Flight Control of a Quadrotor under Model Uncertainties

Abstract

Quadrotor type helicopters continue to grow in popularity for academic research and unmanned aerial vehicle applications. However, the model uncertainties caused by battery voltage drop, payload variation and flight condition change, have rarely been considered in control design. This work proposes a quantitative feedback theory based robust design approach to deal with these uncertainties. By analyzing the rigid body dynamics and aerodynamic forces/moments under different voltages, payloads and flight conditions, we model quadrotor dynamics as a set of linear models with parameter uncertainties, which represent a larger flight envelop than models linearized from hover condition. These model uncertainties, as well as the robust stability requirements and performance specifications of flight control system are then used for designing and tuning the controllers, to perform tradeoff between controller complexity, robust requirements, and performance specifications. We also implemented a prototype system, and conducted a serial of experiments in realtime outdoor flights to evaluate its performance. The results show good, robust, and reliable performances of the designed system in autonomous hovering, takeoff, waypoint navigation and landing flights. NOMENCLATURE INTRODUCTIONQuadrotor has become a popular unmanned aerial vehicle (UAV) platform. Compared to conventional helicopter which has the same capabilities of vertical takeoff and landing (VTOL), hovering, and low speed flying, quadrotor has three main advantages. First, quadrotor is lifted and propelled by four sensorless brushless direct current (SLBLDC) motor driven rotors, eliminating complex mechanical structure such as swashplates and linkages, thus is inherently more reliable and maintainable. Second, for given load capacity, the individual rotor of quadrotor is smaller than the main rotor of helicopter, leading to smaller damage when crashing and safer operation when flying over densely populated areas. Third, due to smaller rotors and symmetrical structure, the vibration of quadrotor is much smaller than that of helicopter. These advantages make quadrotor widely used in applications such as aerial photography, surveillance, patrolling, search and rescue, law enforcement, aerial mapping, and so on. In most of these applications, the automatic flying ability of quadrotor is demanded. However, the open-loop unstable nonlinear dynamics and limited payload of quadrotor bring both theoretical and technical challenges to control system design. Although significant progresses have been made on the control of quadrotor, most works rarely concern about the model uncertainties in practical implementation. These model uncertainties are mainly due to three factors: 1) Battery voltage drop. As will be illustrated in Section 3.2.4, the dynamics of rotors change distinctly with battery voltage. 2) Payload variation. The moments of inertia and mass of quadrotor may change with payload, bringing variation of model parameters. 3) Flight condition change. For model...

Cuckoo search-based method for trajectory planning of quadrotor in an urban environment

“…First, the influence of N is discussed. As reported in Xu et al., 17 CS algorithm can obtain a good result when N = 15 to 50 in most of function optimization problem. To test the reasonably of this point in the trajectory planning problem, different values of N are assigned to run the experiments and other parameters are the same with those from Tables 1 to 3.…”

“…As reported in Xu et al, 17 p a ¼ 0.25 are sufficient for most optimization problem, and this is also true in the trajectory planning problem for quadrotor delivering in urban environment. According to Table 7, a smaller or bigger value of p a both makes the best solution worse because a low probability of replacing the solutions may reduce the possibility of finding a better solution, while bigger probability of replacing the solutions may destroy the good solutions and make the best solution missed.…”

“…Equations (17) and (18) guarantee that the quadrotor can fly without inclining too much and the goods are also in good condition.…”

“…The influence of wind field is taken into account both in trajectory planning and position control problem. 17 When the wind disturbance is considered in position control problem, the flight controller has to compensate for trajectory deviations, and therefore the energy consumption becomes an issue. Based on the above considerations, a linear variation of wind velocity is added to the kinematic model, 18 and a promising path planning strategy is proposed which the quadrotor would take advantage of wind to reach the destination, namely, saving time and energy.…”

Optimal control and state estimation for unmanned aerial vehicle under random vibration and uncertainty

“…Relatedly, Al-Sabban et al 10 presented the Markov decision processes for planning drone flights under uncertain wind direction conditions, in which time-varying wind was considered. Moreover, Xu et al 11 suggested a quantitative feedback theory-based robust design approach by considering the model quadrotor dynamics as a set of linear models with parameter uncertainties. Then, Jaeger et al presented a procedure for reliable and robust optimization of an aircraft at the conceptual design phase.…”