High order sliding mode control for restoration of a population of predators in a Lotka-Volterra system.

Human-induced extinction and rapid ecological changes require the development of techniques that can help avoid extinction of endangered species. The most used strategy to avoid extinction is reintroduction of the endangered species, but only 31% of these attempts are successful and they require up to 15 years for their results to be evaluated. In this research, we propose a novel strategy that improves the chances of survival of endangered predators, like lynx, by controlling only the availability of prey. To simulate the prey-predator relationship we used a Lotka-Volterra model to analyze the effects of varying prey availability on the size of the predator population. We calculate the number of prey necessary to support the predator population using a high-order sliding mode control (HOSMC) that maintains the predator population at the desired level. In the wild, nature introduces significant and complex uncertainties that affect species' survival. This complexity suggests that HOSMC is a good choice of controller because it is robust to variability and does not require prior knowledge of system parameters. These parameters can also be time varying. The output measurement required by the HOSMC is the number of predators. It can be obtained using continuous monitoring of environmental DNA that measures the number of lynxes and prey in a specific geographic area. The controller efficiency in the presence of these parametric uncertainties was demonstrated with a numerical simulation, where random perturbations were forced in all four model parameters at each simulation step, and the controller provides the specific prey input that will maintain the predator population. The simulation demonstrates how HOSMC can increase and maintain an endangered population (lynx) in just 21-26 months by regulating the food supply (hares), with an acceptable maximal steady-state error of 3%.

Accurate diagnosis of sepsis using a neural network: Pilot study using routine clinical variables.

BACKGROUND AND OBJECTIVES: Sepsis is a severe infection that increases mortality risk and is one if the main causes of death in intensive care units. Accurate detection is key to successful interventions, but diagnosis of sepsis is complicated because the initial signs and symptoms are not specific. Biomarkers that have been proposed have low specificity and sensitivity, are expensive, and not available in every hospital. In this study, we propose the use of artificial intelligence in the form of a neural network to diagnose sepsis using only common laboratory tests and vital signs that are routine and widely available. METHODS: A retrospective, cross sectional cohort of 113 patients from an intensive care unit, each with 48 routinely evaluated vital signs and biochemical parameters was used to train, validate and test a neural network with 48 inputs, 10 neurons in a single hidden layer and one output. The sensitivity and specificity of the neural network as a point sampled diagnostic test was calculated. RESULTS: All but one case were correctly diagnosed by the neural network, with 91% sensitivity and 100% specificity in the validation data set, and 100% sensitivity and specificity in the test data set. CONCLUSIONS: The designed neural network system can identify patients with sepsis, with minimal resources using standard laboratory tests widely available in most health care facilities. This should reduce the burden on the medical staff of a difficult diagnosis and should improve outcomes for patients with sepsis.