Today, the work environment is challenged by various factors such as personal health, workplace security, employee safety and performance, and workflow efficiency. In particular, these parameters can have a tremendous impact on the closed work environment. Employees in these areas are also exposed to serious health issues and face many pressures, mainly relating to their long stay in the same environment. It is therefore important to keep their vital and health data in order to be able to follow closely their health to anticipate risk situations and be proactive by sending a warning to through the hierarchy.
Indeed, employees can spend more than eight hours daily in such environments which can deeply affect their psychological and physiological equilibrium. Moreover, closed rooms or areas may pose a particular risk because of the lack of ventilation, unpurified air or humidity problems that can eventually lead to serious illness. Many external factors could also interfere, such as sound with high decibel level, electromagnetic radiations, abnormal temperature, lighting not conducive to the conventional eye wavelength range, etc. [1
Besides, in some cases, long and boring stays in closed environments may lead employees toward misbehaviors, like sleeping while on duty, or a deterioration of health conditions, like surprising cramps.
Therefore, an efficient solution for tackling these issues must be implemented to enhance the workflow performance, guarantee a higher level of security, preserve the healthy state of the employee, and prevent unsafe situations.
Technology can, in this context, play a fundamental role in these difficulties. Recently, sensor technology has confirmed its impact on individuals, society and our daily life in terms of monitoring, process control, awareness, security, and safety. A combination of data stemming from multiple sensors, as well as related information, and data fusion techniques led to the emersion of a hardly novel concept known as multi-sensor data fusion (MDF). This has offered the possibility to attain more specific interpretations and inferences than could be obtained with a single, independent sensor. MDF is the subject of diverse military and nonmilitary applications as a form of automated target recognition (e.g., for smart weapons), guidance for autonomous vehicles, remote sensing, monitoring of manufacturing processes, robotics, and so on [2
Moreover, the Shared Sensor and Actuator Networks (SSANs), allow the sensing and communication infrastructure to be shared among multiple applications [6
]. MDF can be considered the key to reducing the huge number of sharing applications by decreasing the amount of sensor-generated data diffused in wireless channels and, accordingly, saving energy. This is understandable since wireless sensors and actuators usually rely on batteries, as their energy source, whose replacement is either undesirable or unfeasible [4
Additionally, various types of scientific research have discussed and resolved many life problems by introducing sensor technology. This technology has been introduced into industrial applications offering monitoring, surveillance and safety solutions [7
]. Moreover, sensors are the core of the medical equipment involved in diagnostics, control processes, and personal health, as well as critical care [8
]. In Reference [11
], the authors proposed a promising solution based on four sensors to ensure the safety of soldiers working in critical zones.
Sensors can enhance many domains, such as fitness and sporting activities [12
], smart cities [13
], along with energy sources such as solar power, batteries and fuel cells [14
]; it can also be used for environmental monitoring, exploring space [15
], and in the security and safety of people and nations [16
Scalable data processing platforms are needed to handle data collected from various sensors. The cloud is considered a perfect computational platform [18
] for hosting data processing applications for smart devices due to its effectiveness and nimbleness.
In this paper, we propose a novel approach based on multi-sensor technologies, taking into account the multiple parameters described above, so as to enhance system security as well as workflow performance. This approach is far from adopting sensor networks [19
]. Indeed, the sensors are characterized by the hybrid aspects of their categories, communication technology, and geographic location. Therefore, technologies and approaches developed in sensor networks do not fit in our present context.
This paper is organized as follows. Section 2
presents, on the one hand, the high levels of requirements of multi-sensor technology and, on the other hand, the different selected sensors and applications for increasing security in critical closed zones. We divided these sensors into three categories: environmental sensors, indirect contact sensors, and direct body contact sensors. The subsequent section (Section 3
) maps the specified requirements with the adopted MDF approach, combines the results obtained from the sensors, analyzes the main security risks and the corresponding countermeasures that the closed environment may face and evaluates the performance and efficiency of this technology when deployed in the architecture of cloud storage. Finally, our conclusions and future work are the subjects of Section 4
2. Multi-Sensor Technology and Effects on Closed Workplaces
This section scrutinizes the issue of the security and safety of workers in mainly closed environments through multi-sensor technology. Undoubtedly, many employees show indicators like mental and physical deficiencies produced by lack of sleep, excessive physical, mental and/or emotional stress, sickness, medications, or a combination of these conditions. This can stem from workplace loneliness, the restlessness or anxiety typically arising from an imminent event or daily life problems that may reduce, at astounding rates, the effectiveness and productivity of the worker and can give rise to safety issues in critical areas [11
]. In this context, multi-sensor technology can be an efficient solution to overcome these difficulties and to improve the performance of workers, apart from the classical strategies and techniques for increasing the outcomes of employees [21
To investigate multi-sensor technology in a specific context, it is highly recommended to identify the requirements which can play a crucial role in its successful implementation [22
]. For that reason, the specification of the high level requirements associated with multi-sensor technologies can blaze a trail towards efficient solutions for closed workplace security. In the following subsection we investigate these requirements.
2.1. High Level Requirements for Multi-Sensor Technologies
In order to combine several sensors in a specific closed environment, high level requirements have to be met. This guarantees a minimum quality so that the collected information has more pertinent value and prompt timing. The most important requirements can be summarized as follows:
Sensitivity: The concept of sensitivity refers to the minimum input of physical parameters that produce a detectable output change. More precisely, it can be defined as either the input parameter variation needed to generate a standardized output variation or an output voltage variation for a given variation in an input parameter.
Resolution: This defines the lowest detectable incremental variation of an input parameter that can be found in the output signal.
Precision: The precision aspect of sensor technology is related to the degree of reproducibility of a measurement. Ideally, the sensor should produce, for a given scenario, the same value every time. Small variations can be detected in the majority of sensors relative to the correct value.
Accuracy: This focuses on the difference between the concrete value of the physical quantity and the measured quantity. Accuracy can be defined by a percentage or in absolute terms.
Range: This indicator specifies the maximum and minimum values that could be captured by the sensor. This range should encompass the minimum and maximum bounds encountered.
Response time: This can be defined as the time in which the output reaches a certain percentage (e.g., 95%) of its final value in response to a step variation of the input. It depends on the sensor nature and should be short enough, compared to the physical aspect, to capture.
Sampling frequency: Some sensors have continuous output during certain time intervals. Technically speaking, a periodic measurement is performed according to the sensor sampling rate. This sampling frequency has to be greater than double the maximum frequency (Shannon theorem) that may be captured. This ensures that the coincidence between the sensor output curve and the actual physical parameters are measured.
Tolerated physical environment quantities: Some sensors may face extreme weather conditions and it is important that the sensor work properly. These physical conditions can be related to temperature, sunlight, wind, smoke, audio noise, electromagnetic rays, humidity, dust, and so on.
Lifetime and maintainability: It is important to estimate the lifetime of the sensor before its adoption in order to adequately plan all the logistics related to installation and maintenance.
Communication medium: In the cases where sensors are integrated in an all-encompassing system, it is necessary to be precise in regards to the communication medium adopted by the sensor. This is particularly useful for multi-sensor fusion as well as multi-sensor fusion.
Interoperability: In a multi-sensor context, interoperability is highly recommended. This allows the integration of sensors with hybrid communication protocols.
Once the high level requirements are specified, we will plan a full-study of the multi-sensor technology. Overall, sensors can be classified into three different packages. The first includes the environmental sensors including air and temperature sensors, CO2 sensors, humidity sensors, radiation sensors, sound/noise level sensors, and so on. The second category, indirect-contact sensors, includes eye motion sensors, heartbeat sensors, body temperature sensors, camera sensors, etc. The last represents the direct-contact sensors containing motion and temperature sensors, smart-health sensors, and fingerprint sensors.
In what follows, we shed light on the most important sensors to be implemented in the closed workplaces and the enhanced safety levels and degree of security of work and workers
2.2. Environmental Sensors
Environment sensors are ordinarily attached to environment monitors. They offer global perceptions of the environmental parameters needed to construct an accurate library monitored by dedicated data and computation centers. These sensors are available as heat and/or humidity, air flow, and liquid. The computational center triggers a specific action when any sensitive limit is reached.
2.2.1. Air Sensors
It is very valuable to collect information from the workplace atmosphere as in References [15
]. Air sensors permit obtaining, in real time, all information regarding carbon dioxide, temperature and humidity levels (Figure 1
). This may be very useful for handling, without pause, the physical conditions with which the workers may be confronted.
Certainly, the lopsided oxygen and carbon dioxide levels can have a negative impact on personal health, ability to concentrate and make right decision, and overall safety [25
and oxygen levels should always be examined in all closed environments in which many people stay for a longer period of time; otherwise, it becomes a life threatening condition. The sensor outputs deal with diverse values for CO2
, oxygen, air temperature and air humidity levels. When the thresholds are triggered, an alarm sounds and/or other connected safety systems are alerted.
2.2.2. Ionizing Radiation Sensor
This is considered the most important sensor and can be very beneficial in closed environments. Essentially, everybody is conscious of the high degree of risk related to atomic radiation. Ionizing radiation—like Gamma rays and both Alpha and Beta particles transmitted from radioactive material and nuclear reactions—may generate chemical or physical problems when it transmits energy to the human skin [27
]. Therefore, an ionizing radiation sensor (IRS) can be a good solution to prevent this danger in the work environment. This sensor, depicted in Figure 2
, offers the possibility to detect the diverse rays and particles with silicon PIN photodiodes either through straightforward immersion in the crystal lattice or through the quantification of luminescence radiation of a scintillation crystal.
In the IRS sensor, dark current signals and capacitance levels are very low and the fully depleted space-charge regions have minimal series resistance, which guarantees maximum absorption of radiation.
2.2.3. Sound/Noise Level Sensors
Many workplaces need to measure and investigate sound and noise levels [29
] in order to prevent any unexpected risk in the environment that can affect the natural workflow. This can be achieved by using an easily operated sensor called a Sound Level Sensor (SLS) (see Figure 3
), which easily measures sound level, in decibels (dB), in a variety of given settings. No constraints are imposed on the use of SLS sensors; they react to sound loudness as the human ear reacts. They measure sound level to within 3 dB along a specific band from 55 to 110 dB; the frequency ranges from 30 Hz to 10 KHz. This sensor can be a beneficial tool for high schools or colleges to monitor classroom sound levels, detect resonation time, analyze sound insulation, and, thus, provide a good environment for education.
2.3. Indirect-Contact Sensors
The following category examines the sensors that target humans without any direct contact. This class of sensor has confirmed its role in the ubiquitous sensing domain and can play a vital role in the protection and well-being of people.
2.3.1. Eye Motion Sensor
The sleepiness of the employees working in critical workplaces can cause grave risks. Infrared (IR) sensors can be an efficient tool to resolve this problem and increase the performance and security of workers in closed spaces. Recently, this technology has been investigated in References [31
] to avoid unexpected vehicles accidents stemmed from lethargy, drugs, or any narcotic substances. The algorithm in Reference [31
] describes the eye movement detection in detail.
In space, the position and orientation parameters identifying the user’s eye must be computed. This data can only be extracted from the camera’s image using a precise eye model that takes into account the complex eye form and motion features. In Reference [32
], the authors studied the possibility of developing an eye model and integrating with eye-tracking systems. Its consistency depends on the precision of object recognition throughout the image processing, as shown in Figure 4
. To obtain a consistent calculation of the model parameters, efficient algorithms were developed in References [35
Generally, an infrared sensor—a device composed of IR LEDs and IR photodiode—detects infrared radiation shed. It has a positive response in terms of relative sensibility [37
] that sets it apart from other detectors such as UV sensitive, standard camera and human eye responsive (Figure 5
). IR light is emitted into our eyes using the IR emitter and reflected rays are received from the eye through an infrared receiver [38
], depicted in Figure 6
. The receiver signal remains high for unmoving or unwinking eyes and diminishes in other scenarios. This is a straightforward way to determine the eye status so that the alarm may be activated if the signal becomes low.
To detect the number of eye blinking, the IR sensor must be positioned near the eye in the transmitter region. The microcontroller digitizes the received analog data from this sensor and decides whether the alert must be activated. This system can be configured and set up in closed workplaces requiring high levels of vigilance to protect employees from sleepiness.
2.3.2. Heartbeat Sensor
The proposed technology, a heartbeat sensor, is used to detect the pulses or heart beats of a person’s heart. It is considered a powerful tool to monitor employee health. It can improve at astonishing rates the security of workers. The Kinect sensor, depicted in Figure 7
, offers the possibility to distantly sense and monitor a human heart rate from approximately one meter away [39
]. By identifying the blood speed, the camera, recognizing how fast the blood is pumping, can compute the beat rate of the heart.
2.3.3. Body Temperature Sensor
Many infections and diseases can originate from severe variations in body temperature. Similarly, the pathway of some diseases can be controlled through the measuring of a person’s temperature and by evaluating the efficacy of the treatment prescribed by the doctor. Measuring body temperature is considered a key feature of clinical research and medical studies. Recently, due to technological advancements, many studies have used various measurement approaches [40
]. The signals received can characterize the process of continuous physiological control can anticipate and treat many cardiovascular, neurological and pulmonary problems in their infancy.
Therefore, measuring body temperature is considered of paramount medical importance and can be achieved by the IR temperature sensor shown in Figure 8
Indeed, it measures the average temperature of the area in the field of view (FOV), which changes according to the distance from the sensor to the target person. This sensor can be used over longer distances but the influence of atmospheric conditions like dust, humidity and noise can be important.
Evidently, IR waves are too small to be seen by the human eye. They are situated between visible light and microwaves, in terms of wavelength (λ), and characterized by the following:
Electromagnetic spectrum = [0.75 μm, 1000 μm]
Near IR if λ [0.75 μm, 3 μm]
Mid-IR if λ [3 μm, 6 μm]
Far-IR if λ > 6 μm
This sensor does not emit the mentioned IR signals itself. Rather, it passively detects the IR radiations hailing from the human bodies in the surrounding space (see Figure 8
b). For that reason, it can be called a passive IR (PIR) sensor.
The PIR sensor band is up to 10 m at an angle of +15° or −15°. The detected radiation is transformed into electrical charges, which are proportional to the detected radiation level. These charges are enhanced using a field-effect transistor (FET) design and loaded to the output pin of the device, as depicted in Figure 8
c. This is then applied to an external circuit for further triggering and amplification of the alarm stages.
The proposed sensor technology is characterized by a fast response, broad band, and large dynamic range, taking into account the atmospheric conditions. It offers the possibility to protect employees working in closed environments from any unexpected disease.
2.4. Direct-Contact Sensors
Direct-Contact sensors work through direct contact with the human-body. These sensors are able to detect physical variables such as heat, strain, and pressure. We concentrate in this section on the direct-contact sensors that can be used in closed workplaces and which play a significant role in employee safety.
2.4.1. Human-Motion Sensors
The human-motion sensor was introduced to detect the human movement and can play a crucial role in both safety and security purposes, mainly in closed workplaces. Recently, many scientific types of research have been presented based on this technology its applications [42
]. Indeed, a novel approach identifying the daily motions of humans using wearable motion sensor systems is presented in Reference [43
In Reference [42
], the authors present a sensor-based architecture called SensX, permitting the monitoring and investigation of the multi-dimensional action steps. The working of this technology is described in Figure 9
. IR can evidently be detected by this sensor although it cannot be seen by individuals due to the fact that infrared wavelengths are greater than those of visible light.
In its field of detection, the motion sensor captures the variations in IR radiation coming from human temperature. Novel flexible and stretchable physical sensors (see Figure 10
) offer the possibility to monitor human-activity and identify personal healthcare by detecting pressure, strain, and heat parameters [44
]. This technology can be mixed with motion sensors to regularly collect all convenient data of the human body and subsequently enhance the quality of the obtained results.
Therefore, it is very useful to install an efficient system based on motion sensors in closed environments so as to control the psychological and/or physiological status of employees throughout the human actions analysis.
2.4.2. Smart-Health Sensors
Generally speaking, this class of sensor is flexible and attached directly to the human body without any loss of movement freedom. The smart health sensor focuses on the sweat secreted by the skin to measure metabolite and electrolyte levels in the sweat [9
]. More precisely, these measures mainly concern skin temperature, in addition to the glucose, lactate, sodium, and potassium in sweat. The resulting measures are analyzed and sent wirelessly to a phone or other type of device (see Figure 11
). These data can be monitored by a supervisor. An alert can be triggered if the gathered data indicate health issues like high temperature, dehydration and general fatigue.
2.4.3. Fingerprint Sensors
Fingerprint sensors were popularized and adopted for smartphones as well as companies [48
]. This sensor, depicted in Figure 12
, recognizes the fingerprint of an employee and sends information to the supervisor, such as the action time. In order to ensure the security of a specific closed area, such as guards control rooms, it is possible to frequently request a fingerprint validation to make sure that the employees are in a suitable state to fulfill their task and are neither sleepy nor outside their offices.