Technology Focus Areas

TECHNOLOGY FOCUS AREAS:

Robotic/Autonomous Systems
Sensors, Machine Learning, and Connectivity
Data Analytics
Agriculture Data Security

Robotic/Autonomous Systems

Robotic platforms including both ground and aerial systems are being developed and deployed for many agricultural purposes. NDSU research under the FEWS Initiative is focused in several areas:

• High-Throughput Data Acquisition and Analytics Platforms for Plant Breeding

  Thousands of small field plots are planted and harvested every year to develop and evaluate new genetic varieties of crops that optimize desired traits. NDSU has ten plant breeding programs each having nurseries at multiple locations. Traditional methods of data collection are very manual in nature. Ground-based and aerial-based (drone) platforms are being developed to automate data collection for phenotypic traits to improve both the efficiency and accuracy of data collection processes.

• Weed Management

  Site-specific weed management (SSWM) research is focused on increasing the efficiency and precision with which weed mitigation can be applied. Sensors and AI technology onboard ground and aerial vehicles are being developed to accurately detect locations of weeds within various types of crops. In some cases, the focus also includes identifying the species of weed. Methods to mitigate weeds in a precisely-targeted, efficient manner are also being studied. These include chemical (spray), mechanical (tillage), and laser options. The platforms being developed and evaluated for autonomous weed management range from small robotic systems to large ground-based spray applicators and drones.

• Soil Health and Sustainability

  Research involving integration of emerging UAV-based remote sensing technologies with precision soil monitoring (moisture, pH, salinity, and organic matter) is focused on assessment of crop and soil moisture dynamics. The research outcomes will enable improved agricultural decision support tools for dryland agriculture production.

Sensors, Machine Learning, and Connectivity

• Sensors

  Internet-of-Things (IoT) systems have been widely adopted in many industry sectors and also for many home applications. In agriculture, IoT sensors have potential to significantly impact crop and livestock decision making through in-situ measurement and monitoring of many important physical parameters. While technology is very advanced in farm equipment and other areas of agriculture, there are still relatively few in-field sensors deployed in farm and ranch operations.

  NDSU research is focused on adapting existing technologies and developing new sensor technologies to provide solutions for a wide range of use cases in both agricultural research and production operations. Research challenges include developing sensors that can integrate seamlessly into farm operations without being overly burdensome to deploy and maintain. Sensors must be robust to survive harsh conditions, cost-effective, and produce usable insights to inform decision-making.

• Machine Learning

  Machine learning is transforming the way that agricultural data from various sources is analyzed to produce insightful information. Data is available from in-field sensors, drone and satellite imagery, weather information, ag machinery sensors, historical production records, and many other sources. The research involves collecting raw data sets and training machine learning models to analyze data and produce accurate information to inform decision-making.

• Connectivity

  Reliable and affordable data connectivity has become an essential element of farm and ranch operations. While wireless communication technologies have evolved and improved greatly, a single optimal solution for last-acre connectivity has not emerged. Tradeoffs between cost, power consumption, data rates, and communication range factor into this challenge.

  NDSU research is focused on integrating and testing various technologies to provide effective solutions. The use cases vary from low data rate, battery-powered in-field sensors to autonomous vehicles and other systems transmitting high-bandwidth imagery or video. Wireless communication capability from satellite-based systems, terrestrial cellular systems, LoRaWAN, and other technologies each have unique attributes and are included in the research. Test beds have been installed at NDSU Research Extension Center locations to enable data collection and evaluation across multiple farm settings.

Data Analytics

Raw data collected from IoT sensors, equipment, unmanned aerial vehicles, satellite systems, lab testing, and other sources is transformed into meaningful information through analytics. NDSU’s Agricultural Data Analytics team is developing tools that can be broadly useful to agricultural researchers.

• FielD·Hub is an R Shiny design of experiments (DOE) app that aids in the creation of traditional and non-traditional augmented experiment designs, including spatial trials with checks assigned to plots in a diagonal arrangement across the field. Outputs from FielD·Hub include interactive field maps, tables, and field books that can be extracted and saved.
• Mr. Bean is an easy-to-use R-Shiny web-app that simplifies the analysis of large-scale plant breeding experimental analysis by using the power and versatility of Linear Mixed Models (LMM).
• ExLibris, a querying and analytics tool, is designed to streamline access to agronomic data stored in Genovix, a proprietary database from Agronomix (a Cultura Company). ExLibris allows quicker querying across years, locations, and multiple traits by automating data storage, transformation, and access. This method enables users to concentrate on deriving insights instead of managing data. By integrating historical and current data, ExLibris makes variety advancement decisions more efficient.
• AgSkySight is a software platform developed to simplify the process of drone image stitching and vegetation index calculations in agricultural research fields. Drone technology is increasingly utilized in agriculture to capture high-resolution field images. Managing, stitching, and analyzing these images, however, remains complex and time-consuming. Current solutions often lack scalability, robust vegetation index calculations, and seamless integration with high performance computing resources. AgSkySight provides an efficient and user-friendly workflow for UAS data ingestion, image stitching, and analysis, particularly for agricultural research plots. These capabilities will aid researchers' ability to obtain timely insights from drone-captured data from their experimental plots.
• PredictPro aids in analyzing omics data (i.e., genomic and phenomic) collected from field trials. Beyond facilitating downstream bioinformatics analysis of genomic data, PredictPro equips users with powerful performance metrics and rich visualizations that promote thorough model evaluation and interpretation. PredictPro includes a model suite for comparative analysis to pinpoint the most effective predictive breeding strategies. It also features a comprehensive validation protocol to guarantee model reliability and scalability across different datasets and environments. The PredictPro pipeline is a robust process for processing and integrating multi-year, multi-location datasets, facilitating data flow and model training. Beta testing is currently underway.
• IoT Data Hub is a software application for managing, visualizing and downloading data from the AES Sensor Network. Sensors deployed at NDSU research sites transmit data wirelessly through LoRaWAN technology. Data is ingested and stored in a database for access by researchers. Examples of sensors deployed include soil temperature, soil moisture, rainfall, air temperature, relative humidity, windspeed, leaf wetness, tank liquid level, and others.

Agriculture Data Security

The security of agricultural data is a critical element for the safe and reliable production of food. As sensors, unmanned aerial and ground vehicles, artificial intelligence, and many other technologies are applied to crop and livestock production, it is critical to ensure robust approaches to security are integrated. NDSU research is taking a broad approach that analyzes the food production “system” to identify vulnerabilities and impacts. From this system-level analysis, focus areas are identified for research that is impactful towards fielding effective solutions.

• Outcomes from the research will 1) address data security concerns for sensors, tools and technologies that advance sustainable crop, livestock and integrated crop-livestock management and 2) provide improved, transparent, and end-to-end secure data architectures to support integration of new sensors and technologies in agricultural production systems.