CIF: Small: Learning, Optimization & Analysis for Biologically Inspired Community Networks

Bacteria are single-celled organisms that constitute some of the earliest forms of life. Bacteria have an aggregate biomass that is larger than that of all animals and plants combined. Despite their simplicity, the operations and interactions of bacteria are not fully understood, yet microbial communities play a significant role in bioremediation, plant growth promotion, human and animal digestion, disease, the carbon-cycle and the cleaning of water. Thus, it is of significant interest to further understand microbial populations. This project uses methods from communications and signal processing to model, optimize and hopefully further understand the interactions within microbial populations. The particular focus is on the phenomenon of quorum sensing wherein bacteria express new genes, under the right environmental and population conditions, which facilitates collective behavior such as biofilm formation as well as virulent behavior such as infections. Thus, quorum sensing enables these simple unicellular organisms to achieve complex, collaborative tasks. The project will train graduate students in cross-disciplinary research and potentially impact applications and industries such as microbial fuel cells, bacterial infection suppression, commercial cellular systems, search-and-rescue, transportation systems, environmental monitoring, and homeland security. Modeling, learning and optimization strategies for biologically inspired problems will be considered. Several unique challenges will be directly addressed: (1) developing signaling and channel models that enable optimization and control, while capturing key operational behavior of the bacteria; (2) designing meaningful cost functions that are relevant to the microbial application; and (3) designing models addressing the inherent coupling between individual bacteria as well as coupling between the bacteria and their environment. The project views quorum sensing as a critical element of a multi-agent, distributed decision-making system that is the bacterial community. A key challenge is to bring everything together: the incorporation of biological constraints, the coupling between microbes, the challenge of determining the proper objective function that captures biological fitness or other microbial objectives in the context of temporal and spatial evolution of microbial colonies. The following tools will be leveraged: decentralized detection/decision making, sequential decision making, active localization, active sampling, stochastic geometry, causal graph identification, directed information, two-way communication, multi-hopped communication, and packet scheduling methodologies. While the proposed work does not have a significant experimental content, previously collected experimental data in the context of existing collaborations will be leveraged to drive the proposed research. The project aims to close the loop by not only using communication, signal processing and information theory to model and design microbial systems, but to use the outcomes of this biologically inspired research to design engineered systems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.