AlgaSens: An AI-enabled Platform for Cost-effective Automated Characterization of Algae and Other Micro-objects for Optimizing Research andCultivation in Bioproducts & Biofuels.

Algae are rapidly growing in importance as an abundant and renewable source of thousands of valuable components that are cultivated for applications in biofuels and other bioproducts, such as additives for food, nutrition supplements, pharmaceuticals, chemicals, etc. Growing algae is the largest cost in algal products today, and algae crop loss due to microscopic pests represents one of the major reasons for this cost. This necessitates development of cost-efficient and automated technologies to empower methodologies for algae crop protection, as well as to optimize algae research and cultivation processes for biofuels and other bioproducts applications. An innovative platform is developed for automated high-throughput monitoring and characterization of microalgae during cultivation process for biofuel and other bioproducts production in-field or for laboratory research. Specifically, the proposed platform will be able to 1) monitor and characterize the cultivated microalgae samples with resolution down to individual alga (including specific classification of algae, pests and other microobjects); 2) automatically estimate in a label-free way the individual alga characteristics, such as shape, size, color and lipid content, empowering precise decision making to adjust cultivation process or to initiate harvesting; 3) empower crop protection methodologies by early detection of microscopic pests and provide advanced warning to enable the corresponding risk-mitigation strategies. During Phase I, the following objectives were accomplished: 1) developed an initial prototype version of the proposed platform adapted for monitoring and characterization of dense microalgae samples, 2) developed deep learning enabled image processing algorithms to characterize identified microalgae and to detect pests and other micro-objects to empower crop protection methodologies, 3) developed a metric for individual algae lipid content estimation, 4) performed proof of concept evaluation of the platform in the laboratory and in-field during a deployment campaigns at a partner facility where algae were cultivated and monitored. During Phase II the final prototype will be integrated, optimized and developed to be robust and cost- effective. It will incorporate embedded computing to enable portability and long-term unattended operation. New modes of operation, DENSE and NORMAL, will be developed for the device. In DENSE mode the system will enable concentrated sample processing at high flowrate aiming to identify and quantify pests. This mode will be designed for long-term unattended deployments. For this mode, a new sensor will also be developed enabling estimation of biomass concentration. The NORMAL mode will be developed for laboratory operations and for processing of diluted samples with high accuracy. The focus of this mode will be on performing detailed analysis. Furthermore, an advanced lipids measurement algorithm will be developed to significantly complement existing algae analytics. Finally, the device will be integrated and evaluated in the laboratory and during several long-term in-field experimental campaigns at partner facilities. The proposed platform will have significant commercial applications for the intended market of algae-based biofuel and bioproducts commercial producers, research laboratories and government organizations. There also are several secondary markets for the underlying technology, including in aquaculture, public health, drinking water safety, microbial products, clinical research and pharma.