipped by sensors while walking normally on structural glass pedestrian systems [6,18]. In this regard, it is worth to remind that the herein presented experimental strategy aims also at supporting the assessment about the possible use of low-cost, commercial, wearable sensors in support of tests to carry out on various building configurations [6]. Most importantly, the overall analysis is based on the acquisition of body motion features, especially human-induced reaction forces, for the mechanical analysis and quantification of biodynamic parameters of technical interest for structural design [18].
Among others, the vertical reaction force due to pedestrians is in fact certainly of primary interest. At the same time, it is known that is rather hard to calculate and can involve mutual interaction of pedestrians and substructures [23,24]. In the present study, a primary attention is hence spent for the experimental derivation, based on commercial sensors, of the well-known Dynamic Load Factor (DLF) corresponding to human-induced reaction forces, which represents a parameter of utmost interest for structural analysis of floor systems, as well as for comfort analysis and optimization of pedestrians. Most importantly, for structural analysis, average trends of DLF for the examined walking config-urations are the primary input for deterministic approaches like the Fourier series approach, where the DLF is needed to describe the mechanical load on a given structural system. To this aim, three different slab systems as in Table 1 were taken into account (two of them made characterized by transparency and high flexibility).

Sensors
The present analysis starts from the study concept reported in Figure 2b. In addition, moreover, the use of a Micro Electro-Mechanical Systems (MEMS) triaxial, Wi-Fi accelerometer and inclinometer on pedestrian foot is assessed, as it could further facilitate the analysis of body motion during walks. A typical example can be seen in Figure 3, where the vertical acceleration component measured at the body centre of mass of the author (BCoM) is plotted as a function of time, during a walk on a rigid floor. Also, the corresponding rotation for left foot is presented in the figure. To note that rotation angle for foot is scaled to 1/10 th , to facilitate the readability of graphical comparison. For the analysis of DLF trends (amplitudes and sensitivity to floor configuration, under motion of the same pedestrian), three different slabs as in Table 1 were taken into account during tests.
. Figure 3. Analysis of body vertical acceleration and foot rotation during normal walk, based on wearable sensors.

Experimental Evidences
For the present study, based on experimental records like in Figure 3 and others, a major advantage was taken from the use of Matlab ® for curve -fitting and consequent extrapolation of DLF values. The same operations were repeated, over the number of available walking records, for harmonics corresponding to vertical, longitudinal, and lateral human-induced loads during motion. In Figure 4, it is thus possible to see the trend of first five harmonics for the author walking normally on three different substructures, two of them transparent and flexible. The experimentally derived curves are grouped in terms of reaction component. To note that present evidences are proposed for a fixed walking frequency of fp= 1.5 Hz for all the examined substructures.
The average DLF amplitudes reported in Figure 4 and experimentally derived for the transparent / flexible / lightweight SLAB#2 and SLAB#3 systems are relatively small, compared to the rigid concrete system noted as SLAB#1. Furthermore, in Figure 4 (a) it can be noted that he second harmonic of vertical force for SLAB#3 is associated to higher average DLF compared to the corresponding experimental evidence for the first harmonic. This finding is also in line with several literature studies (such as [25][26][27]), where it has been confirmed that flexible floors with high sensitivity to human-induced effects are characterized by typically pronounced second harmonic and associated DLF.
When the presently elaborated DLF values -based on wearable sensors -are compared with a number of literature efforts, typical evidences can be observed in Figure 5.

Summary and Future Developments
With a focus on the first harmonic of vertical reaction force, it is worth to note that the DLF values elaborated from present study are in close correlation with literature, especially for the reinforced concrete SLAB#1. At the same time, it can be seen in Figure 5 that DLF experimental evidences for transparent / flexible substructures (SLAB#2 and #3) are clearly lower than the concrete system, with an average DLF quantified in ≈0.11-0.12 for both, and associated to ≈−37% DLF scatter towards the rigid / opaque system (#1).
Such a kind of output suggests on one side that the use of body sensors for integrating structural design performance indicators is particularly efficient (as demonstrated for example by comparison of literature and present results on SLAB#1). At the same time, all basic assumptions motivating the present experimental investigation are confirmed in Figure 5, where it can be seen that transparent / flexible systems #2 and #3 involve a marked modification in human behaviours and body motion. Future studies will be thus extended to confirm present evidences and support the development of a robust methodology in terms of human comfort-driven structural design optimization.