Progress 07/15/20 to 03/14/21

Outputs
Target Audience: Nothing Reported Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Various calibration models have been developed. We explored linear (PLS) and machine learning models for classification of samples and also to model properties. We validated our models to maximize prediction and avoid overfitting. We developed the software to integrate sensor control middleware board, the sensor, iOS application, and backend software (and database) on the cloud. We demonstrated the end-to-end data trip and decision support in the field conditions.

Publications

Progress 07/15/19 to 03/14/21

Outputs
Target Audience:The addressable market for our product is the broader agricultural testing market which will be worth $6.29 Billion by 2022 (15). In the US, there are 2 million farms covering an area of 922 million acres (i.e., an average of 420 acres per farm). Using the range (1 sample/acre to 40 samples/acre (in precision agriculture)), and a once in 3 years soil assessment assumption, we estimate that the number of soil tests required to reach 10% of the US farms ranges from 30 million to 1.2 billion per year. Our market audiences include several large food & beverage companies, experts at the Rutgers University soil-testing laboratory, the experts at the Northeast Organic Farming Association (NOFA) of NJ, etc. Our customers are independent growers, soil consultants, precision agriculture (PA) equipment manufacturers, and food & beverage companies. Our primary go-to-marketstrategy is to tap into the resources of organizations like NOFA and regional soil-testing laboratories to market directly to our prospective customers. Organizations like NOFA interact closely with growers and are advocates of proactive soil management. Regional soil testing laboratories will be a source of region-specific ground-truth data for our product and we will forge win-win relationships with them. As our product primarily supports precision agriculture and intense spatiotemporal soil analysis, it is not a substitute for the highly accurate soil analyses conducted at the regional soil laboratories. Hence the regional soil testing laboratories and organizations like NOFA and NJ Farm Bureau can be a great channel partner for our product. Given the emergence of many new technology startups in the precision agriculture space, organizations like NOFA are ready to collaborate with budding companies (see support letter) and we will leverage such opportunities to educate a larger pool of end users for our product. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Hardware/spectral sensor modules wereacquired and assembled for Visible, MIR, and NIR spectral acquisition. The hardware, in prototype form, was assembled, tested, and tuned. We analyzed over 1,000 samples of soils. We collected data from Rutgers University's soil testing laboratory. Various calibration models have been developed. We explored linear (PLS) and machine learning models for the classification of samples and also to model properties. We validated our models to maximize prediction and avoid overfitting. We developed the software to integrate sensor control middleware board, the sensor, iOS application, and backend software (and database) on the cloud. We demonstrated the end-to-end data trip and decision support in the field conditions.

Publications

Progress 07/15/19 to 07/14/20

Outputs
Target Audience:The goal of this SBIR Phase I project is to conduct research and prove the feasibility of measuring soil properties using cuttingedge detectors, customized optical accessories, and data processing algorithms. Our overall project goal is to bring to market a portable soil sensor which can analyze soil samples in near real-time, in Phase II and beyond. Growers rely on soil analysis to help them assess the available levels of plant nutrients, in order to plan fertilization, tillage, and planting tasks. Soil analysis and the remedial action taken have a direct bearing on crop quality and yield and it affects downstream manufacturing operations in food & beverage companies (like harvesting, food processing, and packaging) significantly. The conventional approach to soil-analysis involves sending samples to far-off laboratories, and results may be delivered days later. This approach is time-consuming, tedious, and laborious. Today, growers may choose to analyze soils less frequently and only in some fields and, at times, may choose not to analyze soil at all. As a result, soil treatments may not fit the soil/crop requirements and therefore conform to nutrient stewardship principles of proper fertilizer grade, rate, and timing. With over 2 million farms in the US alone, the commercial opportunity for our idea is significant. Our phase I objectives are to 1) assemble a sensor comprising of miniature detectors and optical accessories, 2) develop algorithms to learn the mapping between spectra data and soil properties, and 3) demonstrate feasibility of real-time soil sensing on a bench-scale hardware-software setup. This project directly addresses USDA's / NIFA's soil and water conservation goals. As the world faces increased demand for food, it is imperative that resources like land, water, and fertilizers are managed proactively and optimally. Our product will enable growers to be more proactive in soil treatment and hence consume optimal water and fertilizer amounts while protecting soil and water quality. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals? We will build calibration models. We will validate our models.

Impacts
What was accomplished under these goals? Hardware/spectral sensor modules was acquired and assembled for Visible, MIR, and NIR spectral acquisition. The hardware, in prototype form was assembled, tested, and tuned. We analyzed over 1,000 samples of soils. We collected data from Rutgers University's soil testing laboratory.

Publications