Auri

The Proposal

I wish to establish a project that would improve agriculture (both for food and for medicinal applications) by providing technology assistance that would help “tune” plant environment to meet specific phenotype requirements (the general term is "Precision Agriculture").

Example use cases

The following scenarios are just the tip of the iceberg of what can be achieved with optimized agriculture

Improve Yield

By monitoring the plant important metrics and comparing them to that crop optimal baseline, the farmer can apply accurate amounts of specific nutritients in precise locations, thus reducing the cost while improving crop yield (food quantity per unit of land)

Extend product shelf life

A farmer can provide his crops the right elements, based on knowledge gathered through thousands of crops analysis, such that at the time of the picking, the plant holds the best internal "ingredients" (compounds) to allow it reaching its maximal shelf life. Thus enabling longer logistic transfers and reducing food waste.

"Tuning" medicinal crops

Medicinal Marijuana that is to be prescribed to epileptic children needs to minimize the THC concentration while maximizing the CBD and other theraputic compounds concentration. Comparing the plant measurements to known recipes in the database will help in reaching that goal.

What I actually suggest

Here are the main system components that I see in my mind

On the plant side

Farmer will place a small low-energy-multi-sensor device within proximity to his crop. This device will collect all required measurements (soil, air, perhaps even on-plant measurements) and transmit them (either directly using on-board communication or via close by aggregation node gateway) to the cloud for processing.

Photo by Christian Joudrey

In the cloud

All collected measurments will be cataloged and categorized according to best practices (possibly MIAPPE), thus enriching database and allowing better applying of statisitcal analysis methods (Did I hear someone say Machine Learning?).
Then the readings will be compared to a golden-standard baseline of optimal environment conditions (MIT Open Phenome Project seems like great candidate to start with) while aiming to the user pre-defined deisred phenotype (e.g - Increase tomatoes sweetness, make lettuce more meaty, improve pest resistence of wheat, etc.).

Photo by imgix

Back at the growing site

The system at first is planned to be only indicative (though operative plug-in modules that integrates with an on-site automation system is definitely on the agenda). The cloud analysis results will be accessible via front-facing application (web/mobile) and will enable the farmer to take the recommended steps in order to achieve the desired crop outcome.

Photo by Tim Mossholder

Open questions and challenges

Will it make a difference

How profound is the effect that enviromental changes have on the phenotypic plasticity within the same strain of crop (my small research suggest big impact but I may be wrong).

What should be measured

Moisture, Temperature, pH, EC, NH3, Oragnic matter, minerals presence, Radiation intensity, this are just a few of the attributs that shuold be measured. Find which measurements are important for plant expression and prioritize them by degree of impact on the plant (Soil / Air / On-plant measurements)

How to measure those metrics

Once it is clear which data is the most important, can that data be collected (sensed) in a way which is precise enough, cost effective and phsically small

Find comparison baseline

Professional litrature contains general information on how to grow different strains. Yet this information is hardly publicly available and far from being organized in a meaningful / insightful way. As stated before, I'm currently looking into MIAPPE and MIT Open Phenome Project as starting ground for solving this challange

Extract insights from data

Hopefully, quite fast data will start to pile. Perhaps unsupervised learning models are fitting to classify and extract features from the data in a way that can provide meaningful instrcutions to the farmer

Enginner the entire system

From the low-energy collector sensors, via the gateway to the cloud. The cloud infrastructure (DB / uptime / devops / ML), all the way back to the front application. The entire system requires loving and knowledgable hands to bring together all the components

Final thoughts

Colony infrastucture for open distributed work place

Colony.io is a very promising project that aims to redefine the future of work and demonstrate how open organizations can operate. I believe that this projects would be a great showcase of how individuals from different backgrounds and with various skill sets can come together and colleborate toward a common goal. In addition, if at some point in the future this project will offer any means of monetization (plugins for automated control systems at the farm just to name an example), then the revenue of the project colony will be distributed among the colony members based on their contribution to the project.

About me & contact

My name is Yotam Katznelson, I'm an Electric Engineer (with strong emphasis on the software side of things throughout most of my professional life). After reading an interesting article about the world food problems (resources to yield dwindling ratios, chemical pesticide impact on the environment, unrecycled plastic metarials in the agriculture industry, so on and so forth) I came to a conclusion that this can be a great domain to provide my 10 cents and try to contribute. If you find this proposal interesting, if you hold any knowldge that may beneficial for the cause (Plants researchers, Chemical analysts, Electric engineers, software developers, biologists or any other related human knowledge), if you would like to contribute in any way or even if you just want to speak for no specific reason, don't hesitate and do drop me a line to yotam.katznelson [at] gmail.com

Background