On December 14th, 2019, a group of 11 individuals called UCI TIPPERS: Zero Waste Anteaters hosted a research symposium on their project ZotBins. The event, with about 25 individuals from the UCI community, presented the audience with insight about technology-based zero-waste management. In addition to insights about the project, the symposium also featured two guest speakers: Dr. Roberto Yus of the UCI TIPPERS research group and Dr. Yuanyuan Feng of Carnegie Mellon University.
Introduction
To introduce the Zero Waste Anteaters, PhD candidate Primal Pappachan and 4th year undergraduate students Joshua Cao and Owen Yang kicked off the event with an introduction about ZotBins. In this introduction, they talked about the three main topics that ZotBins addresses: (1) greenhouse gas emissions from landfills, (2) waste contamination issues, and (3) the goal of reaching Zero Waste by 2020 at UCI. Some of the notable highlights of this presentation was the achievements of the ZotBins project. Since the start of the project in 2017, the ZotBins Team now called Zero Waste Anteaters have deployed several smart bins in Donald Bren Hall and the West Food Court. The team has also submitted a proposal to the National Science Foundation’s Smart & Connected Communities with over 6 different organizations from UCI and outside of UCI supporting the project. Just this month, the ZotBins team have also completed their 70th weekly team meeting and there are still many more to come!
CMU IoT Online Presentation: Internet of Things Privacy Infrastructure (IoTPI)
The first presentation was from Professor Feng, who presented an Internet of Things privacy aware infrastructure. The infrastructure featured an IoT device logging system that would advertise itself to all users through an IoT Assistant (IoTA). This IoT device logging system is called the IoT Resources Registries (IRRs). Where if smart buildings, with many IoT applications, are constantly running, then users will be notified through their mobile devices and be given the option to opt out. This allows all IoT applications including ZotBins to be privacy aware and fulfill any privacy requirements by law.
Sensors, Circuits, and Hardware: The Data Collection Mechanism for ZotBins
Kathy Nguyen and Owen Yang presented their work on the electrical side and the hardware side of ZotBins. One of the biggest features of ZotBins is data collection. The specific types of data that ZotBins currently collects is bin fullness data and waste weight data. This data can be used in a variety of scenarios such as for an optimized pickup scheduling or for studying waste trends to improve waste diversion rates (the amount of waste diverted away from landfill) from a systematic approach. The types of sensors talked about in the presentation was the ultrasonic sensor to help collect fullness data and also the loadcell with an HX711 amplifier to collect weight data. Both Kathy and Owen talked about the implementation of the sensors and some of the challenges that came with it. Some of the future work that they plan to do, is to create a custom PCB to reduce the build time of each smart bin and to implement new sensors.
Using Object Detection to Create A Digital Waste Guide
Anthony Luu, Marissa Lafreniere, and Ted Zadouri presented their work on using Object Detection to create a digital waste guide for bin users to properly dispose of their waste. Their current work is using YOLO v3 to create a neural network that identifies and classifies different types of waste. For example, they could use their trained model to identify coffee cups. Once the coffee cups are identified they could then assign specific instructions to the user such as “empty liquids and recycle.” Another way to identify waste objects is to also use barcode scanners. A lot of products such as food wrappers or multipurpose sprays already have a barcode on them. Using a camera, they could also scan the barcode to identify the object. In the future, they plan to have this available on a mobile app and experiment with this idea in the UCI community.
UCI TIPPERS: A Data Management System for Smart Spaces
Dr. Roberto Yus presented about UCI’s TIPPERS system, a data management system for smart spaces. He introduced the concept of IoT applications and how they could all work together. Then he talked about the TIPPERS system, an IoT Framework with multiple IoT applications that is currently implemented at UCI. For example, one IoT application that is currently in place is room monitoring, where people can learn if a certain room is full or occupied. However, this can raise some privacy concerns and so Dr. Yus, talked about how the TIPPERS system is dedicated to having privacy practice by design. Users are notified whenever their data is being collected and also anonymized and then given the choice to opt out of the data collection. He also talked about how ZotBins started as one of the TIPPERS hackathon participants and a 3rd TIPPERS hackathon will be coming very soon.
The ZotBins Interface: Development and Use Cases
Marawin Chheang and Jesse Chong presented their work on the ZotBins website. They went through their entire process of creating the website starting from requirements analysis. There are two end-user groups that they had to keep in mind when developing the web application: (1) the Users group who interact with the bins by disposing items and (2) the Management group who are responsible for waste operations and studying waste trends. Some features of the website include real-time display of waste data, map locators, and more. They also talked about the problems with the old version of the website and redesigned a new set up using React/React Native with Express JS. Their new website features a very clean and straight forward dashboard for all the needs of both end-users and they even came up with a new mobile-app mock-up. In the future, this website and the app will be released to the public.
The Data Collection System with Error Detection and Fail-safes
David Pham presented the back-end data collection system of the physical smart bins in the ZotBins System. Currently the smart bins use a Raspberry Pi Zero W micro-controller to collect data from the ultrasonic sensor and the load cell. Then it pushes it to an external server, the TIPPERS server, to store that data. The old implementation of the code does not consider any errors that could happen with the sensors or the network. If any of those errors occur the continuity of the data collection will be destroyed. To cover these cases, David talked about his implementation of error detections and fail-safes in the original code. The error detections will help notify Smart Bin managers to service the bins if there are any sensor failures. Some fail-safes that were implemented also prevent loss of data. For example, if the network is down, then the ZotBins system code will save all the sensor data locally and check after a given time period to try to push the data back onto the external server. In the future, this code will be deployed and tested in a network of test bins.
Event Photos and Presentations
Photos: https://photos.app.goo.gl/yHqsW12Hh15kjEhy9
Presentations: https://drive.google.com/drive/folders/1_H2YWJvCkcXU7mRrmtlTd1s8V6HNIwHv?usp=sharing