Performing Department
(N/A)
Non Technical Summary
This project seeks to identify and promote Robot Integrated High Tunnels (RobInHighTs) as a major economic driver for successful urban, indoor and emerging agricultural platforms. High tunnels are ephemeral structures that cover and protect crops, thereby extending the growing season by several months in midwestern climate zones. They have gained popularity over the last two decades in the urban and peri-urban setting due to their cost-effectiveness, adaptability, and low environmental costs. Though advantageous, high tunnels require an extra degree of management to ensure a quality crop and good production rates as the urban environment brings new challenges with unique pests and soil quality. Sustenance and economic viability of urban high tunnels are thus limited by the shortage and rising inequities in urban agricultural workforce. The primary objective of the projectis to demonstrate that the RobInHighTs platform can greatly automate routine high tunnel operations leading to increased crop yield, reduced manual intervention, while simultaneously preserving soil quality for sustained use. Our approach, guided by our past research, will integrate compact mobile platforms and novel hybrid soft-rigid manipulators within high tunnel environments in the Illinois Student Sustainability Farm and demonstrate the potential to navigate autonomously through rows of crops in high tunnels, use AI-powered perception to identify targets such as berries, leaves and possible pests, and perform fine dexterous manipulation tasks such as berry harvesting, pruning and precision spraying. Furthermore, the proposal seeks to understand and evaluate the economic implications of such a platform, barriers to entry for urban and minority farmers, and scalable business models for industry involvement through focused group surveys, workshop demonstrations and other extension activities. The proposal brings together a team of roboticists, crop scientists with expertise in high tunnels, agricultural economists, and education and extension experts from University of Illinois Urbana-Champaign and Tuskegee University.
Animal Health Component
50%
Research Effort Categories
Basic
20%
Applied
50%
Developmental
30%
Goals / Objectives
Urban farms can enable efficient and fresh local food production, minimize food miles, and open new avenues of income for minorities and small communities. However, urban farms face unique challenges associated with achieving profitability at small production scales, maintaining air and soil quality, irrigation, pest management and the climate required for efficient and economically viable produce. In this project, we emphasize High Tunnels (HTs), which are low cost, unheated, metal-tube structures covered with one or two layers of greenhouse plastic to create a protected environment for crops, as ideal solutions for urban farming. HTs are gaining in popularity as they prolong the production season especially in midwestern and northeastern climatic zones, increase yields and improve the quality of high-value specialty crops such as fruits, vegetables and cut flowers. The primary goal of the projectis to investigate the feasibility of robot-aided autonomy to streamline labor intensive operations in the urban setting. Specifically, the goal of the Robot Integrated High Tunnels (RobInHighTs) platform is to integrate recent advances in robot hardware design, vision-based perception, autonomous navigation, and manipulation towards automating high tunnel operations such as harvesting, pruning and pest management thereby achieving sustainable increases in yield and profitability. A secondary goal is to map the economic implications that accompany the increase in yield and identify barriers to adoption by engaging with urban and minority farmers in Illinois and Alabama through our Extension activities. Furthermore, we aim to train the next generation agricultural workforce through curriculum development and workshop activities steeped in robotics.
Project Methods
Our approach consists of well-integrated research and Extension activities to conceptualize and evolve the RobInHighTs platform. For the scope and duration of this project, we will create the first iteration of RobInHighTs by integrating matured robotic fundamentals with High Tunnel operations. The evaluation of the technology together with stakeholder involvement will inform our understanding of adoption and profitability. Our long term approach will seek to integrate and incorporate feedback from the stakeholders to guide best HT integration practices, which will in turn will inform the need for new fundamental robotic capabilities. The specific research aims of the project are (i) SA1: Robustification of fundamental robotic operations, (ii) SA2: Demonstrate and evaluate RobInHighTs for intensive farming, and (iii) SA3: Assessment of profitability, socioeconomic barriers for adoption and workforce training.In SA1, we will build on the team's past expertise in automated navigation, vision-based perception and manipulation of fruits and leaves. These operations are fundamental to automating routine high tunnel tasks such as harvesting, pruning, and monitoring for pests. Our robot platform will be comprised of the commercial TerraSentia mobile robot with a unique hybrid (soft-rigid) manipulator and a gripper that is capable of dexterous manipulation and reach. Specifically, we will (i) Use a combination of stereo camera vision, inertial measurement units and LiDAR to navigate through rows of high tunnel crops, (ii) Use images from onboard cameras and tip camera on the manipulator to automate the identification and localization of targets such as berries and potential pests by employing deep learning-based object detection (fast R-CNN and YOLO) or instance segmentation (MaskRCNN, YOLACT), and (iii) Combine trajectory planning and controls, and visual servoing methods to approach and manipulate a target.In SA2, we will integrate and validate the robot platform within high tunnels at the Student Sustainability Farm at the University of Illinois. Two high tunnels will be used, one RobInHighTs, and another control high tunnel, which is similar in all aspects but managed without robotic intervention. Both high tunnels will grow tomatoes during the summer and early fall, and greens (lettuce, spinach etc.) during late fall. Specifically, we will (i) Program subroutines for harvesting, pruning and pest management of tomato vines based on existing best practices in the high tunnel, (ii) Understand the effect of frequent robotic monitoring on crop spacing and density in greens, and (iii) Evaluate the effect of RobInHighTs on the yield and manual labor hours by comparing it with the control high tunnel.In the Extension and education specific aim, we will engage with stakeholders, specifically urban growers and consumers to identify socioeconomic factors that can promote or hinder adoption of RobInHighTs. Specifically, we will (i) Conduct detailed crop budgets for the RobInHighTs by using estimates of yield, and labor effort from SA 2, (ii) Identify barriers for adoption through focused group interviews and (iii) Conduct statewide educational activities and workshops for training personnel to operate high tunnels with robots, leading to faster adoption. We will also engage with industry to better understand the barriers involved in providing robotics-based services for urban growers.