Estimation of dynamic populations for biology, defence and space applications

Estimation problems arise whenever one must infer the value of some physical quantity of interest from a stream of incomplete and/or noisy data. They are relevant to a large and surprisingly diverse set of physical problems. At Heriot-Watt University, the research group in multi-object systems focuses on estimation problems in which the objects of interest form a population whose size is unknown and usually time-varying -- e.g. proteins at the surface of a cell in cellular biology, vehicles on a battlefield in defence systems -- with applications in cellular biology, maritime surveillance and space situational awareness. While these domains appear strikingly different, we will highlight in this talk the similarities of the underlying estimation problems, explain why we are exploiting a single estimation framework to approach them, and illustrate our activity on a few ongoing projects spanning across the three application domains.