The ultimate goal of this project is to systematize agricultural knowledge in order to boost automation in agronomic and agrotechnology research. To achieve this, we have identified a three-step path. Each of these steps constitutes a necessary partial objective of the entire project. The three steps are as follows.

1. Build a standard for agriculture data. You can find more on this here
Building this standard means laying the foundation stone for an IT revolution in agriculture.
2. Make reliable agricultural knowledge available for anyone. This is the first product we are implementing. For more information, see here
Technology is data-driven. To digitize an industry you have to start with data.
3. Improve agriculture efficiency by speeding up agrotechnology reasearch. For a detailed description, see here
Let's help scientific knowledge expand by making it easier to share.

We can summarize our general approach with the following statement

Standards definition was definitely the thing to start with. It is indeed essential to build a solid foundation on which the whole system can rely. Encoding is the real pillar behind data accumulation, classification and analysis. Without a good encoding system it is not possible to talk about an "IT revolution" in a certain sector, e.g. agriculture.

All sectors where the IT revolution has begun, or is still ongoing, have developed some standard format to be able to store and share data. Just think of digital image encoding formats. Without these encodings, digital graphics as we know them today would not exist. Nor would exist generative text-to-image or image classification algorithms.

The same is true for text documents. Character encoding systems have enabled the creation, storage and digital sharing of written text. Character encoding systems are fundamental, then, to all products and services that enable us to handle written text, from word processing programs, to NLP algorithms, to LLMs.

ADF open format

The main purpose of this format definition is to encode agriculture data.

Figure 4.

ADF is the format that agriculture data needs right now: open, portable and extensible

You can find the implementation here in this repository

The table below summarizes the developments completed, and the roadmap for the next ones. A summary description and estimated timeline is provided for each. As new features are developed, sections will be added to this web page to describe them in detail.

Below you can see the schedule diagram of the activities listed in the table above. Some of these can be carried out (in whole or in part) in parallel.

Figure 3. This represents the roadmap of the first 15 months of the project. Currently have implemented the first feature (ADF format), and we are starting to write the TLA+ specifications for the database.

Terius' proposal mainly consists of two services: a distributed database, and a machine learning engine. These two services will be available either individually or in an integrated manner. Let us now look in detail at the main features of both, Terius Database and Terius Engine. In the third section (Comprehensive cloud), we will see the outline of a possible intrgeation between these two services in order to offer a complete product.

Terius Database

A distributed Database that can contain a reliable, tested, and standard-formatted agricultural information. This database must be able to be queried either through public APIs or through a web interface that can display data interactively through graph and table dashboards. Integrated with the database must also be a reporting service as well.

Reliable database

Get access to a wide database, with tons of data about crops. Public APIs are available as well as a web user interface.

Interactive dashboard

Data can be visually organized to compose some elegant and interactive dashboards, where you can visualize your data with custom charts and tables.

Reports

A powerful and flexible engine to print, edit, and schedule reports.

Market placement

Terius AI Engine

An on-demand cloud model that allows you to

• Completely automate greenhouses and crops
• Simulate the growth of a crop under arbitrarily given environmental conditions
• Determine what the optimal environmental conditions are for a given crop, based on analyses conducted on data stored in Terius Database

Comprehensive cloud

An all-in-one Platform that allows both services (Terius DB and Terius AI) to be integrated. In Figure 1 is shown an example of the interaction between clients and terius server. The client on the right is a prototype of smart greenhouse for researchers.

In Figure 2 we can see the interaction between two smart greenhouses. The same peer-to-peer network can be implemented with computers (instead of Terius greenhouses), as you can see in the left client of Figure 1.

Figure 1. A cloud infrastructure that is able to provide information to everyone: from final customers to researchers and industry.

---

Figure 2. This is an example of the P2P feature within Terius Network. Fully autonomous greenhouses (like this two prototypes) must be able to talk to each other, sharing thier own data.

At the moment, Terius is still an independent project. The resources we have are limited, and we are always looking for funding and sponsorship. In particular, it would be great to be able to get funding so that we can become a startup.

Any help in any form is always very welcome, feel free to contact me :)

If you are a programmer or systems engineer and would like to contribute directly to the project, you can find all the repositories here.

The team

Matteo Nicoli

matt@terius.org

Founder and software developer