A Very MAD Blog
Month
Concept
The purpose of this project was to encourage people to produce their own food regardless of their prior skills, knowledge, or time. Eve is a personal plant assistant that will notify you at the very best time in your day to tend to your plants. There were initially three parts to this system, a device that will be implanted into an area or pot of the consumers choice, a receiver unit that will process the information and be able to calculate the best time to notify the user using AI technology, and an app to display the data.
The biggest challenge in this project was regarding the neural network. Creating the network was challenging to complete as there were many psychological factors associated with increasing motivation that could not be easily be defined and used in this project. The data given to the network would then have to be solely based on if the participant followed through with tending to their plants. If they did not, more information would be needed as to why they chose not to (e.g. Notification was not given at a reasonable time, they lost interest in gardening). Unfortunately, we did not end up utilising the neural network in this project as we did not have the time, the required data or the necessary experience with AI software to be able to create this complex system.
Given the time restraints of this project and other uncontrollable circumstances effecting the outcome, we would have liked to develop this further - focusing on seamlessly integrating the system together, spending more time on user interface and experience to ideally increase motivation, and to possibly create a finished marketable product for consumers. Our current prototype uses a sensor to send data to a base unit. This send the data to the computer that would process it through the AI, However, although we finished the fuzzy system we ran out of time to integrate everything together for the showcase. The output goes to an app that easily displays data and tips to the user.
We were able to complete a fuzzy system, an App prototype, a Kickstarter
style video, a website, and the physical device prototype. We will, However, keep developing Eve into the future, adding features, and making it a marketable product.
CONTEXT
The idea for this project stemmed from a need to focus on sustainable agriculture. At the peak production of this concept, farms had vastly developed to using mechanised equipment, becoming dependant on fossil fuels, as well as integrating pesticides and fertilizers into their produce. Farmers introduced these new ideas into their system without regard for future consequences (e.g. soil degradation, reduced profitability, a threat to human and animal health) (Reganold, J. 1990).
Alternative methods of agriculture have been able to avoid these major
repercussions by creating a system that works within a closed environment with
constant monitored conditions. Ideas such as Vertical Farming (farms that
utilise vertical space within an environment – predominantly in urban
environments – combined with the use of grow lights and a strict closed system),
Hydroponics (closed system that diminishes the use for soil and instead
utilises water and nutrient composition within the cycle to grow produce) and
Aquaponics (the use of nutrient-rich fish waste to “feed” the plant roots
submerged in the water) have led the way to a more sustainable future.
A demand for food will increase rapidly given the rate of population growth. Many of the problems expected to occur will cause a strain on food production and consumption. Producers will likely experience competition for their land, water and energy resources given the growth in population in relation to land area available (Godfray, H. 2013). Keeping up with this growth in population as well as being affected by unpredictable energy and oil costs, climate change and pollution creates limits on resources (water supplies, soil and land quality) that farmers rely on (Alexandratos, N. 2012).
An innovative concept has emerged from discussions about sustainable agriculture called Agroecology. Although this term can refer to many ideas such as a scientific discipline, agricultural practice, political or social movements (Wezel, A. 2011). This idea as one principal “regards farm systems as the fundamental units of study and in these systems, mineral cycles, energy transformations, biological processes and socio-economic relationships are analysed as a whole” (Altieri, M. A).
REFERENCES:
Alexandratos, N.; Bruinsma, J. World Agriculture Towards
2030/2050: The 2012 Revision; FAO, Agricultural Development Economics Division:
Rome, Italy, 2012.
Altieri, M. A. (n.d.). Agroecology: The scientific basis of alternative agriculture. Berkeley, California: Division of Biological Control, University of California.
Godfray, H. C. (2013). The challenge of feeding 9–10 billion people equitably and sustainably. The Journal of Agricultural Science,152(S1), 2-8. doi:10.1017/s0021859613000774
Reganold, J. P., Papemdick, R. I., & Parr, J. F. (1990). Sustainable Agriculture. Scientific American, 262(06), 112-121.
Wezel, A. (2011). Agroecology as a Science, a Movement and a Practice. Sustainable Agriculture, 2.
How to use your Eve – a personal plant assistant.
· You should find the following in the box.
· An Eve v1 sensor
· A Wireless Base Unit
Set up is easy!
All you need to do is connect your base unit to power and the internet, install the app from our website and then is it as simple as pressing the paring button on the sensor.
This will open the app where you can setup your plant.
When conditions are poor you will be notified through the app.
This will take you to a screen with information and tips about the plant.
Using Eve
Eve will tell you how your plants are doing and notify you when they need care at a time suitable for you in the app.
We made Eve in a range of colors, these are the finished prototypes version 3.
Fuzzy systems are a type of Artificial Intelligence that strive to conquer the complexities of human logic that isn’t always accurate or exact. A fuzzy system or Fuzzy logic tries to add meaning to words like sometimes and a little or a lot. These terms can then be used in programming or in other AI systems.
“Fuzzy logic is widely used in machine control. The term “fuzzy” refers to the fact that the logic involved can deal with concepts that cannot be expressed as the “true” or “false” but rather as “partially true”. Although alternative approaches such as genetic algorithms and neural networks can perform just as well as fuzzy logic in many cases, fuzzy logic has the advantage that the solution to the problem can be cast in terms that human operators can understand, so that their experience can be used in the design of the controller. This makes it easier to mechanize tasks that are already successfully performed by humans” (Fuzzy Control System, n.d.)
We will be using a Fuzzy Logic System to describe the condition of the plants to the user as well as training our AI.
Fuzzy Control System. (n.d.). Retrieved from wikipedia: https://en.wikipedia.org/wiki/Fuzzy_control_system
According to Dictionary.com (2018) ‘Artificial Intelligence’ is “the capacity of a computer to perform operations analogous to learning and decision making in humans, as by an expert system, a program for CAD or CAM, or a program for the perception and recognition of shapes in computer vision systems.” In summary, a software that can learn in one way or another.
Artificial Intelligence (AI) has been quickly evolving and changing, getting more advanced in recent time. We started with rule-based AI where a set of rules are given for the machine to follow that can be refined and added to. The problem is the world is a complicated place. Next came Learning AI where the AI took these rules and decided on the best outcomes and refined them. Now we are moving to Neural Networks that are almost like a ‘black box’. They work similar to a brain in which they have neurons and process data based on the connections of these neurons. This however, makes it impossible to look at the process it is using and refine it manually. (Barnatt, 2017)
AI can be split into 3 parts in most cases. The learning algorithm, Test data to learn off, and some sort of hardware. In a lot of cases AI is run off a normal computers processor which is what we are using, however, specialized hardware also exists. Test data can either be given to teach the AI then the AI used, and/or The AI can learn while it is in use from the data it is given and outcomes. We are using the latter. (Gerven & Bohte, 2018)
Types of AI:
https://simplicable.com/new/types-of-artificial-intelligence
https://futurism.com/images/types-of-ai-from-reactive-to-self-aware-infographic/
AI Hardware:
https://cloud.google.com/tpu/
https://azure.microsoft.com/en-us/services/machine-learning-studio/
AI Components:
http://www.explainingcomputers.com/ai.html
Dictionary.com. (2018, May 06). artificial intelligence. Retrieved from Dictionary: http://www.dictionary.com/browse/automation
Barnatt, C. (2017). DIGITAL GENESIS - The Future of Computing, Robots and AI. explainingcomputers.com.
Gerven, M. V., & Bohte, S. (2018). Artificial Neural Networks as Models of Neural Information Processing. Frontiers in Computational Neuroscience, 5-30.
Smaller design with IR led before batteries were installed. Batteries to be installed are lithium 3v button cells x2.
Using Picaxe chip with temperature, light, and soil moisture sensors.
433mhz Modules from China, they turned out to be faulty and replacing them would take too long or be too expensive to buy locally.
Hacking 315mhz Remote Module, the transfer rate was too slow ( 1bit per second)
Data over homemade FM radio transmitter including designing my own inductor that took a while. converting the data to sound, the quality of the signal wasn’t good enough to get all the data (data loss too high).
Radio Circuit I built ages ago, too large
As a last resort, I turned to trusty IR and am now designing a protocol based on Sonys IR protocol to send the sensor data. the main downside to this is the range and it needs line of sight.
We are working on two main parts of this project, 1) the sensor that will be placed in the pot/garden, and 2) the base unit that will calculate and display what the user needs to do.
The Sensor Circuit:
I designed this circuit so a picaxe chip could read the sensors and transmit the data to the base unit. This needed to be as small as possible so I kept that in mind when deciding on parts. A Picaxe is a type of microcontroller that is all on one chip and doesn’t need any other components to run. Picaxe is coded in basic over a serial interface.
Bonus points to those who realize that I switched standards partway through drawing that, school vs self-taught knowledge.
From that design I made a mockup on a breadboard to check everything worked.
The Sensor Code:
The code I designed to just read the sensor values then send them over serial where I will connect some sort of transmitter.
Fuzzy Stuff:
Our fuzzy system currently will give us a vague English word for different values, success. Shalom is still working on perfecting this then integrating it to an AI.
Goals:
My next goal is making the physical circuit and then making the wireless data transfer work.
