Introduction to Physiz

This is one of first post in our journey to automate, monitor and optimize farms at Physiz(pronunc. fai — siz).

I started Physiz along with my Co-founders Mohit and Siddharth around mid April this year and it has been an exciting endeavor so far. We started with the idea of controlled environments for growing crops and drew inspiration from OpenAg, an OpenSource initiative by MIT Media Lab, in our initial prototypes.

The initial months as for any startup is like getting dropped into middle of a forest without a map and finding a way out

It has been six months now and we have crossed a few milestones. We have started generating revenue and see ourselves as cashflow positive in six months, moved from working in our living room to our first office, grown from 3 to a team of 6 people in Mumbai and 3 remote developers. Below are the key points that hi-light this journey.

Problem — The first and foremost aspect of building a business is to solve a problem faced by a large set of people/businesses for which they are willing to pay an amount sufficient to sustain the business. After our early exploration into the industry, we stumbled upon A LOT of problems and then narrowed down on a few and developed a solution for it. The problem that we are solving is Data collection at farm level, Automating repetitive tasks, and Optimizing the climate and nutrients to achieve consistent crop productions.

Market — Agriculture is an ENORMOUS market. . If we can classify the market based on our use case(Agriculture Input) it can be divided into two categories, Indoor & Outdoor. The scope of automation is limited to fertigation in an open field where as a controlled environment gives more scope of automation and hence more value addition and revenue. Therefore, it was a logical choice to start from Indoor market, which is Greenhouses and Urban Indoor farms and we conceptualized our products for Indoor while keeping in mind the Open fields. We are currently focussing on about 1500–2000 greenhouses in India, a very few percentage of them are automated and gives us a good opportunity to generate revenue and perfect our product before we take it to international markets.

Product — After realizing a size able opportunity in Indoor/Controlled environment we started developing our products. We were certain on the demand in market after doing an initial market research and sales. We could have developed a more robust system faster if we didn’t chose to decouple the features of hardware into different modules which, in my opinion, is a better design. It was also essential for us to give a foundation to our products which we can scale to open field systems without restructuring our hardware and software in future and it will be a while when we can see the advantage or disadvantage of this design. The products we have developed so far are:

  • Brain — The main controlling unit which runs the client application at farms. It mainly collects data from sensors, gives commands to connected equipment and syncs everything to the cloud.
  • Climate/Climate Pro — Sensor node to record temperature, humidity, light intensity in each spectrum and CO2 concentration and sends it to the brain.
  • Exo — Interface to connect any equipment at farm and turn them on and off.
  • Reservoir — Measure pH and EC of the the water source or circulating water in soil less farming techniques.
  • Dosing — Dose nutrients, pH up and pH down solution in soil less system.

4. Technology — Being a hardware and software solution, the technology landscape is wide at Physiz. Each piece deserves it’s own post but I will try to summarize it in a few words despite the temptation to give a detailed deep down.

  • Hardware — For our existing products the hardware needs were already developed and battle tested by various industries and it really helped in reducing the development and iteration cycle. The main hardware we are using in our products is a micro controller designed on top of esp8266 chip. Our target customers were indoor farms which does not require kilometers of range and we went ahead with wifi mesh networks for ease, reliability, availability and cost. For temperature, humidity and light intensity sensors we experimented with off the shelf sensors from Texas Instrument and SiLabs and finalized TI after one iteration. The major challenge in sensors was to measure pH and EC in a continuos manner without any hassle of calibration and minimum maintenance efforts. Bluelab is the market leader in pH and EC measurement and even their products require an initial calibration and monthly maintainence. We are still perfecting this and by end of this year we will have a definitive spec on the pH and EC probes. Either we will accept the market standard of calibration and maintenance or maybe we can come up with a better standard (fingers crossed)
  • Software — Our main software is the web application Farmfeed, which is an interface for users to see(yes literally!) what is happening at the farms, configure and control equipments. The automation works on a concept called recipe which we borrowed from OpenAg. It is a time series data of parameters affecting crop growth like temperature, humidity, light intensity and uses the equipment at the farm to maintain it at optimum level at different stages of crop growth. There are many parts behind the complete infrastructure mainly divided into Database, Message Broker using Redis and MQTT, TCP tunnel, Two Nodejs Application(one on cloud and another at premise), Firmwares written in C, WebServer, React on front-end. Each of them doing a specific task and to be discussed in detail in other engineering specific blog
  • Machine Learning — Before putting ML algorithms into production our focus is to refine our problem statements and build infrastructure and pipeline to handle large amount of data. It’s a no brainer that output of any AI/ML algorithm is dependent on the data, therefore, if we are first to get largest dataset from farms, we will work down on our way to improve accuracy of our algorithms currently which is not worth mentioning. It will be a while before we reach an accuracy of 95% and most of our automation decision will based on software configuration but it’s a good starting point to have a simple and clear problem statement to decide which equipment to turn on/off or do nothing every fixed interval of x minutes.
  • R&D — We are currently working on developing a soil based probe to measure moisture in a field with certain accuracy and at the same time is economically feasible for indian farmers(500–1000 INR). Sumeet before joining us was already working on a more sophisticated sensor to measure N,P,K from soil based probe without using chemicals but then we took a step back from there and started with soil moisture and will take it gradually from there. The methods we are exploring are Frequency Domain Reflectometry(FDR), Time Domain Reflectometry(TDR) and Theta Probes to measure soil moisture. It is in its early stages and we will have more updates as we make more progress.

5. Customers — We have been lucky to find customers along with customers finding us, organically, early in our journey. Our sales process ramped up after Priaynk joined us and hacked his way into finding our initial customers, which consist of Agri exporters, Agri produce sellers, Individual farmers and Fertilizer Manufacturers. From proprietary business to publicly listed companies we have found a diverse set of people willing to buy our products and it’s the greatest validation for us.

For the next two months we have more than 10 acres of indoor farms to be automated in various geographies and it is going to be a challenging task. There is no certain timeline on when we will release our open field system but it will happen in the very near future. We have two pilots to finish by end of this year and if everything goes well we will ship our open field system for everyone after that.

Lastly, thanks to our friends and family who have supported us in this journey, our early adopters for showing trust in a unproven product, cynics for questioning, people in the industry for their feedback and everyone else who has contributed directly or indirectly in anyway whatsoever.

At the end it is all about people, whatever we chose to do in our life boils down to the fundamental building block of our society “People” and we hope to do our part in impacting lives of people in a good way by staying persistent and dedicated to our mission

Build sustainable and efficient food systems

Automating and Optimizing our farms is the first step in achieving this mission and we have just started.