Chapter 2

REVIEW OF RELATED LITERATURES AND STUDIES

This chapter presents the relevant literature and studies local or foreign taken

from some books, journals, published thesis and dissertation, and from the internet.

A. Waste management optimization approach

Numerous characteristics in waste management systems, such as the rate of

waste production, the location of the disposal site, the cost of the treatment, and

their relationships, might be unknown and have an impact on the optimization

procedures that go along with them. Different imprecise programming techniques

were used to handle these waste management uncertainty challenges. For instance,

methodologies like fuzzy, stochastic programming, and interval programming

were frequently utilized to address uncertainty-related waste management issues.

The investigation showed that by using these optimization strategies correctly, the

effectiveness of the waste management system may be maximized (Ajay Singh.,

2019)[4].

The investigation showed that by using these optimization strategies correctly, the

effectiveness of the waste management system may be maximized. The

management of municipal solid waste (MSW) has become one of many

governments’ most urgent problems nowadays. With the rise of the bioeconomy,

waste-to-energy (WtE) has expanded beyond its traditional association with

incineration to include any processing method that can produce heat, power, or a

fuel produced from trash. In order to ensure that MSW is managed sustainably

within the framework of the circular economy, further effort must be made to
ensure that the most recent data is obtained. In this regard, we have reviewed

various technologies that are used to exploit MSW as a renewable resource

critically, taking into account their most recent advancements. We have also

considered the important benefits and practical restrictions of energy and material

recycling for sustainable MSW management. Due of its questionable reputation

and social criticism, incineration—the most common method—is currently

challenging to continue applying. Meanwhile, the biological approach offers a

compelling solution for the MSW organic part, which is currently largely

unrecycled and generates disposal concerns. As the bioeconomy gains

momentum, it becomes clearer how environmental biotechnologies should be

coupled and integrated for more sustainable MSW management. The conclusion

of this essay includes developments of potential futures, which can act as a timely

reminder to nudge responsible authorities/researchers to continue enhancing the

current MSW management system (To-Hung Tsui, Jonathan WC Wong., 2019)[5].

Smart technology can play a significant part in solving today’s major population

problems and laying the groundwork for a sustainable future. A clever strategy

offers the chance to integrate knowledge, which is essential to addressing the

major issues facing modern civilizations. Today’s key challenges are ensuring a

balanced economic development of society and reducing the effects of global

warming. It is essential for all engaged engineering professions to work closely

together to achieve multidisciplinary synergies and to solve difficult engineering

issues. A lot of research should be done on topics like efficient energy conversion

technologies, integrating renewable energy systems, enabling the circular

economy framework, integrating processes effectively, and other concerns that are

crucial to the general public (Sandro Nižetic., 2019)[6].

6
Many nations are looking for new strategies to manage trash generation

effectively considering the global waste growth. A participatory strategy that

involves the community in the collection, management, and disposal of garbage is

frequently included in an integrated waste management strategy. Understanding

the dynamics of local communities’ waste disposal is crucial to developing an

integrated solid waste management strategy. For a municipal region in Guyana,

this study includes a qualitative evaluation of waste characterization, home waste

management, recycling willingness, and waste management policy. A community

survey of unlawful dump sites and an online survey of resident opinions were

used to conduct the surveys in municipal administrative Region 10 of the country.

Spatial mapping and distribution counts were used to analyze the characterization

of waste. Binary logistic regression modeling was used for the statistical analysis.

Models indicate that many families use unauthorized dump sites as a result of

ineffective trash collection services. The community’s high level of recycling

willingness, particularly among middle-aged women, is demonstrated by

statistical modeling, although it may be limited by a shortage of recycling

facilities. The nation is looking for a new paradigm in the framework of solid

waste policy. The knowledge required to implement this paradigm will come from

an understanding of the problems and requirements faced by each municipality

within a nation (S. Moonsammy, et. al., 2021)[7].

The expanding problem of municipal solid waste management calls for an

integrated system for trash reduction, collection, composting, recycling, and

disposal. Choosing waste-treatment technologies, locating waste processing and

disposal units, and allocating waste flow to processing facilities and landfills are

just a few of the administrative, operational, and managerial considerations that go

7
 into integrated solid waste management (ISWM). By taking into account

economic, environmental, and social factors, this study seeks to satisfy the

sustainability requirements for building an ISWM system. The Technique for

Order of Preference by Similarity to Ideal Solution (TOPSIS) was used in the

present study to create a trustworthy Suitability Indicator for the location of the

ISWM system’s components. In order to reduce fixed costs, reduce transportation

costs, and increase system suitability, a new model is created. The proposed

model has been shown to be applicable, efficient, and case-unspecific for creating

an ISWM system and assisting the current systems in both developing and

developed nations (Hossein Asefi, Samsung Lim., 2017)[8].

B. Technology-based waste segregation system

In today’s world, waste management is a widespread issue that is only

getting worse as urbanization progresses. A vibrant element of maintaining an

ecological ecosystem is waste management. Sorting at the base level depends

heavily on proper garbage disposal at the dump sites. The old method of sorting

rubbish takes more time and requires more labor. Waste can be sorted using a

variety of techniques and tools. Processing waste materials by applying image

processing to analyze and classify the trash can be highly effective. This project

seeks to analyze recently published studies from all over the world. This will

make it possible to identify the issues, the algorithm employed, and the

methodology of the studies mentioned. It can also evaluate the appropriate

algorithm to apply in a subsequent investigation. These papers discuss the many

techniques and systems that were put forth for waste segregation. These also

discuss the shortcomings of the systems and algorithms they employed that were

8
already in place. It provides several opportunities to generate fresh knowledge for

developing a new system (M. Flores, J. Tan., 2019)[9].

The increasing amount of garbage that each individual produces every second

poses a real problem, though. Incredible 100,000 tons of garbage are produced

daily in India. In addition to making landfills harder to recover due to the aimless

method in which waste is dumped, uncontrolled waste dumping on the borders of

towns and urban neighborhoods has had real ecological consequences in terms of

ground water contamination and contribution to global warming. Our project’s

noble goal is to construct a motorized framework that will help save many lives

and leave the earth cleaner and greener. A programmed waste segregator that aims

to isolate the loss at the transfer level itself is proposed in this work (A. Madankar,

M. Patil, P. Khandait., 2019)[10]. The source itself is a crucial degree of

segregation. Even crude segregation at the household level might mark a turning

point in the nation’s approach to trash management. In this nation, the economic

worth of garbage has not been studied. In this article, an automated trash

segregator that uses capacitive and inductive proximity sensors to separate

garbage into wet and dry at the home level and monitor bin filling status is

proposed (M Vamshi, et. al., 2020)[11].

The amount of garbage produced has dramatically increased as a result of the

growing population's rapid increase in resource consumption. The process of

segregation and recycling is difficult due to ineffective waste management at the

beginning of trash disposal from the home and industrial sectors (R Resmi., 2021)

. Only at industrial scales are different segregation techniques and products

available for handling trash. To appropriately remove the garbage, it must be

handled, separated, transported, and disposed of in order to minimize threats to

9
public safety and the environment’s sustainability. When garbage is sequestered,

its economic value is easier to understand. In the context of small societies, this

study tries to implement an automatic waste segregation system. The waste can be

separated into three categories: wet waste, solid waste, and metal waste using this

simple approach.

An automated trash segregator plays a crucial role in preventing this situation And

making recycling easier. Only after segregation do the importance and economic

worth of garbage become apparent. Such a mechanism for separating metal, dry,

and wet wastes does not yet exist. In order to solve this issue, the idea suggests a

spot automatic waste segregation device. A parallel resonance impedance system

is employed to separate the metallic trash, and capacitive sensors are used to

separate the wet and dry waste. The waste has a higher chance of being recovered

as a result of this activity, and the occupational risks for waste separation

personnel are also decreased (V Deepthi., 2018) [13]. Waste segregation and timely

collection provide a significant barrier for the majority of developing nations.

Many recyclable materials are wasted as a result of ineffective trash segregation.

We introduce SmartBin in this paper, a bin that can automatically notify the waste

collection center when it is full and can separate waste at the source without the

need for human interaction. SmartBin will assist in resolving the issue of waste

segregation and will contribute to the creation of a greener society without

sacrificing safety and hygienic standards (M. Jayson, S. Hiremath, HR. Lakshmi.,

2018)[14].

To reduce threats to the general population and the environment, waste

management processes such as segregation, transportation, handling, and disposal

must be appropriately handled. In our project, “automated waste segregator and

10
 monitoring system,” an effective way to dispose of the garbage has been designed.

In order to allow for the direct delivery of wastes for processing, this study

suggests an automatic waste segregator (AWS), which is a simple, low-cost option

for a segregation system at homes. In order to improve trash management, an

automatic waste segregator is used to separate garbage into three main categories:

plastic, organic, and metallic waste. For monitoring the waste collection process,

ultrasonic sensors have been added. All of the trash cans would be equipped with

the sensors. An alert will be sent to a microcontroller when the rubbish reaches the

sensor’s level (S Sreejith, et. al., 2016)[15].

C. Importance of Convolutional Neural Network (CNN)

One of the most important networks in the field of deep learning is the

convolutional neural network (CNN). The fact that CNN made noteworthy

advancements in a variety of fields, including but not limited to computer vision and

natural language processing, has garnered a lot of interest from both business and

academics in recent years. The reviews that are now available primarily concentrate

on CNN’s uses in various contexts without taking CNN into account on a broader

scale, and recent fresh concepts are not explored (Yang, et. al., 2021)[16].

Convolutional Neural Networks (CNNs) have emerged as the de facto standard for

many Computer Vision and Machine Learning operations over the past ten years.

CNNs are alternating convolutional and subsampling layers feed-forward artificial

neural networks (ANNs). If trained on a sizable visual database with ground-truth

labels, deep 2D CNNs with millions of parameters and numerous hidden layers can

learn complex objects and patterns. They may be used as the principal tool for many

engineering applications for 2D signals, such as photos and video frames, with the

11
right training. However, this could not be an effective alternative in many applications

compared to 1D signals, particularly if the training data is limited or application-

specific. 1D CNNs have recently been proposed as a solution to this problem and

have already attained cutting-edge performance levels in a number of applications,

including the classification and early diagnosis of personalized biomedical data, the

monitoring of structural health, the identification and detection of anomalies in power

electronics, and the detection of electrical motor faults. Another significant benefit is

the feasibility of a real-time and affordable hardware implementation due to the

straightforward and compact setup of 1D CNNs that only carry out 1D convolutions

(scalar multiplications and additions). The overall architecture and operating

principles of 1D CNNs, as well as their main technical applications, are covered in

detail in this paper, with a special emphasis on recent advancements in the field.

Finally, their distinctive qualities are highlighted, capping off their cutting-edge

performance. On a special website, the benchmark datasets and the main 1D CNN

software utilized in those applications are also shared with the public. This work fills

a vacuum in the literature that exists on a review of 1D CNNs and their applications

(Kiranyaz, et. al., 2020)[17].

A unique kind of multilayer neural network or deep learning architecture called a

convolutional neural network (or CNN) is influenced by the visual system of living

things. The CNN is excellent for a variety of computer vision and natural language

processing applications. This chapter’s major objective is to thoroughly examine all of

CNN’s fundamental parts. Additionally, it provides a broad overview of CNN’s

history, recent developments, and some of its most important application fields

(Ghosh, et. al., 2020). It has been shown that conventional neural networks provide an

effective framework for background removal in video captured by stationary cameras.

12
The well-known Self-Organizing Background Subtraction (SOBS) approach and its

neural network-based variations have, in fact, long held the top positions on the vast

CDnet 2012 dataset. Deep learning techniques like convolutional neural networks

have lately and overly been applied for foreground identification, background

initialization, and deep learnt features (Bouwmans, et. al., 2019)[18].

Modern image classification algorithms are more accurate and faster than ever, and

they may be applied to a wide range of tasks, such as security features, face

recognition for authorisation and authentication, traffic identification, and medical

diagnostics, among other things. Different methods can be used to solve the problem

of image classification. But among all of them, machine learning algorithms are the

most effective. These algorithms are based on a notion that was proposed in the past

but was never able to be put into practice due to a lack of processing capability. The

models are better trained with the concept of deep learning and are able to recognize

various degrees of visual representation.

Convolutional neural networks changed this subject by learning the fundamental

shapes in the initial layers and developing to learn the picture attributes in the deeper

layers, leading to more precise image classification. Convolutional neural networks

were developed as a result of Hubel and Wiesel’s 1962 work on the hierarchical

representation of neurons; this work was based on the examination of stimuli for the

cat’s visual cortex. Understanding how the visual brain in humans and animals

functions was a major advancement in the science of computer vision. With the use of

a convolution neural network and the idea of deep learning, an image’s feature is

extracted in this article. For diverse purposes, more classification methods are put into

practice (S Apoorva, et. al., 2018)[19]. So that waste may be more easily recycled,

machine learning techniques can be used to sort rubbish into various recycling

13
categories. In this paper, a deep learning algorithm is used to develop an autonomous

waste-classification system that can separate waste into recyclable, non-recyclable,

and metal categories. The convolutional neural network architecture was used in this

computer vision method to accomplish the classification in real time so that the

garbage could be deposited into the appropriate chambers as soon as it was tossed into

trash cans.

This paper explains backpropagation before introducing convolutional neural

networks by starting with multilinear perceptrons. Using this, we move on to discuss

convolutional neural networks, define the term, and offer useful instructions for

training these networks for classification and segmentation tasks (Jonas Teuwen,

Nikita Moriakov., 2020)[20].

Synthesis

According to, R. Resmi, et. al., (2021)[21]. The amount of garbage produced has

dramatically increased as a result of the growing population’s rapid increase in

resource consumption. The process of segregation and recycling is difficult due to

ineffective waste management at the beginning of trash disposal from the home and

industrial sectors (R. Resmi, et. al., 2021) [22]. Proper waste management is one of the

most important components of an urban development plan. Waste collection is a very

difficult process that requires effective management of the entire system, from waste

collection to hygienic waste disposal. Separating collected debris is crucial since

dumping polythene bags, used furniture, and e-waste together in landfills runs the risk

of releasing hazardous chemicals that could contaminate the surrounding area. The

wet waste component is either composted or turned into methane gas. Biogas can be

14
utilized as a source of energy, and compost can replace the need for chemical

fertilizers. The scrap metal, plastic, and paper can all be recycled or reused. The most

fundamental prerequisite for starting the management process is an automated waste

segregation process. We have therefore compared numerous automated waste

segregation procedures that have been deployed using diverse technologies in this

paper (N. Punjabi, et. al., 2019)[23]. Waste should be separated in order to maximize its

economic worth. At dump sites, segregating the waste requires additional time and

labor. An automatic waste segregator (AWS), which is a low-cost, simple-to-use

solution for a segregation system at homes, is proposed in this work. The AWS

distinguishes between wet and dry waste using resistive sensors and an inductance

sensing device for metallic objects.

Ensuring waste segregation and resource recovery is essential to effective waste

management. Waste is typically divided into biodegradable and non-biodegradable

categories. The fact that waste is not separated before collection and is nonetheless

wasted presents a significant obstacle in this regard. These end up as enormous

mountains of rubbish at dump yards and need to be separated. The Municipal

Corporation now separates waste by hand. For those who must carry out this task, this

leads in unhygienic working conditions. Despite having the required tools, employees

still run the danger of getting sick from the waste they handle. By lowering health

risks, automation of this process will benefit those who work on it (R Anitha, et. al.,

2021)[24]. This is an alternative for a municipal trash segregator that can separate the

deposited waste on a fundamental level, hence lowering the need for more

sophisticated machinery. So, this is a smart trash that can spread waste without

assistance from humans. This project also sends a message on the status of the trash

can when it is full in addition to sorting.

15
Here, we recommend the use of an Automatic Waste Segregator (AWS), a convenient

and affordable method for sorting household waste. The trash is to be divided into

three groups, namely. Waste that is moist, dry, and metallic. The system uses an LCD

display to show the results of the segregation and a moisture sensor to distinguish

between wet and dry waste. It also uses an inductive proximity sensor to identify

metallic waste. Experimental investigations have shown that trash segregation

utilizing AWS has been effective (Gangal, et. al., 2019)[25], In order to achieve waste

segregation, we have suggested a completely automated waste management system in

this article. The technique used is a system that can separate municipal waste into

Organic and Recyclable waste using computer vision and deep learning in

conjunction with the internet of things (IOT). By limiting the acquisition and

transmission of transmissible diseases, the elimination of manual segregation in the

waste management process greatly lowers the danger to the health of municipal

workers. The trash segregation process will be substantially faster and less expensive

thanks to automation. The goal of this study was to generate and implement novel,

environmentally friendly solutions for efficient waste management systems that

require little to no human involvement (S Yadav, et. al., 2021)[26].

16
 Conceptual Framework

To illustrate the study to be conducted, the research paradigm is presented on the

below.

Input Process Output

IR sensor Making a waste bin Automated waste

Figure 2. Research Paradigm

Figure 2. Shows the research paradigm the study comprising camera, inductive

sensor module, and the infrared proximity sensor module are all housed in the

enclosure. The IR proximity sensor module is engaged when trash is placed in the

can, which then turns on the camera and inductive sensor. The item falls into a flap

after sliding down an incline, where an inductive sensor and camera sort the trash into

one of four categories: metal, paper, plastic, and other debris. Data about the sort of

waste, the time it was thrown away, and the date are submitted to the database once a

classification has been made. Three servo motors are used to drop and sort waste into

the appropriate bins below. The first servo motor drops the waste into the sensor

region, and the other two servo motors transport the garbage into one of four

receptacles at the bottom third of the container. An ultrasonic sensor is used to gauge

17
the amount of trash in each container once it has been dropped into its corresponding

one. The database’s information on waste container fullness is then updated using the

new data. On the front face of the segregator, there are four RGB LEDs.

Definition of Terms

In order to facilitate a common understanding of the information conveyed in

this research study, these terms are hereby defined operationally and theoretically. For

clarification, the important term used in the study has been defined.

Inductive sensor- this is the sensor that can identify the metallic waste

Infrared sensor- this is the sensor that will trigger the camera and the sensors that

there is a waste has thrown.

Ultrasonic sensor- this is the sensor that will indicate the amount of waste through its

height and it will be showed by the lights in the front face of the machine.

Arduino uno- this is the open-source board which will use for the programming of

the inputs to set them into their different functions.

Waste- an object that will be thrown in the machine which will be categorized by the

sensors that will fall into the target receptacles.

Automation- this is the necessary process that needs to be done to make the trash bin

optimized or connect the trash bin in technology either.

Servo motor- a machine that fixed on the flap to fall the thrown waste into its

receptacle.

Capacitive proximity sensor- this is the sensor that will identify the plastic waste

that has thrown.

Smart optical- this is the camera that can capture the waste and for the details of

waste where this waste threw and what kind of waste it is.

18
Classifications- it refers to the texture or materials that used to make the particular

waste which will be detected by the sensors.

RGB light- this is the light that used as indicators when the specific receptacle is

already full of waste, the connected light in which classification will be illuminated in

red to indicate the full receptacle inside.

19