{"id":{"repo_id":"rowan","oai_identifier":"oai:rdw.rowan.edu:etd-4118"},"canonical_url":"https://search.ndltd.org/etd/rowan/oai:rdw.rowan.edu:etd-4118","repository":{"repo_id":"rowan","name":"Rowan University","base_url":"https://rdw.rowan.edu/do/oai/"},"display":{"title":"DEVELOPMENT OF A MODULAR AGRICULTURAL ROBOTIC SPRAYER","abstract":"Precision Agriculture (PA) increases farm productivity, reduces pollution, and minimizes input costs. However, the wide adoption of existing PA technologies for complex field operations, such as spraying, is slow due to high acquisition costs, low adaptability, and slow operating speed. In this study, we designed, built, optimized, and tested a Modular Agrochemical Precision Sprayer (MAPS), a robotic sprayer with an intelligent machine vision system (MVS). Our work focused on identifying and spraying on the targeted plants with low cost, high speed, and high accuracy in a remote, dynamic, and rugged environment. We first researched and benchmarked combinations of one-stage convolutional neural network (CNN) architectures with embedded or mobile hardware systems. Our analysis revealed that TensorRT-optimized SSD-MobilenetV1 on an NVIDIA Jetson Nano provided sufficient plant detection performance with low cost and power consumption. We also developed an algorithm to determine the maximum operating velocity of a chosen CNN and hardware configuration through modeling and simulation. Based on these results, we developed a CNN-based MVS for real-time plant detection and velocity estimation. We implemented Robot Operating System (ROS) to integrate each module for easy expansion. We also developed a robust dynamic targeting algorithm to synchronize the spray operation with the robot motion, which will increase productivity significantly. The research proved to be successful. We built a MAPS with three independent vision and spray modules. In the lab test, the sprayer recognized and hit all targets with only 2% wrong sprays. In the field test with an unstructured crop layout, such as a broadcast-seeded soybean field, the MAPS also successfully sprayed all targets with only a 7% incorrect spray rate.","abstract_html":"Precision Agriculture (PA) increases farm productivity, reduces pollution, and minimizes input costs. However, the wide adoption of existing PA technologies for complex field operations, such as spraying, is slow due to high acquisition costs, low adaptability, and slow operating speed. In this study, we designed, built, optimized, and tested a Modular Agrochemical Precision Sprayer (MAPS), a robotic sprayer with an intelligent machine vision system (MVS). Our work focused on identifying and spraying on the targeted plants with low cost, high speed, and high accuracy in a remote, dynamic, and rugged environment. We first researched and benchmarked combinations of one-stage convolutional neural network (CNN) architectures with embedded or mobile hardware systems. Our analysis revealed that TensorRT-optimized SSD-MobilenetV1 on an NVIDIA Jetson Nano provided sufficient plant detection performance with low cost and power consumption. We also developed an algorithm to determine the maximum operating velocity of a chosen CNN and hardware configuration through modeling and simulation. Based on these results, we developed a CNN-based MVS for real-time plant detection and velocity estimation. We implemented Robot Operating System (ROS) to integrate each module for easy expansion. We also developed a robust dynamic targeting algorithm to synchronize the spray operation with the robot motion, which will increase productivity significantly. The research proved to be successful. We built a MAPS with three independent vision and spray modules. In the lab test, the sprayer recognized and hit all targets with only 2% wrong sprays. In the field test with an unstructured crop layout, such as a broadcast-seeded soybean field, the MAPS also successfully sprayed all targets with only a 7% incorrect spray rate.","abstract_has_math":false,"creators":["Sanchez, Paolo Rommel P."],"institution":null,"degree_name":"Ph.D. Doctor of Philosophy in Mechanical Engineering","degree_level":"Dissertation","degree_discipline":"Mechanical Engineering","degree_department":null,"school":null,"contributors":["Hong Zhang, Ph.D.","Shen-Shyang Ho, Ph.D.","Shreekanth Mandayam, Ph.D."],"advisors":[],"committee_chairs":[],"committee_members":[],"year":null,"date_issued":"","date_published":null,"updated_at":"2026-11-19T08:59:33Z","subjects":["artificial intelligence","machine learning","modular robotics","precision agriculture","precision spraying","weed control","Robotics","Mechanical Engineering"],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[],"person_identifier_entries":[]},"links":{"outbound_url":"https://rdw.rowan.edu/etd/3115","outbound_label":"Repository record","outbound_source":"dc:identifier","record_url":"https://rdw.rowan.edu/etd/3115","record_label":"Repository record","record_kind":"record","record_source":"dc:identifier","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://rdw.rowan.edu/do/oai/?verb=GetRecord&metadataPrefix=oai_etdms&identifier=oai%3Ardw.rowan.edu%3Aetd-4118","prefix":"oai_etdms"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Hong Zhang, Ph.D.","Shen-Shyang Ho, Ph.D.","Shreekanth Mandayam, Ph.D."]},{"key":"dc:creator","label":"Author","values":["Sanchez, Paolo Rommel P."]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.available","label":"Dc Date Available","values":["2023-05-24T07:00:00Z"]},{"key":"dc:date.created","label":"Dc Date Created","values":["2023-05-23T07:00:00Z"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Mechanical Engineering"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Dissertation"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph.D. Doctor of Philosophy in Mechanical Engineering"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["artificial intelligence","machine learning","modular robotics","precision agriculture","precision spraying","weed control","Robotics","Mechanical Engineering"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://rdw.rowan.edu/etd/3115"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Precision Agriculture (PA) increases farm productivity, reduces pollution, and minimizes input costs. However, the wide adoption of existing PA technologies for complex field operations, such as spraying, is slow due to high acquisition costs, low adaptability, and slow operating speed. In this study, we designed, built, optimized, and tested a Modular Agrochemical Precision Sprayer (MAPS), a robotic sprayer with an intelligent machine vision system (MVS). Our work focused on identifying and spraying on the targeted plants with low cost, high speed, and high accuracy in a remote, dynamic, and rugged environment. We first researched and benchmarked combinations of one-stage convolutional neural network (CNN) architectures with embedded or mobile hardware systems. Our analysis revealed that TensorRT-optimized SSD-MobilenetV1 on an NVIDIA Jetson Nano provided sufficient plant detection performance with low cost and power consumption. We also developed an algorithm to determine the maximum operating velocity of a chosen CNN and hardware configuration through modeling and simulation. Based on these results, we developed a CNN-based MVS for real-time plant detection and velocity estimation. We implemented Robot Operating System (ROS) to integrate each module for easy expansion. We also developed a robust dynamic targeting algorithm to synchronize the spray operation with the robot motion, which will increase productivity significantly. The research proved to be successful. We built a MAPS with three independent vision and spray modules. In the lab test, the sprayer recognized and hit all targets with only 2% wrong sprays. In the field test with an unstructured crop layout, such as a broadcast-seeded soybean field, the MAPS also successfully sprayed all targets with only a 7% incorrect spray rate."]},{"key":"dc:title","label":"Title","values":["DEVELOPMENT OF A MODULAR AGRICULTURAL ROBOTIC SPRAYER"]}]}],"canonical_facts":{"dc:contributor":["Hong Zhang, Ph.D.","Shen-Shyang Ho, Ph.D.","Shreekanth Mandayam, Ph.D."],"dc:creator":["Sanchez, Paolo Rommel P."],"dc:date.available":["2023-05-24T07:00:00Z"],"dc:date.created":["2023-05-23T07:00:00Z"],"dc:description.abstract":["Precision Agriculture (PA) increases farm productivity, reduces pollution, and minimizes input costs. However, the wide adoption of existing PA technologies for complex field operations, such as spraying, is slow due to high acquisition costs, low adaptability, and slow operating speed. In this study, we designed, built, optimized, and tested a Modular Agrochemical Precision Sprayer (MAPS), a robotic sprayer with an intelligent machine vision system (MVS). Our work focused on identifying and spraying on the targeted plants with low cost, high speed, and high accuracy in a remote, dynamic, and rugged environment. We first researched and benchmarked combinations of one-stage convolutional neural network (CNN) architectures with embedded or mobile hardware systems. Our analysis revealed that TensorRT-optimized SSD-MobilenetV1 on an NVIDIA Jetson Nano provided sufficient plant detection performance with low cost and power consumption. We also developed an algorithm to determine the maximum operating velocity of a chosen CNN and hardware configuration through modeling and simulation. Based on these results, we developed a CNN-based MVS for real-time plant detection and velocity estimation. We implemented Robot Operating System (ROS) to integrate each module for easy expansion. We also developed a robust dynamic targeting algorithm to synchronize the spray operation with the robot motion, which will increase productivity significantly. The research proved to be successful. We built a MAPS with three independent vision and spray modules. In the lab test, the sprayer recognized and hit all targets with only 2% wrong sprays. In the field test with an unstructured crop layout, such as a broadcast-seeded soybean field, the MAPS also successfully sprayed all targets with only a 7% incorrect spray rate."],"dc:identifier":["https://rdw.rowan.edu/etd/3115"],"dc:subject":["artificial intelligence","machine learning","modular robotics","precision agriculture","precision spraying","weed control","Robotics","Mechanical Engineering"],"dc:title":["DEVELOPMENT OF A MODULAR AGRICULTURAL ROBOTIC SPRAYER"],"thesis:degree_discipline":["Mechanical Engineering"],"thesis:degree_level":["Dissertation"],"thesis:degree_name":["Ph.D. Doctor of Philosophy in Mechanical Engineering"]},"updated_at":"2026-11-19T08:59:33Z"}