NIA | Seminars | Marwan Bikdash

NIA Seminar by Marwan Bikdash

Date: August 3, 2005
Time: 10:30am
Location: NIA, Rm 137

Collision-free Optimal Path Planning Using Mixed Logical Linear Programming
Marwan Bikdash, North Carolina A&T State University

Scientists now believe that planets and solar systems are very common. Still we have not been able to detect one yet because a telescope capable of detecting a dark planet light years away is prohibitively large. A solution is: Instead of sending a very large spacecraft, one can send a large number of small spacecraft which can cooperatively construct a better image of the sought-after planet using a physics concept known as interferometry. A pioneering effort in this regard is the proposed NASA StarLight mission. This mission requires several formation reconfigurations of spacecraft under severe restrictions on fuel, sun exposure, and fault tolerance.

We present a formation reconfiguration algorithm based on planning collision-free optimal paths. The proposed solution consists of two steps and is suitable for reconfigurations of large formations of spacecraft. First, the formation flying problem is converted to a logical disjunction of linear programs whose unknowns are the parameters of the parameterized trajectories. Then, a two phase hybrid algorithm which uses both sequences of linear programs and mixed integer linear programming to solve the new parameter optimization problem has been established. A method that combines the cutting plane and the bisection ideas is then proposed and the algorithm becomes 30 times faster.

Dr. Marwan Bikdash, Associate Professor of Electrical and Computer Engineering, has many research projects funded by NASA, the National Science Foundation, the Office of Naval Research, the National Renewable Energy Laboratories, the Boeing Company, and others. He received his M.S. and Ph.D. in Electrical Engineering from Virginia Tech in 1990 and 1993 respectively. He has research experience in the control and analysis of nonlinear systems, signal processing, sensor fusion, vibration control, and fuzzy logic, and other intelligent systems. He teaches courses on control theory and practice.