Micro helicopters are heading into a burning building. Unknown terrain. Unknown obstacles. Deploying a network of autonomous vehicles in such a dynamic environment to perform complex and intricate tasks is clearly a challenging control problem. Mazen Farhood, who joined the aerospace and ocean engineering faculty in 2008, focuses on the cooperative control and coordination of multi-vehicle systems in complex environments.
The trajectories have to be designed on-the-fly, and the vehicles must track these trajectories as accurately as possible regardless of various uncertainties and exogenous disturbances. “I propose a unified approach for path planning and regulation,” remarks Farhood. The control strategy also has to change significantly depending on the position of the vehicle with respect to the obstacles. “The vehicles of interest are generally underactuated. There are numerous critical points on the vehicle that require attention but only a limited number of available control inputs. So, the control strategy has to change significantly in order to prioritize the regulation of certain outputs over others according to the vehicle configuration and position relative to the obstacles.”
The vehicle models must capture the possible nonlinear dynamics while being amenable to control synthesis. “We are talking about the aggressive maneuvering of vehicles that have complex dynamics. So, simple linear time-invariant models won’t do in general. I propose a number of models to work with,” adds Farhood.
Such cooperative control problems are usually complex and computationally intensive. “We try to reduce the complexity by reducing the order of the models we are working with and devising simplified control algorithms that exploit the structural properties of the distributed system.”
When the controller is developed, the work isn’t complete, however. It is time for performance analysis of the designed control system. “We have to make sure that the controller performance is satisfactory despite all the system uncertainties that we justifiably ignored to simplify the control design process. Such uncertainties may include model nonlinearities, aerodynamic parameters, communication latency and packet loss, noise and disturbances, etc.”
The theories Farhood works with are not restricted to any particular vehicular system. Military applications of this project include coordinated attacks, such as hitting targets and refueling, while civilian applications include firefighting and security.
Farhood earned his B.Engr. in mechanical engineering from American University of Beirut, Lebanon in 1999, and his M.S.M.E. and Ph.D. from the University of Illinois at Urbana-Champaign in 2001 and 2005, respectively.
Before coming to Virginia Tech, he was a scientific researcher at the Delft University of Technology, The Netherlands. There he developed integral quadratic constraint (IQC) analysis and synthesis tools as part of a European Space Agency (ESA) project. Before that, he served as a postdoctoral fellow in aerospace engineering at the Georgia Institute of Technology, where he developed linear parameter-varying (LPV) techniques for the control of autonomous aerial vehicles about trajectories in obstacle environments.