The project is hosted at https://gitlab.com/trustedplastic/trusted-plastic

There is a dedicated folder with a thorough report on achieved progress. Reviewers can check the results of the computer vision part with provided step by step visualizations and check the corresponding results on-chain with provided scripts.

Submission Deadline: April 15, 2024
Project Start Date: April 22, 2024
Project End Date: July 22, 2024

1. Infrastructure

   • Brief Description: Set up servers for backend, oracle, and project infrastructure needs.
   • Measure of Completion: Servers are fully operational and support project requirements.
   • Estimated Date of Completion: May 1, 2024
   • Budget: $3,000

2. Detailed Technical Specifications

   • Brief Description: Create detailed documentation of the development process, including specifications for smart contract interfaces and architecture.
   • Measure of Completion: Documentation is published in the open-source repository.
   • Estimated Date of Completion: July 24, 2024
   • Budget: $4,000

3. Investigator Smart Contracts

   • Brief Description: Develop smart contracts for decentralized decision-making by investigators, with multisig and consensus decision-making functionalities.
   • Measure of Completion: Smart contracts are deployed and can successfully execute the decision-making process of inspection
   • Estimated Date of Completion: May 22, 2024
   • Budget: $10,000

4. Plastic Credits Smart Contracts Modifications

   • Brief Description: Modify smart contracts for plastic credits, including features for merging shipments, splitting profits for individual collectors and merger, optimizing to prevent memory issues, and separating tokens into plastic and carbon credits.
   • Measure of Completion: Smart contracts are updated and functioning according to new specifications.
   • Estimated Date of Completion: June 5, 2024
   • Budget: $12,000

5. Real-time Oracle Deploy and Optimization

   • Brief Description: Deploy and optimize a real-time computer vision oracle for plastic waste verification to handle larger volumes efficiently.
   • Measure of Completion: Oracle can verify a minimum of 500 items in under 20 minutes.
   • Estimated Date of Completion: June 19, 2024
   • Budget: $18,000

6. Prototype of the Verification Toolbox

   • Brief Description: Assemble and configure hardware and software for data collection and verification at recycler facilities.
   • Measure of Completion: Hardware setup is operational and software collects and verifies data accurately.
   • Estimated Date of Completion: July 3, 2024
   • Budget: $4,000

7. API Integration with Pfandomats

   • Brief Description: Develop API and UI for the collection process using pfandomats, allowing data transmission to the backend.
   • Measure of Completion: API is functional and UI is integrated and responsive.
   • Estimated Date of Completion: July 10, 2024
   • Budget: $6,000

8. UI: Collector's / Merger's / Recycler's Interface MVP

   • Brief Description: Create user interfaces for collector, merger, and recycler profiles, showcasing transaction histories and financial information.
   • Measure of Completion: Interfaces are accessible online and display accurate data.
   • Estimated Date of Completion: July 17, 2024
   • Budget: $12,000

9. Unit Tests

   • Brief Description: Cover all smart contract methods and backend services with unit tests, ensuring full functionality.
   • Measure of Completion: All tests are written, pass successfully, and are documented.
   • Estimated Date of Completion: July 24, 2024
   • Budget: $8,000

10. Android Application for Individual Collectors

    • Brief Description: Develop an Android application with secure camera functionality to capture and track collected plastic amounts.
    • Measure of Completion: The app is operational, secure, and accurately tracks shipments.
    • Estimated Date of Completion: July 24, 2024
    • Budget: $12,000

Total Development Hours: 480 hours
Total Budget: $99,000

First Name: Alex

Last Name: Ametov

Role: CEO

LinkedIn: https://www.linkedin.com/in/amyote/

First Name: Artur

Last Name: Abliazimov

Role: Senior Backend/Blockchain Developer

LinkedIn: https://ru.linkedin.com/in/artur-abliazimov

First Name: Alexander

Last Name: Demin

Role: Senior Frontend Develolper

LinedIn: https://linkedin.com/in/alexander-s-demin

First Name: Evgenii

Last Name: Safronov

Role: CTO

LinkedIn: https://www.linkedin.com/in/evgenii-safronov-06ba171b6

Plastic recycling market:
For many companies, replacing plastic from a product or a full transition to recycled plastic is not currently possible, but we can give them the opportunity to contribute to recycling and plastic removal through 'plastic credits' similar to carbon credits, which allow companies to achieve carbon neutrality while still in the transition to renewable energy. The global recycled plastic market size was USD 46.89 billion in 2022 and is projected to grow from USD 50.78 billion in 2023 to USD 88.96 billion in 2030 at a CAGR of 8.3% during the forecast period.

Briefly about TrustedPlastic:
Our platform allows us to transparently account for the amount of plastic collected by individuals, companies, and NGOs involved in land, coastal, ocean and river clean-up and generate credits for this volume. Plastic and carbon credits are issued as tokens and transparently sold directly to entities interested in off-setting the amount of plastic they release with their products. As the amount of plastic credits sold grows, we shall establish an exchange to move away from direct sales.

Related projects:
There are several projects that use blockchain technologies in recycling. Below are few links:
1.https://businessnorway.com/articles/cleaning-up-plastic-in-the-ocean-with-blockchain-technology-and-ocean-plastic-certificates

2.https://plasticbank.com/

And research suggesting blockchain based approach for 

https://www.sciencedirect.com/science/article/abs/pii/S0013935122009586

Partners:
We recently partnered with Sparklo, a company that manufactures reverse vending machines (called “pfandomat” in Germany) for collecting plastics. In several European countries, laws mandate a surcharge on plastic bottles to encourage recycling, while other regions rely on voluntary recycling efforts. Sparklo's collaboration with us supports the concept of plastic credits, a financial tool designed to enhance plastic recycling profitability when direct sales of collected plastic aren't sufficient. Sparklo’s machines automatically classify bottles, allowing us to integrate our software for additional verification without affecting transparency. They currently handle about 150 tons of plastic per month, potentially generating $75,000 monthly from plastic credits. Sparklo's data also helps address potential fraud in plastic collection, confirming varied bottle types across different locations. Plastic credits generated and sold by TrustedPlastic shall enable rapid growth of Sparklo’s network.
We are also discussing potential plastic credit purchases with several other companie. Therefore, we added few objectives to support manual plastic collection which is still prevalent in many regions.

From the potential buyers side we have few first customers waiting for TrustedPlastic to start generating credits.

Plastic collection and recycling process:

Let us first describe the typical process of plastics waste collection and risks associated with different steps.

On the picture above steps are:

1. Individual collectors delivers plastic to a local collecting point
2. Collector entity merge shipments into a batch and put it into delivery
3. Delivery service transfer plastic waste to recycling facility
4. Recycling factory sorts out recyclable plastic and utilize non-recyclable

As for the reverse vending machines, the process is quite similar. You only need to replace a human collector (1) with a reverse vending machine, and take into account that shipments are merged during delivery (2-3).

The verification process lies in

1. Inspection of visual evidence of collected plastic  
2. Comparison of visual evidences and weight for shipped plastic from collectors (1) and received on recycling facility (4)

We find these two steps sufficient to eliminate two potential frauds:

1. Bringing plastic waste from other regions
2. Accounting collected plastic more than once

More detailed description on this topic can be found in our Activation Award submission. Note that (1) “Bringing plastic waste from other regions” is less likely than (2) “Accounting collected plastic more than once” as it involves a high level of organization. We target to mitigate this risk on a later stages of the project as we gather more data on collected plastic. An example of shipment logical flow can be found in the detailed technical architecture document.

Blockchain:

We benefit from implementing the whole business process flow on Soroban smart contracts. The plastic waste collection happens mostly in developing countries, therefore we are happy to choose Stellar due to its focus on developing countries, huge throughput and readiness for micropayments. 

Simplified scheme of the technical architecture is below.

To scale the process, we have to move from performing “Investigation” ourselves but outsource it to independent investigators who shall receive a small reward for inspecting shipments. While we target that computer vision oracle shall cover most of the comparison cases, investigators shall not only cover remaining ones but provide valuable feedback for improving computer vision.

Once the verification of plastic collection succeeds, the platform triggers token minting. Every minting transaction is linked with the previously verified collection transaction and it issues two types tokens: plastic credit and carbon credit tokens. Plastic credits are directly associated with amount of removed plastic:

1kg of removed plastic leads => 1 plastic credit token

Carbon credits are related to the fact that building from recycled plastic emits less CO2  than from new plastic:

925.9 kg confirmed recyclable plastic = 1 carbon credit token

However, to issue an exact number of carbon credits we need to estimate the amount of recyclable plastic among collected plastic waste. It requires efforts from computer vision oracle that are out of the scope for this application and planned for later development.

The tokens are distributed among all participants according to rules defined in smart contract. Proportions might depend on the e.g. plastic collection process - reverse vending machines or manual collection. Below you can find approximate shares:

UI, UX and tools for verification:

Collectors and Recycler shall be able to see their history and earning through dedicated UI. They will also have some controls over business flow, e.g. altering collectors can specify recycling facility they are planning to use. On the recyclers side we have to place a dedicated hardware (industrial camera and an edge computer). It allows recording visual evidence of delivered plastic waste in industrial conditions of recycling facilities. Reverse vending machines are already equipped with cameras and software to identify plastic bottles. As for individual collectors we target to release a dedicated application for android-based mobile phones. It will guide collectors and securely capture collected plastic waste on camera.

Real-time oracle based on computer vision:
Once we have visual data of collected plastic and received on a recycling facility, we verify if the plastic waste is the same and of same amount. We do it through computer vision service oracle, that is deployed on our servers. 

We developed a prototype that shows a possibility to match collected plastic bottles received at the recycling facility for the case of reverse vending machines. The approach is based on segmenting the plastic bottles and applying modern feature matching, and therefore didn’t require any additional pretraining. Feature matching allows to match parts of the bottles or plastic waste in general even in the case when plastic waste was unpredictably deformed or only partially seen on the pictures. We show here some highlights and welcome to check out more results at our project repository. 

Examples of matching and non-matching bottles. Note the amount of matching features on left and right pairs.

There are many other simpler checks such as total amount of detected objects, object type distribution, etc. Further work here focus on making this oracle “real-time” in terms of performance and generalization of this approach to other objects beyond plastic bottles.

At the end of 2023, we kicked off the development of the blockchain-based platform and made our MVP for computer vision real-world oracle. 

In Q1 2024, we finalized the partnership with Sparklo.com—the details, roles, and expectations are all set. We also defined the technical specifications and functionalities of our blockchain platform using Stellar. We started discussions with potential buyers of plastic credits and other stakeholders.

In Q2 2024, we will launch a beta version of the platform to test it in real-world scenarios. We refined the platform based on user feedback, and expanded our network of plastic waste collectors and potential credit buyers. We conducted initial testing with a closed group to gather early feedback, and launched an informational campaign to raise awareness among potential users and additional partners.

We are now working on the official launch of the completed platform. We're facilitating the first transactions of plastic credits on the platform.

In Q3 2024, we'll assess the platform's performance and the impact of the plastic credits market. We'll focus on building a community around the platform, and explore opportunities to expand the service to additional countries with new collecting/recycling partners. We will be ready to implement a full-scale marketing strategy to attract users and stakeholders. 

Q4 2024 is for developing additional features based on user demand and technological advancements. We'll also initiate the development of a secondary marketplace for direct trading of plastic waste. We'll integrate the capability to issue carbon credits based on the recyclable content of the plastic collected. We'll conduct an annual review of the project's progress and strategize for 2025, as well as enhance the technological backbone to support increased transactions and user base. 

In January 2025, we will release our first sustainability report detailing the environmental impact achieved with the help of our community of brands, collectors and recyclers in collaboration with the Stellar community of makers.