GrowBot
Team information
Category:
Hazel Westingham
Bachelor
Simon Fraser University
Connor Horii
Bachelor
Simon Fraser University
Ethan Brown
Bachelor
Simon Fraser University
Cam Foster
Bachelor
Simon Fraser University
Atiqa Mohsin Ali Khan
Master
Simon Fraser University
This project is being coached by
About the team
GrowBot is a team of 4th year mechatronic systems engineering students from SFU, Vancouver, Canada. We have backgrounds in additive manufacturing, control systems, mechanical design, CAD, model simulation, and hundreds of collective hours in design teams, including rocketry and student F1. Each of our four core members have completed 8 or more months of industrial or research engineering co-ops, and are driven towards sustainable solutions such as automated hydroponics.
Our vision
Hydroponics is an indoor farming method which replaces the growing medium of soil for water. This change not only decreases water usage, but also provides a high level of adjustable control over the growing conditions. With this heightened control, it is possible to maintain a higher level of crop health, and thus a higher yield compared to traditional farming techniques. However, hydroponic farms are limited by the frequent attention needed to maintain nutrient levels. As such, the development of automated systems in order to minimise the required labour and attention these farms require is of interest. With an ever increasing population, and decreases in resources and farmable land, high density and high yield farms are predicted to become increasingly critical, only further increasing the demand for and interest in hydroponic systems. Therefore, the development of key components of the hydroponic automation process became the vision of this team.
Our solution
Advances in machine learning have enabled its application in the health monitoring and farming control of plants. GrowBot intends to automate the growth of Albion strawberries in a hydroponic chamber, where the nitrogen-phosphorus-potassium (NPK) macronutrient administration ratio and delivery is overseen by a computational neural network (CNN), facilitated by NPK sensors. Automation is achieved through a supervised machine learning approach which reads the uptake rate of NPK macronutrients and commands for the administration of nutrients of the necessary NPK ratio, based on the plant’s demands. Since the NPK demands change as a function of the plant’s growth stage (from transplant to harvest), and other factors such as water pH and concentration of dissolved solids, a machine learning algorithm is necessary for precisely pampering the strawberry over its lifecycle.