Smart Glasses-Based Personnel Proximity Warning System for Improving Pedestrian Safety in Construction and Mining Sites

A smart glasses-based wearable personnel proximity warning system (PWS) was developed for pedestrian safety in construction and mining sites. The smart glasses receive signals transmitted by Bluetooth beacons attached to heavy equipment or vehicles, with the proximity determined by the signal strength. A visual alert is displayed to the wearer when in close proximity. The media access control address of the Bluetooth beacon provides information on the approaching equipment or vehicle, which is displayed to the wearer so that they can respond appropriately. There was a detection distance of at least 10 m regardless of the direction the pedestrian was looking and the alert was successful in all 40 trials at ≥10 meters. The subjective workload was evaluated using the NASA task load index on ten subjects, either without a personal PWS, with a smartphone-based PWS, or with the smart glasses-based PWS. The mental, temporal, and physical stresses were lowest when using the smart glasses-based PWS. Smart glasses-based PWSs can improve work efficiency by freeing both hands of the pedestrians, and various functions can be supported through application development. Therefore, they are particularly useful for pedestrian safety in construction and mining sites.


Introduction
Collisions between heavy construction equipment and workers occur frequently in construction and mining sites around the world. According to the U.S. Bureau of Labor Statistics, 27% of the 1083 casualties in construction and mining in the USA in 2017 were related to collisions with equipment [1]. Eurostat reported 31 fatalities related to collisions with equipment at construction and mining sites in Europe between 2014 and 2018 [2]. In Korea, there were 121 fatalities due to collisions with the power machinery used at mining and construction sites in 2018 [3]. To prevent such collisions, proximity warning systems (PWSs) should be implemented at construction and mining sites [4].
PWSs provide proximity alarms to warn workers about approaching equipment or pedestrians [5]. They employ various wireless communication technologies, including electromagnetic fields (e.g., HazardAvert, Strata Worldwide, Sandy Springs, GA, USA [6]), global positioning system (GPS) (e.g., MineAlert, Modular Mining, Tucson, AZ, USA [7]), radio-frequency identification (RFID) (e.g., GE Collision Awareness System, GE Mining, Brisbane, Queensland, Australia [8]), and Wi-Fi (e.g., Proximity Detection & Collision Avoidance, Newtrax, Montreal, Quebec, Canada [9]). Electromagnetic PWSs comprise transmitters that generate electromagnetic signals using a wire-loop antenna and receivers that detect the signals when the two components are in close proximity [10,11]. GPS PWSs use GPS devices that are attached to equipment and workers to determine their locations, and provide alerts when collisions are likely [12]. They operate best in open-pit mine environments. RFID PWSs are the Table 1. Specifications of commercially available smart glasses [39]. In this study, we developed a smart glasses-based personal PWS for pedestrian safety in construction and mining sites. By attaching Bluetooth beacons to heavy equipment or vehicles, which transmit signals that are received by the smart glasses, the proximity of the vehicle can be determined based on the signal strength. The pedestrian receives a visual alert when in close proximity to the vehicle. We tested the developed personal PWS at real mining sites to determine its performance, and analyzed the subjective workload felt by pedestrians using the NASA task load index (NASA-TLX) [40].

Development of Personal PWS Using Smart Glasses
The design of the personal PWS is summarized in Figure 1. Smart glasses worn by pedestrians receive BLE signals transmitted from Bluetooth beacons and provide visual alarms when in close proximity to a beacon. The Bluetooth beacons can be attached to heavy equipment, vehicles, or hazardous areas in construction or mining sites, and they continuously transmit BLE signals. Hence, as well as alerting the wearer to approaching heavy equipment or vehicles, the smart glasses can alert the wearer to the fact that they are approaching a hazardous area. The visual proximity alerts are received through the smart glasses while performing on-site work such as safety checks and maintenance, allowing workers to quickly detect and respond to dangerous situations. The design of the personal PWS is summarized in Figure 1. Smart glasses worn by pedestrians receive BLE signals transmitted from Bluetooth beacons and provide visual alarms when in close proximity to a beacon. The Bluetooth beacons can be attached to heavy equipment, vehicles, or hazardous areas in construction or mining sites, and they continuously transmit BLE signals. Hence, as well as alerting the wearer to approaching heavy equipment or vehicles, the smart glasses can alert the wearer to the fact that they are approaching a hazardous area. The visual proximity alerts are received through the smart glasses while performing on-site work such as safety checks and maintenance, allowing workers to quickly detect and respond to dangerous situations. BLE technology can provide point-to-point, broadcast, or mesh wireless communication [41]. In point-to-point communication, data are exchanged by connecting the "master" device (which transmits data) and the "slave" device (which receives data) in a 1:1 configuration. In broadcast communication, several observers are able to receive the ID information transmitted periodically by a single broadcaster. In mesh communication, numerous master and slave devices are connected to exchange data.
Bluetooth beacons periodically transmit BLE signals with information on the universally unique identifier of the beacon, the media access control (MAC) address, and more. Therefore, the Bluetooth beacons act as broadcasters, while computers, smartphones, and smart glasses that can receive BLE signals are used as observers. The reception strength of the BLE signal can be quantified by the received signal strength index (RSSI), expressed as a number between −99 and −35 dBm. The propagation distance of the BLE signal can vary depending on the signal transmission strength and the signal propagation direction of the Bluetooth beacon. As the BLE signal transmission strength increases, the signal propagation distance increases. While the signal propagation direction can be set to be nondirectional to propagate the signal uniformly in all directions, this limits the propagation distance. If the signal is set to be directional, the BLE signal propagates relatively far ahead of the Bluetooth beacon. The changes in RSSI according to the BLE signal transmission strength and the propagation direction of the Bluetooth beacon have been analyzed previously [42].
RECO beacons (Perples, Seoul, South Korea) were used as the BLE transmission unit. RECO beacons are certified by agencies in Korea, the USA, Europe, Japan, etc., and they are in line with global beacon standards. The main specifications of the RECO beacons are listed in Table 2. Figure 2 shows an example of heavy equipment at construction and mining sites with the RECO beacons attached. Two Bluetooth beacons were installed on the front, rear, and sides of the heavy equipment, giving it a total of eight beacons. The Bluetooth beacons were set to transmit directional signals, so that the signals would propagate farther. The signal transmission strength and period of the beacons BLE technology can provide point-to-point, broadcast, or mesh wireless communication [41]. In point-to-point communication, data are exchanged by connecting the "master" device (which transmits data) and the "slave" device (which receives data) in a 1:1 configuration. In broadcast communication, several observers are able to receive the ID information transmitted periodically by a single broadcaster. In mesh communication, numerous master and slave devices are connected to exchange data.
Bluetooth beacons periodically transmit BLE signals with information on the universally unique identifier of the beacon, the media access control (MAC) address, and more. Therefore, the Bluetooth beacons act as broadcasters, while computers, smartphones, and smart glasses that can receive BLE signals are used as observers. The reception strength of the BLE signal can be quantified by the received signal strength index (RSSI), expressed as a number between −99 and −35 dBm. The propagation distance of the BLE signal can vary depending on the signal transmission strength and the signal propagation direction of the Bluetooth beacon. As the BLE signal transmission strength increases, the signal propagation distance increases. While the signal propagation direction can be set to be nondirectional to propagate the signal uniformly in all directions, this limits the propagation distance. If the signal is set to be directional, the BLE signal propagates relatively far ahead of the Bluetooth beacon. The changes in RSSI according to the BLE signal transmission strength and the propagation direction of the Bluetooth beacon have been analyzed previously [42].
RECO beacons (Perples, Seoul, South Korea) were used as the BLE transmission unit. RECO beacons are certified by agencies in Korea, the USA, Europe, Japan, etc., and they are in line with global beacon standards. The main specifications of the RECO beacons are listed in Table 2. Figure 2 shows 4 of 15 an example of heavy equipment at construction and mining sites with the RECO beacons attached. Two Bluetooth beacons were installed on the front, rear, and sides of the heavy equipment, giving it a total of eight beacons. The Bluetooth beacons were set to transmit directional signals, so that the signals would propagate farther. The signal transmission strength and period of the beacons were set to 4 dBm and 10 ms, respectively.    Smart glasses typically use optical see-through display technology to overlap the real-world and virtual environments. The virtual image is projected through the microdisplay into the lens and reflects light through a flat or a curved mirror to an optical combiner located in front of the human eye. The optical combiner superimposes the virtual image on top of the real-world view and delivers it to the user's eyes. In this process, waveguides are used to guide the light to effectively propagate it to the human eye without any losses. Optical fibers or prisms are often used as waveguides. Waveguides can be classified into reflective-, polarized-, diffractive-, and holographic-type, depending on the arrangement and the reflector material ( Figure 3). Reflective waveguides reflect and transmit virtual images using a semi-reflective mirror, while polarized waveguides reflect light through a series of polarized reflectors to deliver the image to the human eye. Diffractive waveguides reflect light through in-and out-coupling using a diffractive optical element with surface relief grating structures. Holographic waveguides are similar in principle to diffractive waveguides, where a holographic optical element is used to separate and transmit the red, green, and blue components of light. More information on waveguides can be found in Erdenebat et al. [43]. In this study, Moverio BT-350 smart glasses (Epson, Suwa, Japan) were used as the wearable personal PWSs for pedestrians. These smart glasses consist of a headset and a controller, and provide a stereoscopic view to the user by a reflective waveguide method [44]. Figure 4a shows the appearance of the equipment, and Table 3 lists the specifications of the Moverio BT-350 model. This device uses the Android 5.1 operating system and is equipped with various instruments including GPS, a gyroscope, accelerometer, and geomagnetic sensors. In addition, Wi-Fi communication and    In this study, Moverio BT-350 smart glasses (Epson, Suwa, Japan) were used as the wearable personal PWSs for pedestrians. These smart glasses consist of a headset and a controller, and provide a stereoscopic view to the user by a reflective waveguide method [44]. Figure 4a shows the appearance of the equipment, and Table 3 lists the specifications of the Moverio BT-350 model. This device uses the Android 5.1 operating system and is equipped with various instruments including GPS, a gyroscope, accelerometer, and geomagnetic sensors. In addition, Wi-Fi communication and Bluetooth 4.1 (Bluetooth Smart Ready Class 2) communication are possible.
To control the personal PWS, we developed an application for the smart glasses for Android operating systems using the MIT App Inventor application [45]. Figure 4b shows the user interface of the developed application. The application visually alerts the pedestrian when the RSSI value of the BLE signal received by the smart glasses reaches a certain level. In addition, based on the MAC address of the Bluetooth beacon that transmits the BLE signal, information on the approaching device or the vehicle is also provided so that the pedestrians can respond effectively when in close proximity.
personal PWSs for pedestrians. These smart glasses consist of a headset and a controller, and provide a stereoscopic view to the user by a reflective waveguide method [44]. Figure 4a shows the appearance of the equipment, and Table 3 lists the specifications of the Moverio BT-350 model. This device uses the Android 5.1 operating system and is equipped with various instruments including GPS, a gyroscope, accelerometer, and geomagnetic sensors. In addition, Wi-Fi communication and

Performance Assessment of Smart Glasses-Based Personal PWS
To evaluate the performance of the developed personal PWS, field experiments were conducted in the Ilgwang mining site (35 • 18'33" N, 129 • 13'35" E) located in Gijang-gun, Busan, Korea. As shown in Figure 2, a total of eight Bluetooth beacons were attached to the excavator, and pedestrians wore smart glasses installed with the personal PWS application. The controller of the smart glasses with a built-in BLE signal receiving antenna was positioned to face the front of the pedestrian. The excavator approached a pedestrian standing in the center of a mine hauling route from 40 m away at a speed of 10-20 km/h, as shown in Figure 5. We then measured the detection distance at which the personal PWS receiving the BLE signal began alerting the pedestrian. To determine the detection distance according to the direction of the receiver, the facing angles between the pedestrian and the excavator were set at 45 • intervals from 0 • to 315 • , and the detection distance was measured five times for each of the eight angular conditions (a total of 40 measurements). The minimum safety distance between the pedestrian and excavator was set to 10 m, and it was determined whether proximity alerts were provided to pedestrians before the excavator approached within a safe distance. of 10-20 km/h, as shown in Figure 5. We then measured the detection distance at which the personal PWS receiving the BLE signal began alerting the pedestrian. To determine the detection distance according to the direction of the receiver, the facing angles between the pedestrian and the excavator were set at 45° intervals from 0° to 315°, and the detection distance was measured five times for each of the eight angular conditions (a total of 40 measurements). The minimum safety distance between the pedestrian and excavator was set to 10 m, and it was determined whether proximity alerts were provided to pedestrians before the excavator approached within a safe distance.

Subjective Workload Assessment of Smart Glasses-Based Personal PWS
Workload is a quantitative measure of the amount of mental stress a person feels while performing tasks within a particular system [46]. Workload is affected by factors of psychological (focus on work, anxiety), physical (physical difficulties, difficulty in controlling machines), temporal (deadlines), and environmental (noise, relationships with colleagues) nature [47]. If there are many negative factors in the system, over-or under-load can occur, which can reduce work efficiency.

Subjective Workload Assessment of Smart Glasses-Based Personal PWS
Workload is a quantitative measure of the amount of mental stress a person feels while performing tasks within a particular system [46]. Workload is affected by factors of psychological (focus on work, anxiety), physical (physical difficulties, difficulty in controlling machines), temporal (deadlines), and environmental (noise, relationships with colleagues) nature [47]. If there are many negative factors in the system, over-or under-load can occur, which can reduce work efficiency. Therefore, it is necessary to design and operate a system with minimal workload to improve the work efficiency.
Subjective workload evaluation can examine workload through a questionnaire. It is frequently used in human-machine system development [48]. Representative subjective workload evaluation methods include NASA-TLX [40], the subjective workload assessment technique [49], and the workload profile technique [50]. In this study, the subjective workload was evaluated using the NASA-TLX method to evaluate the psychological, physical, and temporal effects of pedestrians wearing smart glasses and using the personal PWS while working on a mining site. NASA-TLX is a multi-dimensional rating procedure that estimates the overall workload score based on a weighted average of six factors [51]: mental demand, physical demand, temporal demand, own performance, effort, and frustration level. These workload parameters are defined as follows:

•
Mental demand: How much mental and cognitive skills were required to perform this task? • Physical demand: How much physical ability was required to perform this task? • Temporal demand: How much time pressure did you feel due to the rate or pace at which you performed multiple tasks? • Own performance: How successfully do you think you have achieved the goal of this task? • Effort: How much mental and physical efforts were required to achieve your work aims? • Frustration level: How many uncomfortable feelings (stress, anger) did you feel while performing this task?
The degree to which the subject felt the six workload parameters is rated by the subject. All parameters except "own performance" (which is scored from Excellent to Poor) are scored from Low to High by a value between 0 and 100 (in increments of 5). Next, the weights of the six parameters are calculated using pairwise comparison, and the overall workload score is calculated by averaging the product of the scores and weight of each factor.
Subjective workload evaluation was performed on ten subjects aged 22-27 years (mean age 23.6 years) in the same location where the performance of personal PWS was evaluated. Most of the subjects (70%) said they had knowledge of smart glasses, and 20% said they had used smart glasses before. Three equivalent experiments were performed with the same experimental conditions to compare the effect on subjective workload, where the subjects (1) did not use a personal PWS, (2) used a smartphone-based personal PWS, and (3) used the smart glasses-based personal PWS. The smartphone-based personal PWS application was developed in this study and was installed on a Samsung Galaxy S9 (Samsung Electronics, Suwon, Korea) that supports Bluetooth 5.1.
In all experiments, the subject stood in the center of the hauling road to investigate the road maintenance status, while an excavator or general vehicle approached the subject. A proximity alert along with an image of the equipment was made visible to the subject using the smartphone or smart glasses. The subject was informed to move to an evacuation area outside the nearby transport road if the type of equipment was an excavator, while if the type of equipment was a normal vehicle, the work was paused and resumed once the vehicle had passed through.
The ten subjects each performed all three experiments (1)-(3) in a random order, and after the experiment, the workload for the experiment was examined according to the NASA-TLX procedure. Figure 6a shows the smart glasses when no BLE signal was detected, and Figure 6b shows the visual alert screen when the BLE signal was recognized. Depending on the MAC address of the Bluetooth beacon that sent the BLE signal, a picture of the excavator already stored in the personal PWS application appears on the alert screen. Therefore, pedestrians wearing smart glasses can not only receive an alarm about the approaching heavy equipment, but they can also be warned about the type of heavy equipment.  Table 4 shows the main statistics of the detection distance measurement when the proximity alert is provided, depending on the facing angle between the pedestrian and the excavator, and Figure 7 is a radial representation of the average detection distance per angle. The average detection distance was more than 30 m when the facing angle was 0°, 45°, 90°, or 315°; more than 20 m when the angle was 270°; and 15-20 m when the angle was 135°, 180°, or 225°. The relatively short distance at the latter angles is likely to be because the BLE signal receiving antenna of the smart glasses and the Bluetooth beacon attached to the excavator are facing opposite directions. Nevertheless, the field experiments show that the developed personal PWS has a detection distance of at least 10 m regardless of the direction in which the pedestrian is looking. In all 40 experiments, alerts were always issued when the excavator approached the pedestrian within the minimum safety distance of 10 m.  Table 4 shows the main statistics of the detection distance measurement when the proximity alert is provided, depending on the facing angle between the pedestrian and the excavator, and Figure 7 is a radial representation of the average detection distance per angle. The average detection distance was more than 30 m when the facing angle was 0 • , 45 • , 90 • , or 315 • ; more than 20 m when the angle was 270 • ; and 15-20 m when the angle was 135 • , 180 • , or 225 • . The relatively short distance at the latter angles is likely to be because the BLE signal receiving antenna of the smart glasses and the Bluetooth beacon attached to the excavator are facing opposite directions. Nevertheless, the field experiments show that the developed personal PWS has a detection distance of at least 10 m regardless of the direction in which the pedestrian is looking. In all 40 experiments, alerts were always issued when the excavator approached the pedestrian within the minimum safety distance of 10 m.   Figure 8 shows the scores of the six workload parameters evaluated in three experiments on ten subjects. When the pedestrians did not use a personal PWS, the mental demands, temporal demands, and mental stress were high. This could be because the pedestrians needed to constantly check by eye for approaching vehicles and equipment while working. When the pedestrians used smartphonebased personal PWSs, they had to check their smartphone screen repeatedly to check for approaching excavators or vehicles, and because of increased eye movements, they felt negative emotions. In addition, since the pedestrians had to work with a smartphone in their hands, they did not have both hands free, so more effort was required to achieve their work task. When the pedestrians wore smart glasses-based personal PWSs, most of the workload factors, except for the frustration level, were low. The reason for the high level of frustration is that the subjects felt uncomfortable because they were not accustomed to wearing the smart glasses (the glasses sliding, wearing regular glasses under the smart glasses).  Figure 8 shows the scores of the six workload parameters evaluated in three experiments on ten subjects. When the pedestrians did not use a personal PWS, the mental demands, temporal demands, and mental stress were high. This could be because the pedestrians needed to constantly check by eye for approaching vehicles and equipment while working. When the pedestrians used smartphone-based personal PWSs, they had to check their smartphone screen repeatedly to check for approaching excavators or vehicles, and because of increased eye movements, they felt negative emotions. In addition, since the pedestrians had to work with a smartphone in their hands, they did not have both hands free, so more effort was required to achieve their work task. When the pedestrians wore smart glasses-based personal PWSs, most of the workload factors, except for the frustration level, were low. The reason for the high level of frustration is that the subjects felt uncomfortable because they were not accustomed to wearing the smart glasses (the glasses sliding, wearing regular glasses under the smart glasses). and mental stress were high. This could be because the pedestrians needed to constantly check by eye for approaching vehicles and equipment while working. When the pedestrians used smartphonebased personal PWSs, they had to check their smartphone screen repeatedly to check for approaching excavators or vehicles, and because of increased eye movements, they felt negative emotions. In addition, since the pedestrians had to work with a smartphone in their hands, they did not have both hands free, so more effort was required to achieve their work task. When the pedestrians wore smart glasses-based personal PWSs, most of the workload factors, except for the frustration level, were low. The reason for the high level of frustration is that the subjects felt uncomfortable because they were not accustomed to wearing the smart glasses (the glasses sliding, wearing regular glasses under the smart glasses).   Figure 9 shows the calculated overall workload scores from the three experiments. The overall workload score was about 72.4 for pedestrians not using a personal PWS, about 57.2 for those using a smartphone-based personal PWS, and about 28.9 for those wearing a smart glasses-based personal PWS. The mental, temporal, and psychological stresses of pedestrians were lower when wearing a smart glasses-based personal PWS than when not using a PWS or when using a smartphone-based PWS. Smart glasses free up both hands of pedestrians while effectively providing proximity alerts about equipment or vehicles, helping pedestrians to increase their work efficiency. However, some subjects felt uncomfortable with the fit of the smart glasses; hence, further study is necessary to improve the smart glasses wearability.

Results
Int. J. Environ. Res. Public Health 2020, 17, x 10 of 15 Figure 9 shows the calculated overall workload scores from the three experiments. The overall workload score was about 72.4 for pedestrians not using a personal PWS, about 57.2 for those using a smartphone-based personal PWS, and about 28.9 for those wearing a smart glasses-based personal PWS. The mental, temporal, and psychological stresses of pedestrians were lower when wearing a smart glasses-based personal PWS than when not using a PWS or when using a smartphone-based PWS. Smart glasses free up both hands of pedestrians while effectively providing proximity alerts about equipment or vehicles, helping pedestrians to increase their work efficiency. However, some subjects felt uncomfortable with the fit of the smart glasses; hence, further study is necessary to improve the smart glasses wearability.

Utilization of Smart Glasses-Based Personal PWS in Underground Tunnel
Collisions between equipment and pedestrians occur frequently in underground mines and tunnels. To verify that the smart glasses-based personal PWS could be used underground, a simple field experiment was additionally conducted in the Yeonhwa tunnel (35°12'55"N, 129°13'2"E) located in Gijang-gun, Busan, Korea.
Eight Bluetooth beacons were attached to a typical vehicle, with the Bluetooth beacons set to send directional signals. The smart glasses application was also updated by uploading the Bluetooth

Utilization of Smart Glasses-Based Personal PWS in Underground Tunnel
Collisions between equipment and pedestrians occur frequently in underground mines and tunnels. To verify that the smart glasses-based personal PWS could be used underground, a simple field experiment was additionally conducted in the Yeonhwa tunnel (35 • 12'55"N, 129 • 13'2"E) located in Gijang-gun, Busan, Korea.
Eight Bluetooth beacons were attached to a typical vehicle, with the Bluetooth beacons set to send directional signals. The smart glasses application was also updated by uploading the Bluetooth beacon's MAC address and the photo of the vehicle to the database. Pedestrians waited inside the tunnel with smart glasses on, and the vehicle ran from the entrance to the end of the tunnel. Figure 10 shows the results of the field experiment inside the tunnel. When the vehicle was not inside the tunnel, no alert was displayed on the smart glasses (see Figure 10a). When the vehicle entered the tunnel and approached the worker, visual proximity alerts and a photo of the vehicle were visualized on the smart glasses (see Figure 10b). The experiment confirmed that the personal PWS could be used underground.

Advantages and Disadvantages of Smart Glasses-Based Personal PWS
There are two main problems with existing PWSs that provide workers with proximity alerts in visual (LED), audible, or vibration forms. First, PWSs are mainly intended for use on construction and mining sites, where vibration and noise are constantly generated by the operation of machines and equipment and frequent blasting work. This makes it difficult for the workers to easily recognize sound and vibration alerts. In addition, if earphones are worn to listen to the sound notifications, sound from the surrounding area is cut off, heightening the risk of various accidents. Second, these PWSs provide constant proximity alarms regardless of the type of equipment or vehicle approaching the operator. Instead, different types of proximity alarms should be provided for different types of equipment, to allow workers to quickly recognize the type of equipment approaching and quickly determine which action to take.
Personal PWSs using smart glasses can solve the problem of existing PWSs by providing workers with visual proximity alarms. Regardless of vibration and noise generated at construction and mining sites, workers can receive visual alerts through smart glasses, allowing the alert to be quickly identified, while hearing the surrounding sounds and vibrations. This helps pedestrian workers to respond quickly to dangerous situations. Smart glasses applications provide not only proximity alerts, but also information about the equipment and vehicles that are approaching pedestrians, allowing them to respond effectively by evacuating or pausing work. In addition, as shown in the subjective workload evaluation results, it has the advantage of not hindering work efficiency and concentration while improving the safety of the pedestrians.
Nevertheless, the smart glasses-based personal PWS has a few drawbacks. Smart glasses can cause discomfort to the wearer when they are worn over regular glasses or if they slip. Moreover, it would be difficult to don these smart glasses over industrial goggles and soundproof headsets, which are typically worn by workers for visual and hearing protection in mining and construction sites. Furthermore, workers may find it difficult to manipulate the smart glasses via the touchpad controller during work. To supplement this problem, the smart glasses should be controlled by motions, gestures, and voice recognition using a camera, microphone, and multi-axis sensors.

Advantages and Disadvantages of Smart Glasses-Based Personal PWS
There are two main problems with existing PWSs that provide workers with proximity alerts in visual (LED), audible, or vibration forms. First, PWSs are mainly intended for use on construction and mining sites, where vibration and noise are constantly generated by the operation of machines and equipment and frequent blasting work. This makes it difficult for the workers to easily recognize sound and vibration alerts. In addition, if earphones are worn to listen to the sound notifications, sound from the surrounding area is cut off, heightening the risk of various accidents. Second, these PWSs provide constant proximity alarms regardless of the type of equipment or vehicle approaching the operator. Instead, different types of proximity alarms should be provided for different types of equipment, to allow workers to quickly recognize the type of equipment approaching and quickly determine which action to take.
Personal PWSs using smart glasses can solve the problem of existing PWSs by providing workers with visual proximity alarms. Regardless of vibration and noise generated at construction and mining sites, workers can receive visual alerts through smart glasses, allowing the alert to be quickly identified, while hearing the surrounding sounds and vibrations. This helps pedestrian workers to respond quickly to dangerous situations. Smart glasses applications provide not only proximity alerts, but also information about the equipment and vehicles that are approaching pedestrians, allowing them to respond effectively by evacuating or pausing work. In addition, as shown in the subjective workload evaluation results, it has the advantage of not hindering work efficiency and concentration while improving the safety of the pedestrians.
Nevertheless, the smart glasses-based personal PWS has a few drawbacks. Smart glasses can cause discomfort to the wearer when they are worn over regular glasses or if they slip. Moreover, it would be difficult to don these smart glasses over industrial goggles and soundproof headsets, which are typically worn by workers for visual and hearing protection in mining and construction sites. Furthermore, workers may find it difficult to manipulate the smart glasses via the touchpad controller during work. To supplement this problem, the smart glasses should be controlled by motions, gestures, and voice recognition using a camera, microphone, and multi-axis sensors.

Comparision of Smart Glasses-Based Personal PWS with the Existing System
To compare the operating performances of the smart glasses-based personal PWS and the existing PWS, the recall for the warning alerts was estimated by investigating the number of warning alerts based on alert type. The smartphone-based PWS, which was developed by Baek and Choi [23], was used in the comparison of operating performance. The field experiment results of two PWSs were used to estimate the recall for the warning alerts. The conditions of this experiment were identical to those for the Bluetooth beacon model (RECO), i.e., the signal transmission strength was 4 dBm and the speed of approaching equipment was 10 m/s. The warning zone and the warning buffer zone were set in the ranges of 0-10 m and 10-20 m between the pedestrian and the equipment, respectively. The facing angles between the pedestrian and the equipment were fixed at 0 • during the 50 trials of the existing PWS and were set at 45 • intervals from 0 • to 315 • during the 40 trials of the proposed personal PWS. The type of warning alert and the recall are defined as follows: • True positive: The alert was activated before the equipment approached the warning zone.

•
False negative: The alert was not activated even after the equipment entered the warning zone.

•
Recall: Ratio of true positives to the sum of true positives and false negatives. Table 5 lists the recall for the warning alerts of the two PWSs. In both PWSs, true positive alarms were issued for all cases, and false negatives were not observed. Moreover, both PWSs had a recall of 100%. Thus, there was no difference in the performance of the false alarms in alert activation. Therefore, it is expected that the smart glasses-based personal PWS can replace existing PWSs and also can be employed to ensure personal safety in construction and mining sites. Table 5. Comparison of warning alert recall between smart phone-based PWS [23] and smart glasses-based personal PWS (from this study).

Type of Warning Alert
Type of PWS

Conclusions
In this study, we developed a personal PWS that uses smart glasses to receive BLE signals from Bluetooth beacons and to provide visual proximity alerts to pedestrians. The performance assessment of the personal PWS at the mining site confirmed that the application provided a visual proximity alert along with a picture of the equipment that was approaching the pedestrian. The average BLE signal recognition distance of the smart glasses was about 37.4 m when the excavator approached from the front of the worker and about 19.4 m when the excavator approached from the rear side of the worker. The workload for the personal PWS on 10 subjects was quantitatively analyzed using the NASA-TLX criteria, which demonstrated that using smart glasses to provide visual proximity alerts led to lower mental efforts, and freed the worker's hands, thus maintaining work efficiency. The overall workload score calculated when using smart glasses was lower than that when using a smartphone-based PWS, suggesting that smart glasses are suitable as devices for implementing personal PWSs in construction and mining sites.
Personal PWSs using smart glasses have the following remarkable advantages in construction and mining sites. First, since smart glasses are worn on the face, workers can use both hands freely. Second, pedestrians can be provided directly with visual proximity alerts, allowing them to quickly determine dangerous situations and quickly evacuate. Finally, the proposed personal PWSs could be implemented and utilized at mining and construction sites by distributing multiple sets of smart glasses and Bluetooth beacons to working sites, regardless of the scale of these working sites.
In future work, the smart glasses-based personal PWS can be extended by developing new functions through Android application programming interface. For example, Wi-Fi communication enables data exchange with remote offices, and GPS sensors can easily recognize the worker's location. The application also recognizes the worker's motion as the glasses contain accelerometers, gyroscopes, and geomagnetic sensors, and they can identify gestures using a camera, making it easier for pedestrians to control the smart glasses at construction and mining sites.