CROWD SENSING USING SMARTPHONES Essay Example
- Category:Engineering and Construction
- Document type:Thesis
Crowd Sensing Using Smartphones
EXPLORING ADVANCEMENT IN CROWD SENSING THROUGH INFRARED BASED PEDESTRIANS COUNTER
KEYWORDS: Crowd Monitoring, Sensors, Data Analysis
The term “crowd sensing” refers to sharing data collected by sensing devices with the aim to measure phenomena of common interest. It is increasingly finding its application in transport and traffic. The recent technological advancement in crowd sensing has opened up new perspectives for cost-effective ways of managing the traffic congestion as well as safety in a critical situation such as evacuation. The project will explore the advancement in crowd sensing especially towards handling the passengers crowds at major transport systems. It will also provide an opportunity to work on data collection and analysis that involves crowd sensing through smartphones. Smartphones will be used in analyzing and recording the most intricate data to attain using simple technologies that almost all people, buildings, and automobiles among others have to provide efficient services to the customers. Crowdsensing is the process that will exploit the application of all technologies such as mobile phone microphones, and internet connections among others to attain unbiased data that can be used in making decisions rapidly and efficiently. The research will exploit and estimate the accuracy of infrared based pedestrian counters under low and high pedestrian flow condition while engaging in an experiment to support the information provided in the research.
Crowd sensing is a word that means sharing the data collected by sensing devices. The devices collect data automatically without the necessity of any human effort. Crowd sensing can be used majorly in Transport & Traffic and organizations by using the Smartphone sensing devices. The current technical development that exists in crowd sensing has exposed new perceptions for cost-effective methods to manage the traffic congestion in addition to the basic safety in a critical location like an evacuation. Crowdsensing through using smartphones can be a cheap and scalable way to implement static wireless sensor networks for Crowdsensing over a large coverage area.
Mobile sensing is envisioned in domains such as health monitoring, social networking and transporting. Mobile phones have developed to the central computer and communication devices. The project presents how crowd sensing through infrared technology is the advanced prototype for collecting data. The report shows how the mobile sensing applications can be exploited to collect data. The project will justify why mobile sensing is important in managing and collecting data in traffic and other locations.
Crowdsensing using the infrared
It is important that individuals have the access of information through technological devices such as infrared technology. Infrared technology has the capability of allowing people to perform different activities such as recording data regularly. Ruser, et al., (2006) state that infrared technology is used in various crowdsensing situations such as personal computers, and for environmental control systems (382). According to (Hashimoto, et al., 1997), the infrared sensors can be used to detect the motion of people whether coming in or out of an area and record the information accurately in the receiver. Thus, the infrared systems will be used and experimented upon to show their effectiveness in counting people among other benefits.
Thus, infrared technology can be utilized for crowdsensing during environmental control challenges among others. According to (Shelke, et al., 2014), infrared technology transmits signals through radiation for data collection, which has high accuracy levels even in a four-way terminal as it collects data from all directions. For instance, when there is a public issue that needs to be controlled such as an outbreak, infrared technologies can be applied to assist in managing the problem. Infrared can be used in in areas with a heavy flow of pedestrians and low flow, and yet provide undeviating recordings (Kahler & Arnberger, n.d.). Through implementing infrared devices in irregular areas, the users can record all forms of data. Infrared devices collect data from all directions (Hashimoto, et al., 1997). It is also not limited by the lack of internet among individuals or transport means, as it only needs motion mainly (Garcfa, et al., 2013). Thus, the infrared devices can be used to collect data from all areas of the country while efficiently and reliably informing people of the areas they should avoid and use. That is; as the infrared technologies can provide real-time information, the information can be used to make decisive decisions that require hastiness (Abuarafah, et al., 2012). Consequently, the environmental issue in question would be managed efficiently.
Active Infrared Sensors (Beam counters)
The beam counters consume a low amount of power in the batteries and are mainly appropriate for pedestrian counting though majorly developed for detecting pedestrians (Fanping, et al., 2007, 2). The infrared light beam counter includes a transmitter, receiver, and a data logger. The primary function of the transmitter is to emit an infrared beam where the receiver intercepts the appropriate position. The interception registers a count by the data logger. Thus, it accurately collects data and records all the people or vehicles among others in a selected location.
The infrared sensors according to (Ruser, et al., 2006) use light scanners to detect motion of objects or people. Mainly two arrays of receivers and emitters are paired together and mounted on opposite sides such as doorways allowing the motion to be detected. Thus, the infrared sensors can be used to determine the number of people passing through a doorway or entering a room (Ruser, et al., 2006). The information can then be used for analysis and trends recognition leading to improved decision-making processes. The infrared sensing devices can be used for several activities such as counting pedestrians and customers. They are reliable counting systems and can be used in several terms. For instance, they can be used as optical systems, radar, ultrasound, mechanical, and capacitive. The infrared sensors are mainly used for counting the people entering and leaving a room (Kerridge, et al., 2003). The statistics from the infrared sensors provide the number of people in a room accurately.
The sensor arrays of infrared have the N pairs, which are of directional IR emitters. As Ruser, et al present the pairs are shielded closely by sensitive phototransistors. The configuration of the two N=3 arrays is mounted vertically facing each other to record data accurately. The distance between the arrays can be big or low as well as the height. The two arrays emit pulses on an 8bit PIC microcontroller (Ruser, et al., 2006). The intensities of the light reflected each sensor in the array depend on the distance and direction of the light source. Once the IR is fixed on the entry of a room, it correctly assesses the motion and movement at the doorway, leading to accurate counts of the people that have used the door (Ruser, et al., 2006, 382). Thus, as (Ruser, et al., 2006) have presented, infrared sensors are the solutions to accurately counting the people within a particular door or area (383).
The automated infrared sensors can also be used to count the vehicles, and pedestrians in different areas (Fanping, et al., 2007, 1). Previous research has lacked to prove the effectiveness of these automated systems on counting the number of people or pedestrians in an area. According to (Shelke, et al., 2014) the IR transmitters and receivers emit infrared radiation (beam signal) as rays. The infrared sensors are linked to high reliability in counting, radiant intensity, and low forward voltage.
Passive Infrared Sensors
The passive sensors differ from the active beam counters in that they count people by tracking the heat emission of objects. That is; as active sensors detect motion, the passive sensors detect heat radiation for counting (Fanping, et al., 2007). The systems only work when the pedestrians are in proximity with the sensor. As Fanping (2007) presents, the sensor can then detect the heat emission of the pedestrians. The sensor mainly has a set heat threshold. Once an object surpasses the threshold, the sensor records a count. The main disadvantage of the sensor is that it cannot measure or count the pedestrians at a big distance. Additionally, when individuals are in a group the device cannot differentiate whether the heat is from one or many individuals. Thus, the count is one person, influencing the accuracy of the beam
According to (Shelke, et al., 2014), the infrared sensors are highly reliable and cost-effective. The devices also consume less power while detecting motion or heat, which is important, useful in crowd control, and counting (Hashimoto, et al., 1997). The infrared sensors are highly reliable for counting pedestrians and monitoring the environment while providing accurate data (Kerridge, et al., 2003).
The laser scanner exists as a high-resolution range finder. It emits the infrared laser pulses, which then detect the replicated pulses. The pulses detected represent a count of a person. However, the principle of measurement is reliant on the time-flight method which identifies the distance to the target is proportional to the interval time that exists between the reception and transmission of the pulses (Fanping, et al., 2007). The scanners have the possibility of rotating at a 360 degrees angle allowing it to regularly detect pulses within the set distance.
Fanping (2006) presents that there are two scanner types the vertical and horizontal scanners. The horizontal scanners can be used for detecting and counting the traffic and can accurately detect and count people that are within a 15m radius of the scanner. The vertical scanner cover a passage width of up to 26m while providing the direction counts and height. The accuracy of the scanners is high despite the challenge in signal processing.
Bluetooth Crowd Sensing
Vassilis Koskatos has evaluated how Bluetooth can be used as a sensing device to capture the trips of passengers on public transport buses. Bluetooth as the author presents can be used to obtain data concerning the trips that passengers make. Under normal circumstances, attaining this form of data is challenging mainly in transport buses where the alighting of travelers is not recorded, as onboard tickets do not show the information. Bluetooth allows for non-participatory, unannounced, and simultaneous tracking of many people in a large crowd. Thus, it provides accurate numbers of individuals in large events or gatherings such as that of the Ghent Festivities (Versichele, et al., 2012, 209).
Wi-Fi Via Crowd Sensing
Farshad and Marina have worked in this section and present that Wi-Fi can be used to leverage the smartphones and mobility of people in a crowd. After a test conducted in Edinburgh, the authors show that using the Wi-Fi Apps is a reliable tactic for doing mobile crowd sensing. It is easier to monitor site differences through Wi-Fi Fingerprinting (Verbee, et al., 2013, 1). Wi-Fi monitoring and fingerprinting are crowd-sensing applications that provide high accuracy and reliability while using the WLAN infrastructure (Verbee, et al., 2013, 1).
2.4 Infrared Sensors in Public Transport and Crowd
Infrared devices help in providing an estimation of the people in an area thus the crowd density (Kerridge, et al., 2003, 3). In the field of public transport, crowd-sensing data can be used to reduce the time people wait for a bus to arrive, increase security in different public stop areas among other benefits that the data collected can enhance. To eliminate the problem of challenges in ensuring efficient public transport, accurate data must be provided regarding the trips that passengers make including the origin and destinations. To acquire this data is challenging and expensive in general mainly because the destinations of the passengers are not recorded on the tickets.
The sensors will accurately record when a passenger enters a public transport vehicle and when and where they alight (Kostakos, 2007, 3) (Kerridge, et al., 2003). The statistics can then be employed in improving public transport by the small spatial resolutions that are attained. Thus, the data is used to plan based on the passenger’s behavior and improve the data collection practices. The data assists in improving public transport by making detailed schedules that consider the passengers actions. It also may lead to increasing buses while enhancing the satisfaction for travelers (Kostakos, 2007, 12).
The infrared sensors can also be used to track semantic data that concerns the transportation mode. That is; it can be used to identify train users, and differentiate them from tram users or car users. The info can be utilized for determining how many workers use the train, tram, or cars. Thus, the number cannot be highly accurate or reliable (Versichele, et al., 2012, 214). Bluetooth scanners are reliable tools for collecting certain forms of data but due to the limitations of having to detect all mobile phones and users on the same route, the data cannot be relied upon, a hundred percent (214).
The research methodology for this study is an experiment conducted to discover the reliability of using Smartphones to collect crowd information. The devices to be utilized for the survey included infrared sensors, data receivers, and Smartphones.
The experiment involves using the infrared sensors to count people in different localities. The data received by the infrared devices will then be sent to the data receiver devices are chosen using Wi-Fi. That part is identified as the sensing part of the experiment.
The second part of the test begins once the data from the infrared devices is received. The data is analyzed and later sent to the network server.
The third and last step begin with the smartphones been activated to collect the information regarding the crowd. That is; it gathers data relating to the number of people within a crowd. The infrared sensor devices check the data. As perceived, the App or Network software used in the phone is good to find the data of a crowd.
The results attained show that Wi-Fi can collect data for a wider range when compared to Bluetooth as it receives information for about 35 meters and 100m outdoors. Thus, the device can be used to collect accurately and reliably data of a crowd.
infrared automatic counting sensors
city council pedestrian counting system
manual counting systems
The experiments contains:
Comparison between high flow and low flow with different height.
Comparison between big, medium and low distance.
Comparison of the accuracy of infrared automatic counting and city council pedestrian counting system (3 locations).
The proportion of grouping.
1. Comparison between high flow and low flow
First Location: West Campus (High flow) Building 202
Compare high flow with three level of height
The figure presents an overview of West campus where the first experiment was conducted.
High flow with height 1.4m
Figure 2 presents the visual preview point where the infrared monitoring technology was located; the height was determined at 1.4 meters while the length was 3meters.
In= 149, Out= 117, total =266
The total=266 Actual manual counting=295. The accuracy= 90.1%
The results collected from the experiment in figure two are presented in figure three. The accuracy of the infrared technologies is high and very reliable. The results derive from the 30-minute experiment, which shows that the accuracy of using infrared devices is 90.1%, which is a reliable and acceptable accuracy level. The number of people coming in is higher compared to those coming out. Thus, the information may stipulate that the business or activities inside the place the people enter are successful as the people once they get it, they do not leave immediately or soon after.
The difference with the counting was low as presented in figure four. The actual counting resulted in 295 while the automatic counting was at 266. Thus, collecting data on such a high flow with a height of 1m and length of 3m will assist in collecting reliable data that can be used in making decisions.
High flow with height 1m
Figure five presents a clear visual representation of where the devices were located. On the sides of the doors at the presented meters. The difference of the experiment with the experiment explained above is that this test was conducted on a height of 1meter and length at 3meters.
In = 204, Out =152, total =356
The total=356 Actual manual counting=507. The accuracy= 70.21%
Figure six presents the results attained when the devices height was changed to 1meter with the same length of 3meters. The accuracy of the counting decreased. That is; the difference of the actual and automatic counting was high given that the accuracy of the device counting was at 70.21%. As the accuracy level is beyond 50%, the statistics attained can be deemed reliable or functional. However, the accuracy in collecting data is low given the differences as presented in figure 7. The figure presents that the people going in are way higher to those coming out as the minutes go on. Thus, the more time there is, the more people get in.
High flow with height 60cm
In= 132, Out =102, total= 234
The total=234 Actual manual counting=322. The accuracy= 70.48% the height 60 cm
Figure 8 provides an outline of the visual view when the devices are located at 60cm height. The accuracy level is better compared to that of 1meter height. The accuracy level is acceptable and can be used in making decisions since it is above 50%, which is the acceptable level of accuracy. Within the first minutes of the experiment, the people coming in were almost equal to the number of those leaving. However, the people getting in were more than those coming out as perceived in the graph.
Figure 10 presents the difference in counting perceived when the sensors were located at 60cm height. The difference appeared a bit high but given the accuracy level; it is an acceptable level.
Figure 11 presents the results attained from the first experiment at the West campus high flow area. The figure is a presentation of the automatic and manual performance achieved at different heights. According to the figure, the accuracy level of automatic vaunting is reliable for decision-making but highly accurate with higher heights.
Figure 12 provides the overview of the percentage differences, which support that accuracy is attained at a higher height in high flow areas. The presentation in Table 1 below presents the error reached when using the three different heights.
The performance supports that the error is low when the height is higher. Thus, the infrared sensing devices can be highly relied upon on high flow areas when the height is high. For instance, at 1.4m, the error percent was -9.830.
Second Location: East Campus Building 251 (Low flow)
Compare low flow with three level of height
Figure 13 presents an overview of the three low flow areas in East Campus that the experiment would be conducted. The same heights as that used in high flow areas will be used for a consistent length of 3.3 meters.
Figure 14 presents the visual outlook of where the sensors will be located. The height is 1.4meters as depicted above.
In=23, Out= 18, total= 41
The total=41. Actual manual counting=43. The accuracy= 95.3%
Figure 15 presents an overview of the results from experiment 1. As presented above, the accuracy level is 95.3%. The accuracy level is highly accurate and thus reliable for crowdsensing. There was no difference to the number of people coming in and out; the difference was significantly low.
As figure 16 presents, the difference with the actual counting was two. Thus, the difference perceived in the experiments is low, stipulating that the sensor device is reliable in such an area fro crowdsensing data.
Low flow with height 1m
Figure 17, presents an overview of the experiment at a 1meter height. The experiment is still in a low flow area.
In=20, Out=15, total 35
The total=35 Actual manual counting=36. The accuracy= 97.2% the height 1 m
The results provided are recorded above. Given the information collected, and accuracy level attained of 97.2%, it is clear that the sensor device was accurate. Thus, it can be relied upon for collecting data in low flow areas. The graph presents that the people getting in were slightly more than those that were leaving. However, it is possible that the difference was an error since the graph shows an insignificant difference.
The difference attained from the 1meter height experiment stipulates that the actual manual counting and automatic are closely similar. Thus, the device can be used in such a low flow area and at such low heights for accurate data collection.
Low flow with height 60cm
The figure above (20) presents an outline showing the graphic application of the sensor at 60cm.
In= 32, Out= 29, total=61
The total=61 Actual manual counting=69. The accuracy= 88.4 % the height 60 cm
The figure 21 presents that the accuracy in counting at 60cm of height is low, yet high since it recorded 88.4%. The sensor devices can be relied upon highly on low flow areas for accurate data collections besides reliability.
Figure 22 presents the differences in automatic and actual manual counting at the 60cm height. The difference recorded as shown, is very low. Consequently, the sensor devices can be used to collect data in low flow areas. Importantly, the accuracy of devices cannot be probed.
Figure 23 presents the results of the low flow area experiments conducted. The three tests conducted all recorded high accurate numbers showing that despite the height, the sensor devices are reliable. However, at the 60cm height, the accuracy level decreased. The result presented that the sensor devices should not be placed so low for accurate data collection.
Figure 24 gives the accuracy attained within the three heights used. At 1 meters height, the data accomplished was highly accurate, followed by the 1.4meters. However, at 88.4 meters, the accuracy decreased. However, all the heights provided highly accurate and reliable data that can be used in major decision-making processes.
2. Comparison between big, medium and low distance
Location: West Campus between Building 204 and Building 210
Figure 25 is a visual presentation of the big distance to be involved in the 1st experiment. The range as presented above is 17 meters. The height is 1m.
In= 28, Out= 23, total= 51
The total=51 Actual manual counting=55. The accuracy= 92.7% (15 min) Length about 17m. The height 1m
Figure 26 provides the results of the experiment. As perceived, the accuracy of the experiment was at 92.7%. Thus, the results show that at such a distance, the sensor device can be relied upon for correct data collection.
Figure 27 is an overview of the difference between manual and automatic data collection. Judging from the numbers in the graph, the difference is low. Thus, the device can be relied on for precise data collection at big distances.
Location: West Campus beside library building 210
Figure 28 presents the visual overview of the medium range to be used in the experiment. The length of 2.7meters.
In=12, Out=10, total 22
The total=22 Actual manual counting=24. The accuracy= 91.6% (15 min) Length about 2.7m
The height 1m
Figure 29 presents the results of the medium distance of 2.7 meters experiment. The results recorded an accuracy level of 91.6%. The accuracy of the sensor device in recording the data was high and acceptable. The graph presents that the people getting out were higher than those getting in. However, as the count presented that those getting in was more, the difference may be related to the errors detected in the literature review (Kerridge, et al., 2003).
Figure 30 presents that the difference in the counting within the automatic and manual counting was low, as perceived above stipulating that the devices should be used in collecting data even for such medium distances.
Location: East Campus Building 253
Figure 31 presents a view of the low distance to be used in the experiment. A height and length of 1 meters each.
In=24, Out 15, total= 39
The total=39 Actual manual counting=42. The accuracy= 92.8% (15 min) Length about 1m
The height 1m
Figure 32 presents the results of the experiment where the accuracy level recorded was 92.8%. The results stipulate that the sensor devices can be used for accurate data collection in low distances. The information in the graph stipulates that those getting in was way higher than those leaving throughout the 15 minutes the experiment was conducted.
The figure 33 present that the difference between automatic and manual data collection in the low distance is almost equal. Thus, the devices can be relied upon for actual data collection in the low distances.
Comparison between high, medium and low distance:
According to figure 34, the difference between manual and automatic data recording within big, medium and low miles is low. Thus, the figure presents that the infrared sensor devices can be relied upon for data collection despite the distance.
Figure 35 and table three also support the results by showing that the accuracy level of the experiments is high. The table also indicates that the error level perceived in all the distances used is low, proposing that the sensor devices can be relied on for accurate data collection.
3. Comparison of the accuracy of infrared automatic counting and city council pedestrian counting system.
Location(1): Lygon St (West)
Figure 36 presents an aerial overview of Lygon St (West and East)
Figure 37 presents an overview of automatic counting infrared systems at a distance of 4.6m and 1.4 meters height.
In=133, Out=104, total= 237
The total=237 Actual manual counting=282. The accuracy= 84.04%
Figure 38 presents an outline of the recorded accuracy of using automatic infrared sensor devices to count the pedestrians at the Lygon set area within an hour. The accuracy is at 84.04%, which is an acceptable level for data collection. Within the first few minutes, those getting in were more than those that were leaving. However, after a few minutes, those getting inside were way more compared to those leaving as the graph proves.
Figure 38. 2
The total=324 Actual manual counting=282. The accuracy= 114.8%
According to figure 38.2, the accuracy level of using the city council pedestrian counting system was beyond 1005. The level recorded was 114.8%. Thus, the device can be used for accurate data collection.
According to figure 39, the city council device gave over count about 15%. Infrared gave under count about 16%, so the both devices are close each other. The difference between them about 30.7%.
Therefore, both devices can be relied upon for accurate data collection.
According to figure 40, the accuracy of a counting technology describes how close the counts it produces are to the actual number of pedestrians that should be counted. When the number of a particular technology is lower than the actual count, the technology is said to undercount. When the technology count is higher than the actual count, it is said to over count (ViaStrada 2009), (NCHRP 2014).
Location(2): Lygon St (East)
Figure 41 presents an overview of the East of Lygon the experiment was conducted.
In=171, Out=104, total= 158
The total=329 Actual manual counting=450. The accuracy= 73.11%
Figure 42 is the presentation of the experiment in East of Lygon using the automatic infrared sensing devices. The accuracy level recorded is 73.11%, which is a figure that is close to the accurate number thus, reliable for data collection.
The total=415 Actual manual counting=450. The accuracy= 92.22%
The city council counting system according to figure 43 provides information that leads to an accuracy of 92.22% in recording data. The numbers recorded by the actual and system recording are slightly different. Thus, they are highly reliable and accurate for data collection.
The difference between the automatic and Melbourne pedestrian counting system is slight. However, as presented in figure 44, the city council system for counting pedestrians is more reliable and accurate than that of automatic systems.
According to figure 45, the city council pedestrian counting system is more accurate than the infrared automatic counting. Thus, the City Council pedestrian counting system is more reliable and exact compared to that of infrared sensor devices.
Location(3): New Quay
Figure 46 presents a visual view of New Quay where another experiment will be conducted.
Figure 47 presents an outlook of the area and meters used in the experiment. The big distance at a height of 1meters.
In=129, Out=92, total= 221
The total=221 Actual manual counting=265. The accuracy= 83.39%
According to figure 48, the accuracy attained through the usage of automatic counting is 83.39%. The difference recorded was slight as perceived above.
The total=242 Actual manual counting=265. The accuracy= 91.32%
On using the city council system, the accuracy attained as presented in figure 49 was 91.32%. Thus, the system is highly reliable for collecting accurate data, as the difference recorded was small.
Figure 50 is a presentation of the accuracy of using the automatic infrared sensor and Melbourne pedestrian counting system. The city council system as presented is more accurate, compared to the infrared system. The figure 51 below also supports the conclusion as it presents that the city system accuracy is high compared to the infrared system.
Table 5 demonstrates the error level accounted for by the systems in the three locations. The city council system has a low error degree in the three areas compared to those of the infrared system as presented. Thus, according to the results, the Melbourne city council system is more reliable for accurate data collection and pedestrian counting.
The proportion of grouping (I have taken the last location as an example) I have monitored the screen in the sensor, and I found the distance less than 5 cm between two people or more the sensor count them as one. Table 6 shows the missing group.
Location 4: Bourke Street Mall (North)
Figure 52 present an aerial overview of the Bourke Street mall where the last experiment will be conducted.
The total=3236 Actual manual counting = 3561. The accuracy= 90.87%
On using the city council system, the accuracy attained was 90.87% as the difference/error recorded was low. Thus, the device can be relied upon for accurate data counting.
According to the experiment as depicted in figure 54, the device which used by city council gave a good accuracy about 90% of high flow areas (3561 persons).
The accuracy of Infrared sensor devices in counting people in high and low flow areas.
Wi-Fi and Bluetooth cannot be accurate when counting people because smartphones used by the users can be off. This is because of battery issues, or the user can turn it off.
Collecting data using Infrared devices does not invade the smartphone user privacy as its means of data collection are not linked to the phone of the user but the motion of the user.
Infrared sensing covers all directions thus, poses an opportunity in public areas where there are many people to be counted.
Will the development of the Smartphones application help to improve crowd sensing in a public place?
How is it possible to use both Infrared technologies in for crowdsensing?
Is it appropriate to use infrared technologies in crowded places such as festivals?
Does the use of Wi-Fi affect the privacy of the smartphones users?
How is it possible to use infrared technology for crowd sensing? And what about the infrared accuracy in low-density and high-density?
Results according to the data above:
Because of the working principle of infrared counting device, the TX transmit the ray, and the RX receive the ray like two lines. When people enter or leave the entrance, they will cut off the line, and the RX will count accordingly. Therefore, if two people enter the door shoulder by shoulder, it is only cut off once, so the count is one. However, if there is about 50mm distance between the two people, HPC005 can count as two.
If some people are crossing the sensors smoothly one by one the accuracy will be more 90% and above.
If some people were crossing the sensors as a group, the accuracy would decrease to about 70%.
60 cm height means the pocket of trousers or thighs. The counter may be affected by both legs and arms and sometimes blocked the sensors. However, this level is suitable to count children.
Accuracy was not affected by changing the distance between the sensors.
No clear impact or effect of indoor or outdoor environment on the precision.
Vertical sensors such as those used by City Council pedestrian counting system are better than horizontal sensors such as we used. Because in vertical sensors more accurate with a group of people when crossing the sensors. Additionally, the vertical type can count children (low height).
City council using two types of sensor – laser, for uncovered areas, and thermal, for covered areas. The entire system is audited every year, and the levels of accuracy are within our allowable thresholds (+/-10%). Infrared beam counter:
A major limitation the research encountered is the lack of automated ways for adjusting sensor-counting error.
Transmitters and receivers require careful alignment that is used to guarantee the receiver of the beam has a perfect reception for recording and receiving data.
The transmitters and receivers do not perform well when installed under a flexible structure.
Cheap and widely available commercially
Low power consumption
In the past, most of the sophisticated technology present such as infrared technologies, Bluetooth and Wi-Fi were not employed in the past as they are of the present. However, using infrared beam counters technologies provides vital information about the customers of a corporation and very useful for the minimization of investment risk (Chen, et al., 2010). The robust positioning technologies have motivated the development of sensors with the capability of monitoring peoples and object movement. According to (Abuarafah, et al., 2012), infrared beam counter systems meet these needs efficiently, as they count motion rather than smartphones with Wi-Fi or Bluetooth on. Human movement behavior analysis has received lots of notice mainly in the turf of visual analytics.
Within research on human behavior, devoted attention to people behavior in mass events such as public assemblies, sporting events, festivals, and exhibitions need to be monitored. The infrared counter systems have a multi-counting systems capability as (Hashimoto, et al., 1997) have put it. Terrible events like the recent stampede that occurred during the 2010 edition of the love parade in Duisburg shows that we need to have precise information on the expected flow of visitors at public occasions (Li, et al., 2015). However, the gathering of quantitative movement data using advanced tracking technologies such as GPS (global positioning system), in these contexts raises critical issues regarding viability. However, using infrared beam counter systems is not challenging, as it is reliable and cost effective (Hashimoto, et al., 1997), as presented in the experiment results and process show above.
Several experiments were set-up at various public places with high people traffic using the infrared beam counter and manual counting to detect the accuracy of using infrared for crowdsensing. The results presented that Infrared Beam counter systems can be used for accurately collecting data despite height or distance. Crowd sensing has proved essential while managing the passenger’s crowds at major transport systems. Collecting this data is challenging and expensive in general, crowd sensing has opened up new ways of collecting data that are novel and cost-effective systems.
Using the infrared to collect data is beneficial since it gathers data on the flow of customers in all directions (Hashimoto, et al., 1997). The data can be used to determine the business trend analysis, evaluation of advertising and promotions among other operations. Exploiting the information gained to upturn sales by leveraging buyer-to-traffic ratio and turnover per visitor can also be performed. It also helps evaluate how many visitors turn into buyers, where the information can be used to determine the effectiveness of the sales team. One can view the number of visitors in a place, and the peak hours of a store. The network helps check data from different stores where the data attained is sent directly to the server without any limitation.
The infrared counting sensors are attractive since they have the power to reduce labor costs linked to the reports provided on pedestrian activity among other benefits (Fanping, et al., 2007). However, infrared devices are more reliable since they do not interfere with the privacy of the customers (Chen, et al., 2010). More importantly, they collect data on the motion not smartphones, where they can be off or undetectable under certain conditions. Thus, the functionality of mobile sensing can be carried out without the disruption of continued communication flow (Radianti, et al., 2014, 2). The significance of crowd sensing can be aligned with the benefits it provides in emergent and hazardous situations (Radianti, et al., 2014, 3). Crowd sensing has proved valuable while managing the passenger’s crowds at major transport systems. Crowd sensing has opened up new perspectives for cost-effective ways of managing the traffic congestion
Crowd sensing provides an opportunity that assists individuals work on increasing data collection methods. For instance, infrared beam counters can be used for estimating the crowd density by accurately recording the people entering and leaving a building or any other area (Garcfa, et al., 2013). Crowd density data is important since it allows individuals to collect functional data for making decisions (Weppener & Lukowicz, 2011, 3). Such information can be used during emergencies presenting the need for using sensing devices to gather information during different situations (Xu, et al., 2013, 1). It is important and can present accuracy when engaging in crowd count processes during any form of monitoring (Xu, et al., 2013, 2).
Socially the App or infrared beam counters can be used to determine the crowd at a restaurant and avoid the hassles of traveling to a restaurant only to find it crowded among other benefits. For business managers and teachers among other stakeholders, the device can be used to determine whether the audience is engaged in a certain topic or not (Xu, et al., 2013, 9). Thus, the process is beneficial in the different setting and can be used officially for crowd counting or unofficially. Using Bluetooth or Wi-Fi for crowd sensing has proved to be reliable tactics for obtaining various forms of data in the large-scale event setting. The key information the project presents is that Smartphones among infrared technologies have advanced to the most reliable and efficient tools for collecting data in a large crowd among other environments.
Using infrared technologies as perceived through the experiment is reliable as it provides accurate datasets from the crowd sensing information collected. Infrared as presented above presents the benefit of detecting data whenever there is motion irrespective of the direction. Thus, the data is highly reliable, and given that it does not interfere with the personal gadgets of individuals, it is highly safe and acceptable. More importantly, using infrared is beneficial as it is cheap and widely available commercially, the consumption of power by the devices is low, easy installation and are highly portable. Thus, the devices can be relied upon for data collection in all directions, and big distances, as well as low and high heights. According to the experiments conducted, the data from the infrared beam counters are accurate despite the high or low flow of pedestrians.
Although there exists Apps for crowd sensing, with advancement in technology, there should be substantial progressions in ensuring efficient and reliable data collecting. Infrared beam counters are the signals that provide reliable data. Therefore, other devices such as Bluetooth, SmartRescue, crowd++, GPS, cameras, and others ought to be advanced to ensure that they also play a significant role in ensuring the attainment of reliable data. It should be developed while ensuring that the devices do not interfere with the privacy of smartphone users. Research on how to increase and ensure the reliability of the information given has to be conducted. For instance, according to the experiment conducted, using infrared beam counters provides accurate information that can be readily used in decision-making, as the figures that were given are almost similar to those manually counted. Therefore, research ought to be conducted on how to determine the accuracy of information from other technologies. Accuracy, reliability and privacy of the users are important factors that must be guaranteed to ensure that in the future it is highly accepted and used for collecting data.
The drawbacks of using infrared bean counters is the lack of the sensors differentiating between the motion of people and objects. For instance, if an insect flying close to the transmitter intercepts the signal, the data logger counts the intercept as a person. Additionally, to attain data accurately, the receivers and transmitters must be aligned as required without a flexible structure. Additionally, when more than one person intercepts the beam simultaneously, the beam counts it as one person (Fanping, et al., 2007). Thus, advances to improve the performance of the devices must be conducted.
Most importantly, in future, methods should be considered to improve the infrared sensors given that not all sensors are reliable for all pedestrian counting. Some of the constraints to be improved include the specifications of the data needed, the budget constraints, and the accuracy level needed. That is; most of the devices were developed for detection not counting. Therefore, improvements should be made on how the devices can now be used for counting accurately rather than sensing (Ruser, et al., 2006).
Additionally, the infrared sensors can be improved in some ways. For instance, the sensors could be improved by adding about six detectors. Three detectors in a horizontal position with three position ( less than 5 cm between the angles), three sensors in a vertical position with three position to detect two positions of height and one sensor work by thermal the detect the heat of human body then join all sensors by using Logarithms. The improvement would ensure that the error of counting two people as one is eliminated.
More Important Things