Chapter 1

Introduction
In recent times, garbage disposal has become a huge cause for concern in the world. A
voluminous amount of waste that is generated is disposed by means which have an adverse effect
on the environment [1:11].
The common method of disposal of the waste is by unplanned and uncontrolled open dumping at
the landfill sites. This method is injurious to human health, plant and animal life.
This harmful method of waste disposal can generate liquid leachate which contaminate surface
and ground waters; can harbour disease vectors which spread harmful diseases; can degrade
aesthetic value of the natural environment and it is an unavailing use of land resources[2:4].
In India, rag pickers play an important role in the recycling of urban solid waste. Rag pickers and
conservancy staff have higher morbidity due to infections of skin, respiratory, gastrointestinal
tract and multisystem allergic disorders, in addition to a high prevalence of bites of rodents, dogs
and other vermin. Dependency on the rag-pickers can be diminished if segregation takes place at
the source of municipal waste generation.
The economic value of the waste generated is not realised unless it is recycled completely.
Several advancements in technology[3] has also allowed the refuse to be processed into useful
entities such as Waste to Energy, where the waste can be used to generate synthetic gas (syngas)
made up of carbon monoxide and hydrogen. The gas is then burnt to produce electricity and
steam; Waste to Fuel, where the waste can be utilized to generate bio fuels.
When the waste is segregated into basic streams such as wet, dry and metallic, the waste has a
higher potential of recovery, and consequently, recycled and reused. The wet waste fraction is
often converted either into compost or methane-gas or both. Compost can replace demand for
chemical fertilisers, and biogas can be used as a source of energy. The metallic waste could be
reused or recycled.
Even though there are large scale industrial waste segregators present, it is always much better to
segregate the waste at the source itself. The benefits of doing so are that a higher quality of the
material is retained for recycling which means that more value could be recovered from the
waste[3]. The occupational hazard for waste workers is reduced. Also, the segregated waste

could be directly sent to the recycling and processing plant instead of sending it to the
segregation plant then to the recycling plant.
Currently there is no system of segregation of dry, wet and metallic wastes at a household level.
J.S. Bajaj[4:12] has recommended that a least cost, most appropriate technological option for
safe management should be developed. The purpose of this project is the realization of a
compact, low cost and user friendly segregation system for urban households to streamline the
waste management process.
The mixed waste is sorted based on the following methods at the industrial level[5]. Larger items
are removed by manual sorting. Then the refuse is sorted based on its size by using large rotating
drums which is perforated with holes of a certain size. Materials smaller than the diameter of the
holes will be able to drop through, but larger particles will remain in the drum.
For metallic objects electromagnets or eddy current based separators can be used. Near infrared
scanners are used to differentiate between various types of plastics based on the ability of the
material to reflect light. X-rays can also be used to segregate materials based on their density.
The methodology adopted in this paper to resolve the issue of waste segregation is by making the
entire process automated and to the reduce cost such that it could be adapted in a household
level.
Waste is pushed through a flap into the proposed system. An IR proximity sensor detects this and
starts the entire system. Waste then falls on the metal detection system. This system is used to
detect metallic waste. After this the object falls into the capacitive sensing module. This module
distinguishes between wet and dry waste. After the identification of waste, a circular base which
holds containers for dry, wet and metallic waste is rotated. The collapsible flap is lowered once
the container corresponding to the type of garbage is positioned under it. The waste falls into the
container and the flap is raised. The waste in the containers now can be collected separately and
sent for further processing.

Chapter 2
Literature review
This section will look into different existing devices that can compare with the proposal. It will
entail several important features that in one way or another help characterize the desired
hardware and software design of the SMART Bin. Sections 2.1 and 2.2 will entail reviews on
automated trash bin systems and garbage collection schemes, respectively. Another section will
discuss about existing self- or controlled-navigations scheme while an additional section will talk
about mobile robots which may serve as a comparison to the desired design and implementation
applicable to the study.
2.1 Automated Trash Bin Systems
Trash bin systems are usually designed on the software level and handled manually. As an
illustration, many cities employ garbage collection by distributing government-issued trash bins
that are manually used and they only use automated systems in canvassing bin collection. It is
then desired in this study to make a similar approach at the Ateneo community. However, it will
be in a smaller-scale and will only occupy a single floor at the College Building. A software
program will be designed to automate a robot in navigating through this particular floor while
disposal of garbage will still be done manually.
But in recent years, automated trash bins existed in the forms of sensor-based trash can
covers and improvised step-open-cap trash bins. Patents in the past decade include the
development of the miniature robots that are limited to recognizing trash using image processing
and employing complicated algorithms to make trash identification possible. Aside from the fact
that this approach is costly, it is also

undesirable in the studys locale wherein available

technology is limited. It is nonetheless important to note that it is possible to substitute costly

components as well as complicated algorithms or programs in achieving the same objective. This
is what the study will emanate from these existing devices. At some institution such as the
Defense Advanced Research Projects Agency (DARPA), a related process to garbage collection
has been patented back in 2003. Barrett (2003) claims a better memory management that
recycles wasted memory by collecting them in a systematic pattern. The concept is similar to an
individual robot recognizing and collecting waste that it encounters. Such learning capacity can
be greatly beneficial when achieved by a robot; however, to apply Barretts research in this way

would require a large memory for the robot. As such, this study will not look into developing a
learning device.
2.2 Garbage Collection Schemes
Collection schemes are not as automated as in garbage collecting robots. As a matter of fact,
most automated collection schemes are implemented in the computer systems level and not in the
desired individual robots.
Multimode garbage collection system is an instance of this implementation. Patented by
OConnor et al in 2005, it discusses garbage collection in distributed computing systems. It is
divided into types of garbage collection which can be done at different times and according to
necessity. Each mode is separated by a unique code based on the routine applicable to each type
of collection. If used in this study, it will demand more complicated codes that may also consume
much memory space and resources available for designing the device. Distributed computing
systems may not also be appropriate in the locale of the SMART Bin application. Once again,
since the SMART Bin is not a learning device, it may as well resort to plainly receiving disposed
garbage rather than identify collections scheme. However, what is available and patented so far
includes relevant fields. A control system for mobile robots has been patented in 2007 by Jones
and Mass. In this patent they claim that controlling robots can be possible in multiple ways such
as following a detected obstacle nearby and a random bouncing motion. It relates to this study by
showing that collection path can be programmed or patterned after these control presented by
Jones and Mass. Similarly, obstacle detection as a basis for device movement is highly regarded
in this study. Instead of resorting to random motion, the collection scheme will be patterned to
the location of the offices where the SMART Bin will be collecting garbage from.

2.3 Navigation Schemes

This study aims to allow the device to navigate on its own and several other devices have
achieved such feat in a variety of ways. In a patent filed by Ramshur (2012), a wire sensor is
used to allow a garbage conveyor to transfer the bins from the garage to the curb where the
garbage truck collects them. It is fixed and efficient as a navigation technique since the conveyor
device follows a set wire which is the only input for the sensor. Although it can be achieved in
this study by creating a line-following robot, painting a line along corridors may not be allowed
by the school administration.

The design can resort to a similar approach. For example, a wall-following robot cleaner has
been designed by using the concept of sensors and algorithms. Hong
and his team (2008) designed a cleaning robot to automatically do its primary function while
traveling on its own by a systematic and simple idea of establishing a reference wall and
calculate its own path to clean a region. Although the latter capacity is likely not to be used in the
software design, but to use a wall as a reference for navigation will be used. Moreover, the
choice of wheels and motors will be important. More importantly, it is very crucial that the
choice fits the navigation constraints of the contraption. There are two options structure-wise,
whether a single motor controls the both wheels on either side of the device or each wheel is
controlled by a separate motor. Servo motors can be used and rotation of the device will depend
on its computed direction. On the other hand, some devices use a motor for each wheel. To
rotate, one motor will rotate relatively faster than the other one (Ramshur, 2012). Motor
controllers are also used in some robotic devices. These are commercially available components
with serial controllers which receives its commands from a computer, thereby controls motor
speed for each individual motors (Lakshmanan, et al, 2008). Also part of the navigation is
detection of hindrances along the devices path. For some existing devices, image processing is
used but it proves to be costly. Another option uses a proximity sensor for detecting external
obstacles such as in the lawn mower with sensors patented by Bernini (2009). There is a problem
if these sensors are used in the study. These sensors are based from infrared technology which is
sensitive to sunlight. In fact, due to the presence of sunlight, infrared sensors may commit errors
in detecting objects.
For this reason, the proponent resorted to ultrasonic sensor that is not limited to the presence of
infrared. It is accurate by using ultrasonic waves that are invulnerable to any noise that may alter
measurements. Such sensors have been used in similar studies such as the Tailgator that is
designed by Nayak and his team (2009). The use of sonar in his study was to detect the trash and
establish a path towards this detected trash. On the other hand, the sensors to be used for the
SMART Bin will be used to detect any objects that may hinder the travel path and it will also be
used to establish a path by detecting the wall present in the vicinity of the robot.

2.4 Mobile Robots

In Peccioli, Italy, several types of garbage collecting robots are designed by researchers at Scuole
Superiore Sant-Anna University to collect garbage from homes and sweep the streets. Since the
streets in this town are narrow for the garbage trucks to pass, the researchers designed these
robots, called Dustcart, to replace them and to be responsive to users through SMS (Grifantini,
2010). It is similar to the SMART Bin in the way that both will navigate on its own to collect
trash. Chen, et al (2010) designed an automated garbage collector robot which aims to collect
intelligently trash and other waste that it will find along its path. It was built for household use
and to be as competent as other existing garbage collecting robots. What was limited among
existing household garbage collecting robots at that time is the limitation in collecting larger
wastes.
Both these indoor and outdoor garbage collecting robots uses sophisticated navigating and
garbage identification modules to achieve their functions. Camera data and search algorithms are
of common use in these systems. This study, however, aims
to use simpler approach to navigation and garbage collection. Using the modules of these related
studies will defeat the purpose of our objectives. A simpler project design was made by Sebelah
(2009) that simply collects trash on the floor and keeps the trash from falling back to the floor.
The device uses three types of sensors instead of using cameras: ultrasonic, light, and sound.
These three sensors aim to respond to different inputs such as detection, determination, and
command. However, the project lacked accuracy in moving towards the detected garbage. It can
be compared to the study of Nayak and his team as both designed robots that navigate upon
detection of waste.
For the Dustcart, Chens, Nayaks and Sebelahs automated garbage collector robots, the
primary limitation of theirs conceptual frameworks is trash had been disposed before being
collected. It is desired in our scheme however that a robot will navigate around to collect waste
directly when they are disposed.

Chapter 3
Design considerations
Figure shows a diagram of the AWS. An upper enclosure ensures waste does not fall out of the
sensing area. Inside the enclosure is an infra-red (IR) proximity sensor module. When the waste
is dumped in by pushing the flap, the IR proximity sensor module gets activated and brings the
micro controller MSP430G2553 out of low power mode. The object slides over the incline to fall
on the inductance coil which is used to sense any metal object. If the object is metallic a change
in parallel resonant impedance of the metal detection system is observed. The object continues
and drops into the capacitive sensing module. Here, a decision is made if the waste is wet or dry
based on its relative permittivity. Two DC geared motors are used to perform the final
segregation based on the identification. One motor moves a circular base to get the
corresponding container under the collapsible base. This collapsible base is controlled by the
second motor. The process flow analysis is given in Figure. The individual modules are
explained here as follows.

Figure1 Waste Segregator

A. Entry System and Initialization

The waste is dumped into the AWS by pushing it through a flap. This flap comes in the proximity
of the IR proximity sensor which marks the entry of the waste. The sensor sends an interrupt to
the microcontroller which comes out of the low power mode. It then initializes the sensor
modules. The initialization of all modules ensures that any dynamic changes in the environment
do not affect the sensing. The sensor modules establish a base count by averaging many samples,
while the waste slides over the first incline. An average of 100 counts is taken to establish a base
count for the LDC1000. The object takes 240 milliseconds to slide down the incline while 100
samples take 200 milliseconds[6:12]. Once, an average for the metal detection system is
established the LDC100 is switched to threshold mode and the capacitive sensing module
initializes its base value by averaging 15 samples. Thus, averages of all base count values are set.

B. Metal Detection System

The object moves over the incline and falls on the inductive coil. The inductive coil is a part of a
parallel inductance and capacitance (LC) circuit. As shown in Figure 2 the circuit is connected to
the LDC1000 inductance to digital convertor. This measures the parallel resonance impedance of
a parallel LC circuit and returns data as a proximity value. This data changes whenever another
metallic object is introduced in the vicinity of the coil.
When an alternating current is passed through a coil it generates a magnetic field. When a
metallic object is introduced in the vicinity of the coil, eddy currents are induced on its surface.
The eddy currents are a function of the distance, size, surface area and composition of the target.
This generates a magnetic field which opposes the original magnetic field which is generated by
the coil. The inductive coupling between the coil and the object creates a mutual inductance
effect on the coil which decreases the parallel resonant impedance of the circuit which in turn is
reflected by an increase in the proximity count value. Magnetic fields do not affect the metal
detection system. It can detect any conducting material irrespective of its magnetic
properties[6:8].
An average of base count is set up by using LDC1000 in data ready mode which generates an
interrupt every time the proximity value is ready to be read[6:21]. After a base count is
established the LDC1000 is switched to threshold mode which generates an interrupt once the
proximity data crosses a set threshold[6:20]. Hence, the object is inferred as metallic. The waste

continues down the second incline towards the apex. If the type of garbage is not metallic then
the capacitive sensing module continues to sense the object, else the sensing module is stopped
and the actuators are activated.

C. Capacitive sensing module

A base count value is established once the LDC1000 has been switched to the threshold mode.
Pin Oscillator method determines a count value for each pair of capacitive plates[7:330]. Three
pairs of copper plates are placed along the walls of the structure which are inclined to each other
at an angle of 45. This arrangement is made to ensure that waste of all sizes can be sensed. The
area of each pair of plates increases as it moves away from the apex of the structure. The
sensitivity of the plate decreases with its increase in area, hence smaller plates would accurately
sense objects of smaller size. Even though the sensitivity of the larger plate is decreased, it is
designed to detect larger objects which will yield a change sufficient to be identified.
The principle of working for the pin oscillator method in the microcontroller is shown as in
Figure 4. The microcontroller uses two timers. One timer is used to control a fixed time window,
while the second timer uses the output of the schmitt trigger as clock. The input of the Schmitt
trigger is connected to a resistor and a plate of the capacitor. The other plate of the capacitor is
connected to ground. The second terminal of the resistor is connected to a multiplexer output
which supplies the rail voltages alternately based on the output of the Schmitt trigger. As a result,
the capacitor charges and discharges and the Schmitt trigger toggles which generates a train of
pulses at the output. These pulses are applied to the second timer as a clock input. This process
continues till the first timer counting window is complete. The count value in the second timer is
the capacitive count value. When an object is introduced between the plates, the capacitance
increases and thus the capacitor takes a longer time to charge and discharge. Hence the count
value stored in the second timer is lower than the base count value. The difference between the
two count values is used to identify whether the waste is dry or wet.
The property used for segregation of waste is the relative dielectric constant. Once a dielectric is
introduced between the plates of the capacitor the capacitance increases. Wet waste has a higher
relative dielectric constant than that of dry waste because of the moisture, oil and fat[8], content
present in kitchen waste. If the change in the capacitive count is greater than threshold then the

type of garbage is inferred as wet waste else it is dry waste. Since the capacitance value of the
plates is different, the change observed for the same object by the different plates is different.
Hence different threshold levels are assigned for each pair of capacitors. Thus, the type of waste
is identified as either wet or dry and the actuators are activated.

Chapter 4

Detail Design
Components
LCD display
A liquid crystal display (LCD) is a flat panel display, electronic visual display, based on on
Liquid Crystal Technology. A liquid crystal display consists of an array of tiny segments (called
pixels) that can be manipulated to present information. Liquid crystals do not emit light directly
instead they use light modulating techniques.
LCDs are used in a wide range of applications, including computer monitors, television,
instrument panels, aircraft cockpit displays, signage, etc. They are common in consumer devices
such as video players, gaming devices, clocks, watches, calculators, and telephones.
LCDs are preferred to cathode ray tube (CRT) displays in most applications because of
1. The size of LCDs comes in wider varieties.
2. They do not use Phosphor; hence images are not burnt-in.
3. Safer disposal
4. Energy Efficient
5. Low Power Consumption
It is an electronically modulated optical device made up of any number of segments filled with
liquid crystals and arrayed in front of a light source (backlight) or reflector to produce images in
color or monochrome.

Figure2 Reflective twisted pneumatic liquid crystal display.

1. Polarizing filter film with a vertical axis to polarize light as it enters.
2. Glass substrate with ITO electrodes. The shapes of these electrodes will determine the
shapes that will appear when the LCD is turned ON. Vertical ridges etched on the surface
are smooth.
3. Twisted pneumatic liquid crystal.
4. Glass substrate with common electrode film (ITO) with horizontal ridges to line up with
the horizontal filter.
5. Polarizing filter film with a horizontal axis to block/pass light.
6. Reflective surface to send light back to viewer. (In a backlit LCD, this layer is replaced
with a light source.)
JHD162A is one such LCD which is used here.
It has a Panel with 2 rows and 16 columns and with blocks as shown below with 5x8 pixelselection pattern.

Operating Voltage

Pin Details

Timing Diagram

Write Mode

Read Mode

Pixel selection pattern

Interfacing JHD162A with Microcontroller

Alpha Numeric displays form an integral part of the Embedded Systems.
The Data displayed here is controlled by the Microcontroller.
The Control pins like Read Strobe, Read/Write and Enable are controlled through the
Microcontroller Ports as per the waveforms above.
The 8 data is also provided through a Microcontroller Port.

General Circuit and settings of jhd162a

10K

VR 1

ALC D _R S
A L C D _ R /W
ALC D _E
D B0
D B1
D B2
D B3
D B4
D B5
D B6
D B7

+5V
U 1
VSS
VD D
V0
R S
R /W
E
D B0
D B1
D B2
D B3
D B4
D B5
D B6
D B7
LED A
LED K
16x2 ALPH A LC D

ALPHA NUMERIC LCD

Make sure that 5V and GND lines are properly connected otherwise you may end up in
damaging parallel port.
If you want backlight than connect pin 15 of LCD to 5V and pin 16 of LCD to GND. By
adjusting 10k resistor make pin 3 of LCD at 0V. If connections are proper you will see this after
power on.

Flow Chart

Connect VDD, VSS, LEDA,LEDK

Pins for Power and Contrast
Adjustments
RS , EN, R/W And data Pins
Assigned to Ports

EN=1, RS=1 , R/W=0 for data

write

Data in DB0-7 is sent through Port

Pins

Data is Displayed on the Panel

LCD commands:

Light dependent resistor

Working:
Light-dependent resistor alternatively called an LDR, photo resistor, photoconductor, or
photocell, is a variable resistor whose value decreases with increasing incident light intensity.
An LDR is made of a high-resistance semiconductor. If light falling on the device is of high
enough frequency, photons absorbed by the semiconductor give bound electrons enough energy
to jump into the conduction band. The resulting free electron (and its hole partner) conduct
electricity, thereby lowering resistance.
A photoelectric device can be either intrinsic or extrinsic. In intrinsic devices, the only available
electrons are in the valence band, and hence the photon must have enough energy to excite the
electron across the entire band gap. Extrinsic devices have impurities added, which have a
ground state energy closer to the conduction band - since the electrons don't have as far to jump,
lower energy photons (i.e. longer wavelengths and lower frequencies) are sufficient to trigger the
device.

Two of its earliest applications were as part of smoke and fire detection systems and camera light
meters. Because cadmium sulfide cells are inexpensive and widely available, LDRs are still used
in electronic devices that need light detection capability, such as security alarms, street light
controller.
Proximity sensor:
A proximity sensor is a sensor able to detect the presence of nearby objects without any physical
contact. A proximity sensor often emits an electromagnetic or electrostatic field, or a beam of
electromagnetic radiation (infrared, for instance), and looks for changes in the field or return
signal. The object being sensed is often referred to as the proximity sensor's target. Different
proximity sensor targets demand different sensors. For example, a capacitive or photoelectric
sensor might be suitable for a plastic target; an inductive proximity sensor requires a metal
target. The maximum distance that this sensor can detect is defined "nominal range". Some
sensors have adjustments of the nominal range or means to report a graduated detection distance.

Proximity sensors can have a high reliability and long functional life because of the
absence of mechanical parts and lack of physical contact between sensor and the sensed object.
Sensors positioned close to their moving targets are far more likely to suffer damage from
physical contact. With machinery and equipment becoming increasingly compact, there isn't
always room for additional mounting brackets or assemblies to properly hold sensors in position.
A sensor that shifts or vibrates in position is more prone to collision damage. In this instance,
many users prefer to use "extended-range" sensors to reduce the possibility of damage.

Driver circuits
Here the driver circuits are used to control the operations of firing unit, laser unit & audio
reception unit present on the robotic module. Here three types of driver circuits are used they are
ground driver, laser driver and motor driver circuit.

100RPM 12V DC motors with Gearbox

6mm shaft diameter with internal hole

125gm weight

Stall Torque = 1.5kgcm torque

No-load current = 60 mA(Max), Load current = 300 mA(Max)

Motors
NR-DC-ECO is high quality low cost DC geared motor. It contains Brass gears and steel
pinions to ensure longer life and better wear and tear properties. The gears are fixed on hardened
steel spindles polished to a mirror finish. These spindles rotate between bronze plates which
ensures silent running. The output shaft rotates in a sintered bushing. The whole assembly is
covered with a plastic ring. All the bearings are permanently lubricated and therefore require no
maintenance. The motor is screwed to the gear box from inside.

Figure: 12v 100 rpm DC Geared Motor

L293D
The Device is a monolithic integrated high voltage, high current four channel driver designed to
accept standard DTL or TTL logic levels and drive inductive loads ( such as relays, solenoids,
DC and Stepper motor) and switching power transistors.
To simplify use as two bridges each pair of channels is equipped with an enabled input. A
separate supply input is provided for the logic, allowing operation at a lower voltage and internal
clamp diodes are included.
This device is suitable for use in switching applications at frequencies up to 5 KHz.
The L293D is assembled in a 16 lead plastic package which has 4 center pins connected together
and used for heat sinking. Since L293D is an integrated circuit motor driver it can be used for
simultaneous bidirectional control of two small motors. L293D is limited to 600 mA.

Block diagram of L239 motor driver

L293D is an H-bridge IC used to drive the DC motor. This driver is also used to vary the speed
of the DC motor, cpu fan and to control the brightness of the led light. L293 requires +12V, and

+5V power supply. It has a dual channel ie we can control 2 motors with single IC. For every
channel there is a 1 enable pin (connected to +5V) and 2 input pins are connected to
microcontroller port __ and __ respectively. The second channel 2 input pins are connected to
microcontroller port __ and port __ respectively.
To control the speed of the DC motor, CPU fan or to control the brightness of the led
light, the enable pin will be connected to the PWM pin of the microcontroller.

Figure : Process flow of the Automated Waste Segregator

Features

General-purpose register: 8 bits 32 registers (8 bits 8 registers 4 banks)

ROM: 512 KB, RAM: 32 KB, Data flash memory: 8 KB

On-chip high-speed on-chip oscillator

On-chip single-power-supply flash memory (with prohibition of block erase/writing

function)

On-chip debug function

On-chip power-on-reset (POR) circuit and voltage detector (LVD)

On-chip watchdog timer (operable with the dedicated low-speed on-chip oscillator)

I/O ports: 16 to 120 (N-ch open drain: 0 to 4)

Timer 16-bit timer: 8 to 16 channels, Watchdog timer: 1 channel

Different potential interface: Can connect to a 1.8/2.5/3 V device

8/10-bit resolution A/D converter (VDD = EVDD =1.6 to 5.5 V): 6 to 26 channels

Power supply voltage: VDD = 1.6 to 5.5 V

Part drawings and assembly

Induction coil

Dustbin

Motor base

Assembly of waste segregator

Results

Table : Look up table for the rotation of the base control motor

Table : Proximity count values for different metal objects

Figure: Change in proximity count value plotted for metal objects