Method for simultaneous localization and parameter estimation in particle tracking experiments.

We present a numerical method for the simultaneous localization and parameter estimation of a fluorescent particle undergoing a discrete-time continuous-state Markov process. In particular, implementation of the method proposed in this work yields an approximation to the posterior density of the particle positions over time in addition to maximum likelihood estimates of fixed, unknown parameters. The method employs sequential Monte Carlo methods and can take into account complex, potentially nonlinear noise models, including shot noise and camera-specific readout noise, as well as a wide variety of motion models and observation models, including those representing recent engineered point spread functions. We demonstrate the technique by applying it to four scenarios, including a particle undergoing free, confined, and tethered diffusions.