Integrated Guidance and Control of UAVs for Reactive Collision Avoidance

Abstract : Unmanned aerial vehicles (UAVs) employed in low altitude flights are liable to collide with urban structures. The present work focuses on the reactive obstacle avoidance problem for unaccountable obstacles like urban edifices, poles, etc. Unlike existing literature that mainly proposes avoidance maneuvers using kinematic and point mass models, an innovative Six-Degree of Freedom model based partial integrated guidance and control (PIGC) approach was used. PIGC performs the avoidance maneuver in the cascaded two loop structure to overcome IGC approach shortcomings, and reduces the delay in multiple loop tracking. The PIGC Six-DOF model uses nonlinear dynamic inversion technique, is computationally inexpensive, and can be implemented on onboard UAV microcomputers. A robustness study for large numbers of simulations was performed by randomly perturbing coefficients and inertia terms and clearly shows that the neuro-adaptive augmented PIGC design is more robust to parameter perturbations compared to nominal control when applied to the perturbed plant model. In all simulations, all constraints posed by the vehicle capability are very well met within the available time-to-go.