A 6-Axis Robot with Machine Vision to Solve the Migrant Worker Crisis

Overview

Due to massive labor shortages caused by a lack of immigrants coming to the US to harvest crops, farm owners are reporting rapidly increasing food production prices. Food insecure low-income Americans that rely on low-cost food to feed their families are having to pay higher prices because of this labor shortage. This project’s purpose is to create a 6-axis robot that incorporates machine vision algorithms to pick apples. This system locates the centroid of the apple’s location and moves the robotic arm to grab the apple. The main structural components of the robot were constructed with 3D printed parts designed in Autodesk Fusion 360. To drive the 6 axes, 4 stepper motors were used for the first 3 axes and 3 servo motors were used for the last 3 axes. The apple detection algorithm utilizes machine vision from a PixyCam in order to determine the coordinate position of the apple. The algorithms are written in the C programming language on the Arduino IDE and downloaded to the flash memory of the micro-controller. A control algorithm aligned the robot to the position of the apple so that the robot could grab it. The final iteration of the control algorithm calculated the position of the apple in real time before moving the robot one step in the direction of the apple. It repeated until the apple was in the center of the frame. It aligned the X-axis and the Y-axis in parallel to improve time. The test results show that real-time processing of the apple location and parallel motion of the axes lead to the best robustness and time results. The robustness was over 80% with algorithm 2 and 3 and the time was under 15 seconds for the 3rd algorithm. This shows statistically significant improvement from the first version of the algorithm. The force exertion test identified the optimal force needed to grab an apple without bruising it. Literature showed that the threshold for bruising was 0.9 Newtons and a softer padding was added to the gripper of the robotic arm to decrease the force exerted to 0.6 Newtons. Simulations were run on Autodesk Fusion 360 to determine whether a scaled-up version of the robot would be structurally sound. The simulations determined that the designs were suitable for a larger scale. A cost analysis was also done on a larger system to determine the cost-effectivity of the robot. Compared to the rising costs of migrant labor, a robotic system would quickly reach a break-even point and save money over longer periods. This robotic system could be used to replace the missing labor that causes higher costs of produce for low-income Americans.