1 of 26

Trashinator

Sort it Out!

Team 4 - Kiyaranu | Xaviera | Irfan | Quamran | Nishant

2 of 26

Background

Only

13.5%

Waste

Managers

High Cost

of the total waste

undergoes recycling

use a lot of resource to manually sort the trash.

cost of trash collection

is higher than the cost of processing trash.

3 of 26

401,658

418,214

429,128

429,531

446,102

497,238

2014

2015

2016

2017

2018

2019

Background

Total amount

of Waste

Plastic Waste

Sorting Problems

Source: Ministry of Environment

Unit: tons per day

Total amount of Waste

Surge in food deliveries by 17.7%

during COVID

Shows a proportional increase with increasing trash

4 of 26

GOAL?

Efficient garbage sorting for garbage collection

5 of 26

6 of 26

Trashinator

Your trash-sorting companion

A minimalistic and affordable appliance that uses AI to categorize and sort trash based on their type and provides its users with analytics (powered by SAP) on their waste disposal habits

What is it?

Feature set

AI-based waste classification and sorting mechanism

7 Distinct Waste Container Suitable for All Household Waste types

Real-time waste level monitoring and waste usage analytics

7 of 26

Convenient and Easy-to-Use

Automatic Waste Sorting

Provides Insights on Waste Sorting Benefits

Easily Accessible

Reduces the time-cost of waste sorting

Eliminates any confusion on how to categorize waste

Reduces chances of waste missorting

Increases awareness of waste missorting impact and importance of reusable materials

Friendly design allows the placement of Trashinator inside of homes

8 of 26

I am Irfan!

(22)

Occupation

Student

Hobby

Studying, working out

Goal

To efficiently use his time while studying, working out, and performing household chores.

Pain Points

Too much time spend on studying and therefore too lazy to sort his trash

Unable to identify trash properly, and properly manage them.

9 of 26

10 of 26

11 of 26

12 of 26

390M

SOM

1.94B

SAM

22.2B

TAM

Household in Korea

Student enrolled in university in Korea

Student live in dormitory (20% of SAM)

TAM SAM SOM

*Assuming the product price of 100.000 won

13 of 26

Technology Architecture Diagram

Docker

Dataset

Training Result

Detect trash type

detection result

Sort based

on the type

User

trained model

result and statistic

SAP AppGyver

SAC

Camera System

PlatformIO

Sorter System

AWS S3 Bucket

SAP HANA DB

SAP CAP

SAP AI Core

SAP AI Launchpad

Hardware

AI

Backend and Frontend

OData

14 of 26

Docker

Dataset

Training Result

Detect trash type

detection result

Sort based

on the type

User

trained model

result and statistic

SAP AppGyver

SAC

Camera System

PlatformIO

Sorter System

AWS S3 Bucket

SAP HANA DB

SAP CAP

SAP AI Core

SAP AI Launchpad

Hardware

AI

Backend and Frontend

OData

15 of 26

AI

Two trash type: plastic, paper

Use small size CNN

Docker

SAP AI Core

SAP AI Launchpad

HC-05 Bluetooth Module

H5 → tflite → cc

Upload the cc file to ESP eye

Detect the trash

Upload the result to Kyma

16 of 26

17 of 26

Hardware

HC-05 Bluetooth Module

Arduino Uno MCU

Servo Motor

Espressif ESP32-EYE

18 of 26

Back-end and Front-end

SAC

Primary key

Type of trash

SAP HANA CLOUD

SAP AI Launchpad

19 of 26

Back-end and Front-end

AppGyver

Model Creation

SAC

SAP HANA Cloud

20 of 26

Back-end and Front-end

SAP HANA CLOUD

AppGyver

Story Creation

SAC

21 of 26

Back-end and Front-end

SAP HANA CLOUD

AppGyver

Filters in Stories

SAC

22 of 26

Back-end and Front-end

AppGyver

23 of 26

Demo

24 of 26

25 of 26

Scalability and Future

Partnering with Government

Expanding Product Line

R&D to improve performance

Educational Materials

26 of 26

Thank you