The objective of this project is to adapt and apply modern nonlinear control techniques to models of HIV infection, leading to new methods of patient evaluation and treatment. Experimental data and mathematical models concerned with the interaction between HIV, T cell responses, and Highly Active Anti-Retroviral Therapy (HAART) suggest that long-term immunological control of HIV may be established by a short-term schedule of HAART and then maintained in the absence of further treatment. In order to evaluate this possibility, techniques must be developed to match the model behavior to measured patient data; schedules of HAART which are predicted to induce long-term immunological control of HIV must be calculated. These schedules will need to take into account unmodeled disturbances and inaccurate measurements, as well as the systemic cost and other risk factors, using measurements representing only a subset of the critical compartments in the models. Nonlinear Control techniques have been successfully applied to similar problems in a variety of other disciplines. The specific application proposed in this research introduces a number of elements not previously accounted for. The project will extend and apply the framework of Model Predictive Control and the techniques of online identification and output feedback to the design of treatment schedules for HIV infected patients.
| Zurakowski, Ryan (2011) Nonlinear observer output-feedback MPC treatment scheduling for HIV. Biomed Eng Online 10:40 |
| Zurakowski, Ryan; Wodarz, Dominik (2007) Model-driven approaches for in vitro combination therapy using ONYX-015 replicating oncolytic adenovirus. J Theor Biol 245:1-8 |
| Zurakowski, Ryan; Teel, Andrew R (2006) A model predictive control based scheduling method for HIV therapy. J Theor Biol 238:368-82 |
