This thesis addresses the guidance an‎d formation control of a group of agents around an unknown source in GPS-denied environments, where only field signal measurements are available. The objective is to achieve both stable convergence to the maximum point of the field an‎d the preservation of a desired geometric formation. The proposed framework is based on a “boundary–interior” architecture: boundary agents are guided toward the source via a biased extremum-seeking law, while interior agents maintain the formation using neighborhood-based control rules. As a foundation, for a single double-integrator agent, a biased extremum-seeking law combined with a lead–lag compensator is designed an‎d analyzed; the framework is then extended to multi-agent systems with both single- an‎d double-integrator dynamics. By employing averaging theory an‎d Lyapunov functions, sufficient conditions for exponential stability are derived, including the spectral requirement that “the largest eigenvalue of the symmetric part of the interaction matrix must be negative.” Clear mappings are also established between bias parameters an‎d geometric properties of the formation (such as damping an‎d phase), as well as practical constraints on the gain, zero, an‎d poles of the compensator. Simulation results demonstrate that, with minimal information exchange an‎d without requiring source coordinates, the agents stably converge around the maximum point, yielding circular, linear, spiral, an‎d double-line formations with adjustable radii an‎d smooth transient responses.

محسن مجيري فروشاني

ايمان ايزدي نجف آبادي , مرضيه كمالي