A WASTE LEVEL DETECTION SYSTEM USING IOT SENSORS …...measure the amount trash of a bin at a low...

10
http://iaeme.com/Home/journal/IJARET 1482 [email protected] International Journal of Advanced Research in Engineering and Technology (IJARET) Volume 11, Issue 10, October 2020, pp. 1482-1491, Article ID: IJARET_ _10_142 11 Available online at http://iaeme.com/Home/issue/IJARET?Volume=11&Issue=10 ISSN Print: 0976-6480 and ISSN Online: 0976-6499 DOI: 10.34218/IJARET.11.10.2020.142 © IAEME Publication Indexed Scopus A WASTE LEVEL DETECTION SYSTEM USING IOT SENSORS AND VIBRATION MOTOR SangGu Na and Eun-Kyu Lee* Department of Information and Telecommunication Engineering Incheon National University, Incheon, Korea Junghee Jo* Department of Computer Education Busan National University of Education, Busan, Korea *Corresponding Authors ABSTRACT Trash is an inevitable by-product in human life. However, emptying a trash bin takes labor and time to identify and vacate at home and on the streets. To address the issue, this paper proposes a waste level detection system that can automatically measure the amount trash of a bin at a low cost. The proposed system employs the state-of-the-art Internet of Things (IoT) technology; it consists of an inexpensive mini vibration motor, acceleration and weight sensors, and IoT platform. The motor generates vibrations periodically that are captured by the accelerometer. Data with weight information is analyzed to detect -level of a trash bin. To this end, we fill employ a k-nearest neighbor algorithm. This paper develops a prototype and runs extensive experiments to evaluate accuracy performance of the proposed system. Results show that the system can measure waste level with accuracy of 75.39% on average and 94.7% on maximum. Key words: Internet o Things, Distributed system, Waste level detection, Sensors, f Vibration. Cite this Article: SangGu Na, Eun-Kyu Lee and Junghee Jo, A Waste Level Detection System Using IoT Sensors and Vibration Motor, International Journal of Advanced Research in Engineering and Technology, 11(10), 2020, pp. 1482-1491. http://iaeme.com/Home/issue/IJARET?Volume=11&Issue=10 1. INTRODUCTION Internet of Things (IoT) is starting to increase in usage globally. Not only individuals like students, but also businesses have begun to use it, with more than half (57%) of companies already introducing IoT technology, and the ratio is expected to reach 85% by 2021 [1]. Among many IoT services, waste management is one of the most important appealing applications since everyone must deal with waste at home and/or at office. Statistics from the

Transcript of A WASTE LEVEL DETECTION SYSTEM USING IOT SENSORS …...measure the amount trash of a bin at a low...

Page 1: A WASTE LEVEL DETECTION SYSTEM USING IOT SENSORS …...measure the amount trash of a bin at a low cost. The proposed system employs the state-of-the-art Internet of Things (IoT) technology;

http://iaeme.com/Home/journal/IJARET 1482 [email protected]

International Journal of Advanced Research in Engineering and Technology (IJARET) Volume 11, Issue 10, October 2020, pp. 1482-1491, Article ID: IJARET_ _10_142 11Available online at http://iaeme.com/Home/issue/IJARET?Volume=11&Issue=10 ISSN Print: 0976-6480 and ISSN Online: 0976-6499 DOI: 10.34218/IJARET.11.10.2020.142

© IAEME Publication Indexed Scopus

A WASTE LEVEL DETECTION SYSTEM USING IOT SENSORS AND VIBRATION MOTOR

SangGu Na and Eun-Kyu Lee* Department of Information and Telecommunication Engineering

Incheon National University, Incheon, Korea

Junghee Jo* Department of Computer Education

Busan National University of Education, Busan, Korea *Corresponding Authors

ABSTRACT Trash is an inevitable by-product in human life. However, emptying a trash bin

takes labor and time to identify and vacate at home and on the streets. To address the issue, this paper proposes a waste level detection system that can automatically

measure the amount trash of a bin at a low cost. The proposed system employs the state-of-the-art Internet of Things (IoT) technology; it consists of an inexpensive mini

vibration motor, acceleration and weight sensors, and IoT platform. The motor generates vibrations periodically that are captured by the accelerometer. Data with

weight information is analyzed to detect -level of a trash bin. To this end, we fill employ a k-nearest neighbor algorithm. This paper develops a prototype and runs

extensive experiments to evaluate accuracy performance of the proposed system. Results show that the system can measure waste level with accuracy of 75.39% on

average and 94.7% on maximum.

Key words: Internet o Things, Distributed system, Waste level detection, Sensors, f Vibration.

Cite this Article: SangGu Na, Eun-Kyu Lee and Junghee Jo, A Waste Level Detection System Using IoT Sensors and Vibration Motor, International Journal of Advanced Research in Engineering and Technology, 11(10), 2020, pp. 1482-1491. http://iaeme.com/Home/issue/IJARET?Volume=11&Issue=10

1. INTRODUCTION Internet of Things (IoT) is starting to increase in usage globally. Not only individuals like students, but also businesses have begun to use it, with more than half (57%) of companies

already introducing IoT technology, and the ratio is expected to reach 85% by 2021 [1]. Among many IoT services, waste management is one of the most important appealing

applications since everyone must deal with waste at home and/or at office. Statistics from the

Page 2: A WASTE LEVEL DETECTION SYSTEM USING IOT SENSORS …...measure the amount trash of a bin at a low cost. The proposed system employs the state-of-the-art Internet of Things (IoT) technology;

A Waste Level Detection System Using IoT Sensors and Vibration Motor

http://iaeme.com/Home/journal/IJARET 1483 [email protected]

Ministry of Environment in Korea reports that the amount of waste increasing every year. isAs of 2018, 430,713 tons of waste are generated everyday [2]. The fact that so much garbage is being discharged is also a sign of many emptying bins. Failure to clean up the trash on time will result in disadvantages. home, we see trash directly with our eyes and empty a waste Atbin in a short time, but the external environment is different. Inside university buildings, there

are many bins placed on each floor, and there are many bins on streets. A short cycle of checking and emptying consumes a lot of labor and time, while a long cycle makes it look bad and unhealthy.

To address the issue, this paper proposes a waste level detection system that measures how much a trash bin is filled with. The system makes use of off-the-shelf IoT devices - mini-

vibration motors, accelerometers, and weight sensors. Attached to the bin, these devices record values periodically and transmit them to an IoT platform where data is collected and processed. The platform implements and runs a machine learning algorithm that performs a classification process and determines whether the bin is in empty, half-full, or full. If full, it sends a signal so that you can empty the trash bin. To demonstrate the feasibility, this paper develops a prototype with a plastic bin. Performance of the proposed system is evaluated by

extensive experiments where 4,010 data are used for the learning phase and 317 data are tested in the detection phase. Experimental results report that our system detect waste can

level of a bin successfully with 75.39%. We note that using cheap off-the-shelf devices is one of benefits of the proposed system. Thus, it is expected that labor and time can be reduced by detecting the level of filling of garbage using relatively inexpensive equipment and sensors.

The rest of the paper is organized as follows. Section 2 reviews related works and provides fundamentals of a machine learning algorithm used in the system. In Section 3,

details of the proposed system are described including design, prototype, and data processing. Section 4 explains our experiments, reports results, and evaluates accuracy performance of the proposed system. Finally, we conclude this paper in Section 5.

2. PRELIMINARY 2.1. Related Works A bin level detection system has been an interesting theme in academia [3]. Many approaches

employ image processing technologies [4-6]. They analyze images and compare the last image with the previous one. They get images of differences from the two images and

subtract each other to investigate only the latest trash thrown into the trash bin. Comparing the two images and finding the difference is one of the simple methods of image processing

techniques, but it has many shortcomings. For example, if a new image is affected by some factors, it is difficult to estimate and also affects image quality. Al Mamun et al., proposed a waste state management system using ultrasonic sensors, acceleration sensors, weight sensors and gas sensors [7]. The amount of garbage inside the Bin is measured using the travel time of ultrasonic waves, or round-trip time. Force Sensitive Resistor (FSR) is a sensor that can detect physical pressure, squeeze and weight, acceleration sensor is a detection device that detects the state of the wastebasket, and gas sensor is a detection device that detects the presence of

various gases. The system measures the height of garbage with an error of 5%. However, composite sensors are expensive and require careful installation and calibration. Rovetta et al., developed a bin that was equipped with camera, ultrasonic, weight, pressure, LED, and temperature sensor [8]. Using a list of sensors causes some problems; for instance, it leads to high cost and makes it difficult to calibration.

Vibration characteristics have been used in such systems that track and monitor human activity -13]. Mokaya et al., proposed MyoVibe that uses accelerometer-based sensors to [9detect activation of skeletal muscle during intense motion and motion [9]. The authors also

Page 3: A WASTE LEVEL DETECTION SYSTEM USING IOT SENSORS …...measure the amount trash of a bin at a low cost. The proposed system employs the state-of-the-art Internet of Things (IoT) technology;

SangGu Na, Eun-Kyu Lee and Junghee Jo

http://iaeme.com/Home/journal/IJARET 1484 [email protected]

proposed Burnout, a system that prevent injury by detecting and analyzing vibrations in ed muscles that occurred in motion [10]. The system also uses accelerometers to detect

vibrations in skeletal muscles. Although our system also use vibrations in common, the human body and a bin are different, so we did not use muscle vibration sensors. Our system isrelatively easy to implement, do not use complex sensors, do not use difficult experiments es esand complex analyses, and installed outside a bin, so there is less damage. is

2.2 Machine Learning Algorithm Used This paper makes use of a machine learning algorithm, K-Nearest Neighbors (KNN) [14]. The input data in KNN is divided into groups that are less than or equal to , where each cluster k nforms a cluster. That is, data is divided into groups of one or more data objects. When the kcenter of the ith cluster is ui, the set of points belonging to the cluster is Si, the total variance is calculated as follows:

Finding Si that minimizes this value is the goal of the algorithm. The algorithm starts by first setting the initial value, ui . Then repeat the following two steps:

Cluster Settings: Each cluster calculates the Euclidean distance from each data to ui and

assigns the data to the nearest cluster from that data.

Cluster-centric reset: Reset ui to the center of gravity value of the data in each cluster.

Two steps are repeated until the cluster does not change. KNN has already been used in a wide range of areas, including market division and image

segmentation, and the technique is mainly used to pre-process data. We grouped the empty, half full, and full cases when the weight is light or heavy, and used the group data to find out how many sections of the group are / what the average is.

3 PROPOSED WASTE LEVEL DETECTION SYSTEM . 3.1 Overview This paper proposes a waste level detection system. Figure 1 shows a schematic diagram for its operation. An IoT node is attached to a bin; it comprises of Inertial Measurement Unit an(IMU) shield with the main computing device of the work, accelerometer, and a load cell anweight sensor located at the bottom of the bin. To measure vibration effectively, motors and shields should be attached to the surface of the wastebasket vertically and securely.

Upon powered on, the node connects to the Internet via -Fi. When connected normally, Withe RGB LED on the node board flashes blue like it breathes. If it is not connected, go back to the beginning and try to connect continuously. When the connection is complete, initialize the load cell weight sensor. This process is similar to adjusting the zero point of the scale. Then, it measures the sensor value on the accelerometer and the load cell weight sensor in the IMU shield. Vibration and weight depend on the level at which the trash bin is filled. The node collects these sensor values and sends the data to an IoT platform on the Cloud for further

Page 4: A WASTE LEVEL DETECTION SYSTEM USING IOT SENSORS …...measure the amount trash of a bin at a low cost. The proposed system employs the state-of-the-art Internet of Things (IoT) technology;

A Waste Level Detection System Using IoT Sensors and Vibration Motor

http://iaeme.com/Home/journal/IJARET 1485 [email protected]

processing and management. analysis on the platform eventually distinguish whether the Antrash bin is empty, half, or full.

Figure 1. This paper proposes a waste level detection system. This figure shows a schematic diagram

3.2. Waste Level Detection System 3.2.1. Using Motor and Sensor in the Proposed System Our system makes use of a vibration motor and an accelerometer in order to detect waste level of a trash bin.

Vibration is the motion or shaking of a machine or component's equilibrium position. It may be cyclical, such as pendulum movement, or irregular, such as a dent in a gravel road tire. Units of vibration are expressed in metric (m/s/s) or gravitational constant unit "g" (1g =

9.81 m/s/s Vibration generally measured using piezoelectric ceramic sensors or ). isaccelerometers, where the experiment was measured using accelerometers. An accelerometer

is generally referred to as general-purpose "vibration sensor" capable of measuring a vibration. This sensor analyzes the dynamic acceleration of a physical device by measuring it as a voltage. Inside the accelerometer, there is a piezoelectric material that causes a charge when applied. Attached on the upper or side of this piezoelectric material, a seismatic mass responds to acceleration, which in turn puts pressure on the piezoelectric material and uses the amount of charge generated from the piezoelectric material.

Figure 2 The proposed system attaches a motor and sensors the surface of a waste bin. to

Figure 2 sketches the proposed system using a motor and sensors. The IMU Shield and mini-vibration motor are attached to the surface. Vibration generated by the vibration motor is

sensed by the accelerometers in the IMU Shield, and measured values are output as the piezoelectric material changes within the accelerometer.

3.2.2. Prototype Development This paper develops the prototype of the proposed system on a plastic waste bin of 18cm x 14cm x 25cm in size and 9L of capacity. It uses Particle Photon as a main board [15]. It has a

Page 5: A WASTE LEVEL DETECTION SYSTEM USING IOT SENSORS …...measure the amount trash of a bin at a low cost. The proposed system employs the state-of-the-art Internet of Things (IoT) technology;

SangGu Na, Eun-Kyu Lee and Junghee Jo

http://iaeme.com/Home/journal/IJARET 1486 [email protected]

Wi-Fi chip with STM32 ARM Cortex M3 Micro Controller (MCU), which operates at 120 MHz. It also has 1MB of flash memory and 128KB of RAM. The board is also equipped with RGB LEDs that can indicate the status of the board. The code used in Photon is compatible with the Arduino code, and code and firmware can be uploaded and flash remotely through a cloud platform.

Figure 3. A prototype of the proposed waste level detection system (left) and a load cell weight sensor with HX711 module (right)

Particle Photon is mounted on SparkFun LSM9DS1 IMU Shield with a 3-axis accelerometer sample [16]. Because it is mounted vertically on the surface of the bin, only the information on the z-axis is used. A mini-vibration motor is attached to the underside of the

attached device. Vibration motors can control motor operation by giving HIGH and LOW signals to the Arduino code. Vibration generated by the vibration motor is measured on the

accelerometer of the IMU Shield. At the bottom of the bin, a load cell weight sensor is located. A weight sensor is a sensor that measures the weight of an object. On the surface of the road cell there is a resistance called "strain gauge". This resistance changes as much as it changes shape by external forces, which in turn results in a change in the electrical signal output, allowing the weight to be measured. This signal is very small, so the HX711 module

is used together to amplify the signal [17]. The vibrations and weights from the vibrating motors are collected and calculated from the Particle Photon and transmitted to an IoT

platform.

3.2.3 Data Processing and Analysis for Waste Level Detection In order to collect, store, process and perform analysis, our system uses an IoT platform,

ThingSpeak [18]. is a cloud-based data analytics solution from Matlab [19 that works on It ] an open source basis. Since using the RESTful API [20], it is relatively easy to send data to the platform. Although we can analyze experimental data to support Matlab scripts on the web, we used them for data collection and storage. For data analysis, we implement the KNN clustering algorithm in Python. We use algorism because it less complex than other KNN isalgorithms and suitable for separating the filling level of the trash can. Upon receiving sensor data from the accelerometer and load cell weight sensor, it divides groups of empty, half, and

full and performs analysis to identify the intervals, and compare the intervals and data to distinguish whether the trash can is empty, half and full. The node collects these sensor values and sends the data to an IoT platform on the Cloud for further processing and management. An analysis on the platform eventually distinguish whether the trash bin is empty, half, or full.

4. EXPERIMENTS AND PERFORMANCE EVALUATI ONIn order to evaluate performance of the proposed waste level detection system, this section

sets up the system and runs experiments that include two phases as follows. In a learning phase, collected 4,010 data and performed analysis. The data was also fed into the KNN we

Page 6: A WASTE LEVEL DETECTION SYSTEM USING IOT SENSORS …...measure the amount trash of a bin at a low cost. The proposed system employs the state-of-the-art Internet of Things (IoT) technology;

A Waste Level Detection System Using IoT Sensors and Vibration Motor

http://iaeme.com/Home/journal/IJARET 1487 [email protected]

algorithm, from which we created 6 clusters. In a detection phase, the proposed system detected the waste level of a bin. 317 data were tested in total. Based on the detection results, performance of the system was evaluated.

4.1 Learning Phase with Raw Data 4.1.1 Observation on Raw Data This paper first draws raw data of acceleration and weight transmitted from the waste bin to understand their changes. Figure 4 shows collected data over time. On the left, weight of the bin [gram] varies over time [minutes]. Increasing weight indicates that new trash is added into

the bin (the blue arrow) while the weight decreases when a user empties the bin (the red arrow). On the right, the acceleration value varies accordingly over time. Note that the timelines on two graphs (i.e., X-axes) are synchronized. It is observed that the acceleration

value reduces along with increasing weight in general, because a heavy weight attenuates vibration.

Figure 4. Changes of weight (left) of the bin and acceleration values (right) over time. The X-axis represents time in minute in both graphs, and units of Y-axes are gram and m/s/s in left and right

graphs, respectively.

Interestingly, however, acceleration increases as the weight increases in the case of heavy waste. In our experiments, notes and books are added to the bin making it heavy. Since those are hard in surface, we believe that vibration is transmitted well to the inside of the trash bin. To confirm our observation, we run an additional experiment, where we add heavy wastes

Figure 5 displays the results indicating that acceleration increases along with the increasing weight.

Figure 5. In the case of heavy waste, acceleration increases as the weight increases.

4.1.2 Learning with KNN In this phase, the raw data is fed into the KNN algorithm to make it learn samples. To this end, the value must be determined, representing the waste level of a bin. The bin may be kWith this simple distinction, however, it is difficult to make

Page 7: A WASTE LEVEL DETECTION SYSTEM USING IOT SENSORS …...measure the amount trash of a bin at a low cost. The proposed system employs the state-of-the-art Internet of Things (IoT) technology;

SangGu Na, Eun-Kyu Lee and Junghee Jo

http://iaeme.com/Home/journal/IJARET 1488 [email protected]

accurate judgments, which can cause confusion for users to see if the criteria are crossed, as they will be displayed as empty and full immediately. If a simple system error signals that it is not yet full, but only once, the user knows it is full, and he will go empty the bin, checking the actual bin, and then realizing that he has gone o The more states the system

defines, the more accurate level it can capture. However, too many states make the system complicated and may introduce detection errors.

To address this issue, a level of a waste bin in the proposed system classified into three isstates empty , half full , and full . If the bin is empty, is just empty or there is still less : ittrash in it. This condition is a distant priority among the bins that we should care about. The following is half full. A certain amount of trash informs users that the trash can is half full, and they have to pay attention because it can be full over time. Even in this state, one may go empty, but it is not efficient. Our system encourages users to go empty when signals that the ittrash can is full. Finally, the bin is full. It has determined that the waste bin is full of trash, and when the user receives the signal, he or she can go to collect the trash. Users only need to collect the signal from the trash bin, so they can efficiently plan the connection. If garbage is not collected, it is recommended that it be emptied for appearance and hygiene, as the garbage overflows.

Figure 6. Experiments run the KNN algorithm with =6. This creates 6 clusters (groups) where the red kdots represent their centroids.

Our experiments set the value as 6 and run the KNN algorithm with the raw data. Figure k 6 demonstrates the results forming 6 clusters. The X-axis of the graph represents weight,

while the Y-axis represents acceleration. Each dot represents acceleration when the weight is measured. Red dots are the centroids of the clusters, and the collection of different colors

means the cluster and the data that belongs to the cluster. Using the clusters formed, three clusters from the left can be used separately from the empty, half, and full of light waste, and three clusters from the right can be used to separate the empty, half and full of heavy waste.

4.2. Detection Phase In the detection phase, 317 data are tested to assess how accurately the proposed system is able to detect the waste level of a bin. Figure 7 shows an example of experimental results. The

X-axis represents time in minute in all graphs, while the Y-axes represents weight [g], acceleration [m/s/s], and waste level (0 is empty and 100 is full) in the left, middle, and right

graphs. The left figure tells that the waste bin was emptied at 20:45 after filled (i.e., the weight is 2 kg). At this moment, the acceleration curve drops in the middle figure, and

eventually the proposed system detect the waste level successfully. After the moment, edwaste was added to the bin gradually. Between 21:00 and 21:15, the system captured that the bin was

Page 8: A WASTE LEVEL DETECTION SYSTEM USING IOT SENSORS …...measure the amount trash of a bin at a low cost. The proposed system employs the state-of-the-art Internet of Things (IoT) technology;

A Waste Level Detection System Using IoT Sensors and Vibration Motor

http://iaeme.com/Home/journal/IJARET 1489 [email protected]

Figure 7. Experiments run the KNN algorithm with =6. This creates 6 clusters (groups) where the red kdots represent their centroids.

After testing all the 317 data, this subsection collects and reports experimental results, enabling us to evaluate how accurately the proposed system can detect the waste level of a bin. Table 1 summarizes detection performance. States in the first column represent true states that the bin is really in, whereas states in the first row represent the ones that our system determines. The numbers in the internal cells represent the number of times that the system determines in the experiment. Given 168 empty bins, for instance, our system determines that it is empty 167 times accurately and determines falsely that it is half full once. It is highly accurate to judge an empty condition, but when the bin is half full or fully filled, detection accuracy appears to be relatively poor. The accuracy of these collected data is approximately 75.39%.

Table 1 The proposed system performs detection for all 317 data and reports that a bin is either empty, half full, or full.

Waste level of bin Empty Half full Full Empty 167 1 0

Half full 47 43 4 Full 22 4 29

In order to investigate details of accuracy performance, this paper conducts additional experiments, where we fill the bin with either light trash or heavy trash. The results tell us how much weight affects performance.

Table 2 reports results of an experiment where the bin is filled by heavy trash. A total of 170 data is tested. In the case of the empty state, the system perfectly detects the waste level.

For bins with half full, the system determines only 8 times falsely, which performs better comparing to the values in the previous table. In the case of heavy waste, detection accuracy goes up to 94.7%.

Table 2 Accuracy performance when a bin is filled only by heavy trash.

Waste level of bin Empty Half full Full Empty 51 0 0

Half full 3 49 5 Full 0 1 61

Table 3 reports results of an experiment with the bin being filled by light trash. As shown, the overall performance is not as good as that in the case of heavy waste. An average accuracy is 48.51%. When a bin has light trash only, it does not have a greater frequency of vibrations and often oscillates out of boundaries. Thus, this results in more errors. In the case of heavy

waste, however, the distinction between vibration width and vibration width is relatively clear, so the classification of waste level is good.

Page 9: A WASTE LEVEL DETECTION SYSTEM USING IOT SENSORS …...measure the amount trash of a bin at a low cost. The proposed system employs the state-of-the-art Internet of Things (IoT) technology;

SangGu Na, Eun-Kyu Lee and Junghee Jo

http://iaeme.com/Home/journal/IJARET 1490 [email protected]

Table 3 Accuracy performance when a bin is filled only by light trash.

Waste level of bin Empty Half full Full Empty 26 3 2

Half full 26 9 5 Full 11 5 14

6. CONCLUSION Based on vibration and weight, this paper proposed a waste level detection system that detects

the filling level of a waste bin. It used inexpensive off-the-shelf devices, vibration mini motors, and IMU shield devices to calculate and process vibration and weight that

distinguish the level at which the bin is filled. Compared to other methods, the proposed edsystem provides a list of benefits as follows. (1) Easy to install. Instead of creating a complex bin, you just need to attach equipment to it. Some had the idea of attaching a camera to the

inside or installing an ultrasound, but it would damage the device and require complex software. However, the experiment is convenient because only the device needs to be installed outside the bin. (2) Low power consumption. Because the Particle Photon of the experiment

uses less power, it can last many years if lithium-polymer batteries are used. (3) Less expensive. Less equipment is needed and less expensive compared to other complex sensors. The price of IoT equipment can be purchased at about $30, the price of IMU Shield at about $20 and the mini vibration motor and load cell weight sensors at about $3 each. The paper also run experiments to evaluate accuracy performance of the proposed system. Experimental results show that the system can detect waste level of a bin with 75.39%.

ACKNOWLEDGEMENT This work was supported by a research grant from Busan National University of Education in 2020.

REFERENCES [1] Hewlett Packard Enterprise. Internet of Things, Today and Tomorrow. 2017.

http://chiefit.me/wp-content/uploads/2017/03/HPE-Aruba_IoT_Research_Report.pdf

[2] Ministry of Environment in Korea. Report on national waste in 2018: generation and disposal. 2019. http://www.kwaste.or.kr/bbs/content.php?co_id=sub0401

[3] MA Hannan, Md. A. A. Mamun, A. Hussain, H. Basri, and R. A. Begum. A review on technologies and their usage in solid waste monitoring and management systems: Issues and challenges. Waste Management, 43, 2015, pp. 509 523.

[4] M. Arebey, MA Hannan, R. A. Begum, and H. Basri. Solid waste bin level detection using gray level co-occurrence matrix feature extraction approach. Journal of environmental

management, 104, 2012, pp. 9 18.

[5] . Hannan, M. Arebey, R. A. Begum, and H. Basri. An automated solid waste bin level MA detection system using a gray level aura matrix. Waste management, 32, 2012, pp. 2229

2238.

[6] . S. Islam, MA Hannan, H. Basri, A. Hussain, and M. Arebey. Solid waste bin detection Md and classification using Dynamic Time Warping and MLP classifier. Waste management,

34(2), 2014, pp. 281 290.

[7] . A. A. Mamun, M. Hannan, A. Hussain, and H. Basri. Integrated sensing systems and Mdalgorithms for solid waste bin state management automation. IEEE Sensors Journal, 15(1), 2015, pp. 561 567.

Page 10: A WASTE LEVEL DETECTION SYSTEM USING IOT SENSORS …...measure the amount trash of a bin at a low cost. The proposed system employs the state-of-the-art Internet of Things (IoT) technology;

A Waste Level Detection System Using IoT Sensors and Vibration Motor

http://iaeme.com/Home/journal/IJARET 1491 [email protected]

[8] A. Rovetta, F. Xiumin, F. Vicentini, Z. Minghua, A. Giusti, and H. Qichang. Early detection and evaluation of waste through sensorized containers for a collection monitoring application. Waste Management 29(12), 2009, pp. 2939 2949.

[9] F. Mokaya, R. Lucas, H. Y. Noh, and P. Zhang. Myovibe: Vibration based wearable muscle activation detection in high mobility exercises. Proceedings of ACM Int l Joint Conf. on

Pervasive and Ubiquitous Computing. 2015, pp. 27 38.

[10] F. Mokaya, R. Lucas, H. Y. Noh, and P. Zhang. Burnout: A Wearable System for Unobtrusive Skeletal Muscle Fatigue Estimation. Proceedings of ACM/IEEE Int l Conf. on Information Processing in Sensor Networks (IPSN). 2016, pp. 1 12.

[11] Z. Jia, M. Alaziz, X. Chi, R. E Howard, Y. Zhang, P. Zhang, W. Trappe, A. Sivasubramaniam, and N. An. 2016. HB-Phone: A Bed-Mounted Geophone-Based Heartbeat Monitoring System. Proceedings of ACM/IEEE Int l Conference on Information Processing in Sensor Networks

(IPSN). 2016, pp. 1 12.

[12] J. Ranjan and K. Whitehouse. Object hallmarks: Identifying object users using wearable wrist sensors. Proceedings of ACM International Joint Conference on Pervasive and Ubiquitous

Computing. 2015, pp. 51 61.

[13] A. Khan, S. Mellor, E. Berlin, R. Thompson, R. McNaney, P. Olivier, and T. activity recognition: skill assessment from accelerometer data. Proceedings of ACM Int l Joint Conf on Pervasive and Ubiquitous Computing. 2015, pp. 1155 1166.

[14] N. Altman, An introduction to kernel and nearest-neighbor nonparametric regression. The American Statistician, 46(3), 1992, pp. 175 185.

[15] Particle Photon. https://docs.particle.io/. [16] SparkFun IMU Shield GitHub.

https://github.com/sparkfun/SparkFun_LSM9DS1_Particle_Library

[17] Loadcell hx711 module GitHub. https://github.com/bogde/HX711 [18] IoT Analytics - ThingSpeak Internet of Things. https://thingspeak.com/. [19] MATLAB - MathWorks - MATLAB & Simulink.

https://www.mathworks.com/products/matlab.html.

[20] R. Thomas Fielding. Ph.D. Dissertation, Architectural Styles and the Design of Network- based Software Architectures, University of California, Irvine, 2000.