Smart Trash Collecting and Segregating Robot

Smart Trash Collecting and Segregating Robot
Sanjay B
Electronics and Communication Engineering
Velammal College of Engineering and Technology
Madurai
Sanju Vikasini RM
Electrical and Electronics Engineering
Velammal College of Engineering and Technology
Madurai

Abstract— The Earth is not a garbage can to litter everywhere.
Pollution is a serious issue that goes alarming every now and
then. Pollution increases subsequently with increase in
population. Garbage has to be strictly discarded to avoid
pollution. Here on emphasizing land pollution, “Garbage
Collecting Robot” can clean up all the garbages that are thrown
on the roads using Raspberry Pi. This robot will be built on a
metallic base and powered by a battery and contains a Pi cam to
segregate wastes on image processing. The bot will contain a
sensor for obstacle detection and hence has a good path
determination. On implementation, manual requirement can be
avoided and thus prevents people from sanitary issues.
Keywords— Image processing, Raspberry Pi, Recycling,
Robot, Ultrasonic Sensor.
1.INTRODUCTION
“Swacch Bharat Abhiyan”, a need for a clean India is a
dream that is yet to be fulfilled. Land pollution is one of the
major problems that we face especially in cities, malls and
even places like railway stations and airports. Sanitary
workers who clean up the garbages are tend to be exposed to
various skin ailments and health issues. Hence these manual
workers must be replaced by a bot which can be in turn
controlled by the workers to ensure employment doesn’t fail.
2.LITERATURE REVIEW
2.1) In a system that collects garbage automatically through
line. For which we are designing an arena so that the machine
will follow the specific path and also manually by using
Arduino and Bluetooth module. The machine can be
controlled by software which will give command to the
robotic jaw to collect the stationary waste. It also has the
characteristic to get controlled by voice command which
makes the robot to collect the garbage according to the given
command. Findings: We have designed a semi-autonomous
garbage collector robot which can do multiple functions. This
robot has one robotic arm in which it can pick the garbage
and dispense it in main basket attached to the robot. The
camera placed on robot helps the administrator to remotely
monitor the robot while collecting garbage. The prototype has
an electronic mechanism by which the robot can dispense it
collected garbage to the dispensing point. This robot has
installed batteries in which there is no fuel or electricity
required to complete the operation.
2.2) In an existing system, they have combined the usage of
proximity sensors with computer vision to accomplish our
purpose. The camera, mounted on the chassis of the robot,
remains inactive until an object is detected by the proximity
sensor. When an object is detected, the camera captures a
single image and sends it to the server for image processing
classification to detect it as garbage or non-garbage. The
advantage of this method is that a lot of time, memory and
power is saved by performing processing on a single image,
rather than continuously performing processing on every
frame of a live video feed. Thus we can solve this purpose by
using low end processors and hence the hardware cost shall
be substantially reduced. For classification of the image, we
can use two approaches: A. Perform image processing on the
robot processor itself: This will require a moderate cost
processor to be on the body of the robot. There may be some
lag in output since processor won’t be as high end. But
because of the same reason, the cost will be less. B. Send
images from multiple robots on a central server: The central
server will use a high end processor while the robot
processors can be extremely low end. The robots will simply
capture images and send them wirelessly to the server.
2.3) The robot Garbage collection system consists of a set of
a ARM connected to the motors. The mechanism will not
operate for entirety of the vehicle operation and will rotate
only for predetermined set of conditions. The main aim of the
mechanism is to collect garbage which is of similar
dimensions to that of juice cartons, crushed papers, and all
light items whose height is between 5 to 10 cms. Mechanism
is mounted on the front side of the base with an appropriate
ground clearance. When the sensor detects an obstacle, the
image processing is used to categorized the object as garbage
or any living organism .The object detection is specifically
used for safety of animals ,so that they don’t get harm . The
garbage is pick and drop into a bin which is placed right
behind the mechanism. The robot keeps collecting the
garbage until it reaches certain height in the bin. Once the bin
is filled the collected garbage is disposed to a selected place.
3.PROPOSED SYSTEM
The bot can travel in any land terrain and operates on a
battery. Commercial stainless trash cans are used to dump the
waste. It will also contain a scrubber at the bottom to sweep
along the way it travels. The motion of the bot is controlled
by Python programming and Raspberry Pi 3b+. The bot
platform is controlled by a 24v DC geared motor and rubber
wheels for smooth motion. The bot is trained to segregate bio
and non-bio degradable wastes separately on different trash
cans using image processing. Precision long range ultrasonic
distance sensors can be used to know the distance from
obstacles around so that the robot can autonomously navigate
around. A GPS module on board can be used for getting the
accurate location of the robot. This helps to find the main
garbage bin in which the location of it is pre saved. These
robots can have a load cell along with a proximity sensor to
tell the amount of waste in the trashcan.
3.1 COMPONENTS :
i) Raspberry Pi 3b+ :
International Journal of Engineering Research & Technology (IJERT)
http://www.ijert.org ISSN: 2278-0181
IJERTV9IS110152
(This work is licensed under a Creative Commons Attribution 4.0 International License.)
Published by :
www.ijert.org
Vol. 9 Issue 11, November-2020
325
The Raspberry Pi 3 Model B+ is the latest product in the
Raspberry Pi 3 range, boasting a 64-bit quad core processor
running at 1.4GHz, dual-band 2.4GHz and 5GHz wireless
LAN, Bluetooth 4.2/BLE, faster Ethernet, and PoE capability
via a separate PoE HAT. The Raspberry Pi is a low
cost, credit-card sized computer that plugs into a computer
monitor or TV, and uses a standard keyboard and mouse. It is
a capable little device that enables people of all ages to
explore computing, and to learn how to program in languages
like Scratch and Python.

ii) Ultrasonic Sensor:
The ultrasonic sensor measures the distance using
ultrasonic waves. The sensor head emits an ultrasonic wave
and receives the wave reflected back from the target. Hence
the distance to the target can be measure by measuring the
time between the emission and reception. Hence it is much
effective in detecting an obstacle that is encountered on the
path of the bot.
Distance L = ½ x T x C

iii) ADC ADS1115 Module :
The ADS1115 are great analog to digital converters that
are easy to use with the Raspberry Pi using its I2C
communication bus. The the ADS1115 is a higher
precision 16-bit ADC with 4 channels. It has a
programmable gain from 2/3x to 16x so you can amplify
small signals and read them with higher precision.

iv) Proximity Sensor:
Proximity sensors are sensors that detect presence of objects
without physical contact and relay that information captured
into an electrical signal. Proximity sensors are suitable for
damp conditions and wide temperature range usage, unlike
your traditional optical detection.

v) GPS Module:
GPS module has four connection pins, namely: Vcc, GND,
TX and RX. It requires a power input 5V 100mA ,which can
be provided using any suitable supply source. The RX pin
of Pi, i.e the 10th GPIO pin on the RPi board, should be
connected to the TX pin of the GPS module.
4. BLOCK DIAGRAM
The bot will move on the terrain and ultrasonic sensor will
alert on encountering any obstacle. The bot will then identify
if the object is trash and also classify it as bio degradable and
non-bio degradable based on a pre-trained model which is
trained images using tensor flow and keras. The robotic arm
of the bot will then pick up the trash and put in the
corresponding trash can. Proximity sensors which are fitted
on the top the bin will alert once the can is full and the bot
will immediately move to the nearby common bin to discard
the non-bio wastes. The bio- degradable wastes can be further
collected in a separate bin totally for manure use. GPS
International Journal of Engineering Research & Technology (IJERT)
http://www.ijert.org ISSN: 2278-0181
IJERTV9IS110152
(This work is licensed under a Creative Commons Attribution 4.0 International License.)
Published by :
www.ijert.org
Vol. 9 Issue 11, November-2020
326
module in the robot will be used to monitor the location of
the bot.

5.CONSTRUCTION AND OUTPUTS
5.1) Construction :
Actuation of bot in progress
5.2) Outputs:
The system showed good accuracy and could work its best
on the trained objects. The motion of the bot is also good and
its path determination seemed to be good. The novelty of this
paper lies in the concept of a cost-effective system that uses
IoT to optimize the working of a network of garbage
collectors.

6. CONCLUSION
 It helps the humans working in that field to relieve
from the health problems and can be prevented from
cancer, etc.
 The sensor network that can help the robot
manipulate and isolate wanted material means that
this pure yield will amount to a better price in the
market.
 This would be an added advantage to the Swacch
Bharath Abhiyaan by our Prime Minister.
REFERENCES
[1] Rama Prabha D, S. Mahindru, A. Srivastava, P. Nilalohita,
“Autonomous Garbage Collector-Robodumpster”, International Journal
of Civil Engineering and Technology(IJCIET), Volume 9, Issue
12(December 2018), pp. 545-552
[2] Technomentis, “Automated Garbage Collector”,
instructables.com.[Online].Available:
https://www.instructables.com/id/Automated-Garbage-Collector/ .
[Accessed April 10, 2019
[3] J. Bai, S. Lian, Member, IEEE, Z. Liu, K. Wang, D. Liu, “Deep
Learning Based Robot for Picking up Garbage on the Grass
Automatically” In Proc. IEEE Transaction on Consumer Electronics
TCE.2018.2859629.
[4] S. Nolfi, “Evolving non-Trivial Behaviors on Real Robots: a garbage
collecting robot” , Institute of Psychology, National Research Council,
Rome, Italy. Technical Report 96-04, March
[5] N. Kelly. A Guide to Ultrasonic Sensor Set Up and Testing
Instructions, Limitations, and Sample Applications.[Online].Available:
http://www.egr.msu.edu/classes/ece480/capstone/fall09/group05/docs
/ece480_dt5_application_note_nkelly.pdf
[6] Anonymous. (October 20, 2012). [Online]. Available:
http://www.atmel.com.
[7] Dakshit Chalagulla, Jeevanigi Jayateertha, Tamanneeta Giri and V.
Sailaja, “Gesture Controlled Bomb Diffusing Mobile
Robot”, Intelligent Computing and Control Systems (ICICCS) 2018
Second International Conference on, pp. 1-5, 2018.
International Journal of Engineering Research & Technology (IJERT)
http://www.ijert.org ISSN: 2278-0181
IJERTV9IS110152
(This work is licensed under a Creative Commons Attribution 4.0 International License.)
Published by :
www.ijert.org
Vol. 9 Issue 11, November-2020
327
[8] C. Tejas, V. Tejashwini, S. Dhal and P. S. Sirisha, “Flex controlled
robotic arm for the amputees”, 2017 2nd International Conference on
Communication and Electronics Systems (ICCES), 2017

Smart Trash Collecting and Segregating Robot
Sanjay B
Electronics and Communication Engineering
Velammal College of Engineering and Technology
Madurai
Sanju Vikasini RM
Electrical and Electronics Engineering
Velammal College of Engineering and Technology
Madurai
Abstract— The Earth is not a garbage can to litter everywhere.
Pollution is a serious issue that goes alarming every now and
then. Pollution increases subsequently with increase in
population. Garbage has to be strictly discarded to avoid
pollution. Here on emphasizing land pollution, “Garbage
Collecting Robot” can clean up all the garbages that are thrown
on the roads using Raspberry Pi. This robot will be built on a
metallic base and powered by a battery and contains a Pi cam to
segregate wastes on image processing. The bot will contain a
sensor for obstacle detection and hence has a good path
determination. On implementation, manual requirement can be
avoided and thus prevents people from sanitary issues.
Keywords— Image processing, Raspberry Pi, Recycling,
Robot, Ultrasonic Sensor.
1.INTRODUCTION
“Swacch Bharat Abhiyan”, a need for a clean India is a
dream that is yet to be fulfilled. Land pollution is one of the
major problems that we face especially in cities, malls and
even places like railway stations and airports. Sanitary
workers who clean up the garbages are tend to be exposed to
various skin ailments and health issues. Hence these manual
workers must be replaced by a bot which can be in turn
controlled by the workers to ensure employment doesn’t fail.
2.LITERATURE REVIEW
2.1) In a system that collects garbage automatically through
line. For which we are designing an arena so that the machine
will follow the specific path and also manually by using
Arduino and Bluetooth module. The machine can be
controlled by software which will give command to the
robotic jaw to collect the stationary waste. It also has the
characteristic to get controlled by voice command which
makes the robot to collect the garbage according to the given
command. Findings: We have designed a semi-autonomous
garbage collector robot which can do multiple functions. This
robot has one robotic arm in which it can pick the garbage
and dispense it in main basket attached to the robot. The
camera placed on robot helps the administrator to remotely
monitor the robot while collecting garbage. The prototype has
an electronic mechanism by which the robot can dispense it
collected garbage to the dispensing point. This robot has
installed batteries in which there is no fuel or electricity
required to complete the operation.
2.2) In an existing system, they have combined the usage of
proximity sensors with computer vision to accomplish our
purpose. The camera, mounted on the chassis of the robot,
remains inactive until an object is detected by the proximity
sensor. When an object is detected, the camera captures a
single image and sends it to the server for image processing
classification to detect it as garbage or non-garbage. The
advantage of this method is that a lot of time, memory and
power is saved by performing processing on a single image,
rather than continuously performing processing on every
frame of a live video feed. Thus we can solve this purpose by
using low end processors and hence the hardware cost shall
be substantially reduced. For classification of the image, we
can use two approaches: A. Perform image processing on the
robot processor itself: This will require a moderate cost
processor to be on the body of the robot. There may be some
lag in output since processor won’t be as high end. But
because of the same reason, the cost will be less. B. Send
images from multiple robots on a central server: The central
server will use a high end processor while the robot
processors can be extremely low end. The robots will simply
capture images and send them wirelessly to the server.
2.3) The robot Garbage collection system consists of a set of
a ARM connected to the motors. The mechanism will not
operate for entirety of the vehicle operation and will rotate
only for predetermined set of conditions. The main aim of the
mechanism is to collect garbage which is of similar
dimensions to that of juice cartons, crushed papers, and all
light items whose height is between 5 to 10 cms. Mechanism
is mounted on the front side of the base with an appropriate
ground clearance. When the sensor detects an obstacle, the
image processing is used to categorized the object as garbage
or any living organism .The object detection is specifically
used for safety of animals ,so that they don’t get harm . The
garbage is pick and drop into a bin which is placed right
behind the mechanism. The robot keeps collecting the
garbage until it reaches certain height in the bin. Once the bin
is filled the collected garbage is disposed to a selected place.
3.PROPOSED SYSTEM
The bot can travel in any land terrain and operates on a
battery. Commercial stainless trash cans are used to dump the
waste. It will also contain a scrubber at the bottom to sweep
along the way it travels. The motion of the bot is controlled
by Python programming and Raspberry Pi 3b+. The bot
platform is controlled by a 24v DC geared motor and rubber
wheels for smooth motion. The bot is trained to segregate bio
and non-bio degradable wastes separately on different trash
cans using image processing. Precision long range ultrasonic
distance sensors can be used to know the distance from
obstacles around so that the robot can autonomously navigate
around. A GPS module on board can be used for getting the
accurate location of the robot. This helps to find the main
garbage bin in which the location of it is pre saved. These
robots can have a load cell along with a proximity sensor to
tell the amount of waste in the trashcan.
3.1 COMPONENTS :
i) Raspberry Pi 3b+ :
International Journal of Engineering Research & Technology (IJERT)
http://www.ijert.org ISSN: 2278-0181
IJERTV9IS110152
(This work is licensed under a Creative Commons Attribution 4.0 International License.)
Published by :
www.ijert.org
Vol. 9 Issue 11, November-2020
325
The Raspberry Pi 3 Model B+ is the latest product in the
Raspberry Pi 3 range, boasting a 64-bit quad core processor
running at 1.4GHz, dual-band 2.4GHz and 5GHz wireless
LAN, Bluetooth 4.2/BLE, faster Ethernet, and PoE capability
via a separate PoE HAT. The Raspberry Pi is a low
cost, credit-card sized computer that plugs into a computer
monitor or TV, and uses a standard keyboard and mouse. It is
a capable little device that enables people of all ages to
explore computing, and to learn how to program in languages
like Scratch and Python.

ii) Ultrasonic Sensor:
The ultrasonic sensor measures the distance using
ultrasonic waves. The sensor head emits an ultrasonic wave
and receives the wave reflected back from the target. Hence
the distance to the target can be measure by measuring the
time between the emission and reception. Hence it is much
effective in detecting an obstacle that is encountered on the
path of the bot.
Distance L = ½ x T x C

iii) ADC ADS1115 Module :
The ADS1115 are great analog to digital converters that
are easy to use with the Raspberry Pi using its I2C
communication bus. The the ADS1115 is a higher
precision 16-bit ADC with 4 channels. It has a
programmable gain from 2/3x to 16x so you can amplify
small signals and read them with higher precision.

iv) Proximity Sensor:
Proximity sensors are sensors that detect presence of objects
without physical contact and relay that information captured
into an electrical signal. Proximity sensors are suitable for
damp conditions and wide temperature range usage, unlike
your traditional optical detection.

v) GPS Module:
GPS module has four connection pins, namely: Vcc, GND,
TX and RX. It requires a power input 5V 100mA ,which can
be provided using any suitable supply source. The RX pin
of Pi, i.e the 10th GPIO pin on the RPi board, should be
connected to the TX pin of the GPS module.
4. BLOCK DIAGRAM
The bot will move on the terrain and ultrasonic sensor will
alert on encountering any obstacle. The bot will then identify
if the object is trash and also classify it as bio degradable and
non-bio degradable based on a pre-trained model which is
trained images using tensor flow and keras. The robotic arm
of the bot will then pick up the trash and put in the
corresponding trash can. Proximity sensors which are fitted
on the top the bin will alert once the can is full and the bot
will immediately move to the nearby common bin to discard
the non-bio wastes. The bio- degradable wastes can be further
collected in a separate bin totally for manure use. GPS
International Journal of Engineering Research & Technology (IJERT)
http://www.ijert.org ISSN: 2278-0181
IJERTV9IS110152
(This work is licensed under a Creative Commons Attribution 4.0 International License.)
Published by :
www.ijert.org
Vol. 9 Issue 11, November-2020
326
module in the robot will be used to monitor the location of
the bot.

5.CONSTRUCTION AND OUTPUTS
5.1) Construction :
Actuation of bot in progress
5.2) Outputs:
The system showed good accuracy and could work its best
on the trained objects. The motion of the bot is also good and
its path determination seemed to be good. The novelty of this
paper lies in the concept of a cost-effective system that uses
IoT to optimize the working of a network of garbage
collectors.

6. CONCLUSION
 It helps the humans working in that field to relieve
from the health problems and can be prevented from
cancer, etc.
 The sensor network that can help the robot
manipulate and isolate wanted material means that
this pure yield will amount to a better price in the
market.
 This would be an added advantage to the Swacch
Bharath Abhiyaan by our Prime Minister.
REFERENCES
[1] Rama Prabha D, S. Mahindru, A. Srivastava, P. Nilalohita,
“Autonomous Garbage Collector-Robodumpster”, International Journal
of Civil Engineering and Technology(IJCIET), Volume 9, Issue
12(December 2018), pp. 545-552
[2] Technomentis, “Automated Garbage Collector”,
instructables.com.[Online].Available:
https://www.instructables.com/id/Automated-Garbage-Collector/ .
[Accessed April 10, 2019
[3] J. Bai, S. Lian, Member, IEEE, Z. Liu, K. Wang, D. Liu, “Deep
Learning Based Robot for Picking up Garbage on the Grass
Automatically” In Proc. IEEE Transaction on Consumer Electronics
TCE.2018.2859629.
[4] S. Nolfi, “Evolving non-Trivial Behaviors on Real Robots: a garbage
collecting robot” , Institute of Psychology, National Research Council,
Rome, Italy. Technical Report 96-04, March
[5] N. Kelly. A Guide to Ultrasonic Sensor Set Up and Testing
Instructions, Limitations, and Sample Applications.[Online].Available:
http://www.egr.msu.edu/classes/ece480/capstone/fall09/group05/docs
/ece480_dt5_application_note_nkelly.pdf
[6] Anonymous. (October 20, 2012). [Online]. Available:
http://www.atmel.com.
[7] Dakshit Chalagulla, Jeevanigi Jayateertha, Tamanneeta Giri and V.
Sailaja, “Gesture Controlled Bomb Diffusing Mobile
Robot”, Intelligent Computing and Control Systems (ICICCS) 2018
Second International Conference on, pp. 1-5, 2018.
International Journal of Engineering Research & Technology (IJERT)
http://www.ijert.org ISSN: 2278-0181
IJERTV9IS110152
(This work is licensed under a Creative Commons Attribution 4.0 International License.)
Published by :
www.ijert.org
Vol. 9 Issue 11, November-2020
327
[8] C. Tejas, V. Tejashwini, S. Dhal and P. S. Sirisha, “Flex controlled
robotic arm for the amputees”, 2017 2nd International Conference on
Communication and Electronics Systems (ICCES), 2017