Supervising Control Mechanism for UAVs to Enhance Robustness

Abstract

Small UAVs, due to their easy handling and lesser logistic support requirements are attracting lot of interest for applications ranging from military to commercial, and even for scientific data collection. In present day scenario, they are mostly used to conduct surveillance, reconnaissance, search and rescue missions for military and civilian outfits. In order to complete the assigned tasks they are usually equipped with surveillance cameras. Unlike their larger counterparts, smaller UAVs by virtue of their size cannot carry sophisticated gyro stabilized cameras. It therefore becomes important to address this shortcoming, in terms of non availability of gyro stabilization by ensuring a smoother flight even in turbulent weather to receive a stable and good quality video from UAV. Growing interest in small UAVs has motivated many equipment manufacturers to develop Off The Shelf autopilots. Being cost effective, simple, flexible and reliable, PID based autopilots are generally preferred for the purpose. Besides number of advantages, PID controllers have a limitation that they sometimes require retuning of the gains when subjected to external disturbance. UAVs perform satisfactorily if the atmospheric conditions do not change much. However, if the UAV moves to a location where the atmospheric conditions change substantially, the PID gains may require readjustment / retuning, which is a tedious job and requires numerous UAV flights; otherwise performance degradations in terms of pitching and rolling oscillations are observed. The above premise formed the basis and source of motivation for this research. It was considered that instead of going for retuning, gain scheduling or replacing the existing, conveniently available PID controller with some intricate or complex controller, a simpler, realizable and effective scheme should be adopted. The basic idea of the proposed scheme was derived from a situation in which a boy sometimes requires a bit of extra hand from his father sitting beside him, while driving through a bumpy and wavy patch of road. Father closely monitors the driving performance of his son on bumpy road and when he thinks that slight help in terms of applying an extra force by holding the steering can improve the capability of his son to negotiate the bad patch, he does so. After passing through bumpy part of road when father senses that the extra effort is no more required, he gradually removes it and his son takes over the car back again as a solo driver. Following the same analogy, the idea was conceived to propose a supervising mechanism which can act like a human pilot and can be augmented with an existing control system of a UAV for improving its performance in presence of atmospheric disturbances without going into exercise of retuning the controller gains. Our supervisory mechanism is composed of two modules, “observer module” and “correction generator module”. We developed a human thinking like logic for observer module so that it keeps monitoring the status of flight through specified inputs and outputs of the system and instructs the correction generator module to augment main controller by adding compensation commands when required i.e. in case undesired attitude error is observed. The second module for correction generation is based on Fuzzy rules, where rule set was developed using input from a UAV pilot. Effort is made to maintain a logical layout of this thesis so that the reader can be familiarized with the theoretical concepts and analysis tools related to UAVs. A small UAV with already published aerodynamic data is taken as a reference bird. In chapter two and three of this thesis detailed explanation about the formulation of its nonlinear mathematical modeling, linearization process to achieve decoupling and simplicity, conducting static and dynamic stability analysis, criteria for flying qualities assessment is given. Designing of PID controllers for attitude, altitude and heading control using linearized UAV models is covered in chapter four. Performance of designed controller is also validated by using nonlinear UAV model. Atmospheric disturbances such as wind shear and turbulence significantly influence the attitude of UAVs. Rotary moments, rolling moments, vertical and lateral wind gusts generated due to moderate to severe turbulence are simulated as atmospheric disturbance. Pitch and roll response in presence of these disturbances are presented. In chapter five the functionality of proposed scheme is discussed in detail and simulations are run to demonstrate improved pitch and roll response under atmospheric disturbance by augmenting the supervisory mechanism. Simulation results showing significant reduction in attitude errors with slight increase in control effort after augmenting the supervisory mechanism are presented. Simulations are run using linear as well as nonlinear model of UAV. Comparison of results obtained from standalone PID controller and augmented system is presented to validate the efficacy of proposed scheme Keywords; UAV Modeling, Stability analysis of UAVs, UAV autopilots, Disturbance rejection in UAVs