Integral ISS-Based Cascade Stabilization for Vectored-Thrust UAVs

Abstract

We address stabilization of vectored-thrust Unmanned Aerial Vehicles (UAVs): a challenging task due to the peculiar nonlinear underactuated dynamics. According to a wellestablished approach based on the selection of suitable error variables, the error dynamics is described as a pseudo cascade connection where the attitude subsystem indirectly stabilizes the position dynamics. Unlike existing works, we address stability of this cascade using integral Input to State Stability (iISS). With this extension it is possible to employ a quasi time-optimal control law to achieve global asymptotic stabilization of any desired position, thus improving the transient performance with respect to existing control designs. A simulation example shows the performance improvement in comparison with a nested saturations stabilizer.

An Underactuated Control System Design for Adaptive Autopilot of Fixed-Wing Drones

“…The dynamics in Gazebo essentially are analogous to the dynamics described in Section II, as it can be found in most flight control books. The UAV model generated in Gazebo is a 1.0 kg standard structure fixed-wing UAV with aileron, rudder, and elevator, and where the rotor can generate airspeed in the range [10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25]m/s. Primitive paths are used during testing, i.e., straight-line paths and orbit paths [and their combination, cf.…”

“…In order to validate the benefits of the proposed adaptation, we also run a standard sliding mode version of the proposed approach, which amounts to a control action similar to ( 19)-( 20), but where the adaptation laws are switched OFF. In particular, we take θ1 = θ2 = 0 (i.e., as in standard sliding mode literature), and with θ0 = const, for various constants = (1,2,5,10,15,20). The performance of the different autopilots (called SMC autopilots) is compared with the proposed idea in Table II, in terms of tracking error with the desired path.…”

“…The autopilot layer should compensate for uncertain aerodynamics unaccounted in guidance models, but this is challenging in general [12]- [15]. For example, when roll, pitch, and yaw dynamics are linearized and simplified for cascaded loops, cross-couplings are neglected and the knowledge of equilibrium points of the Euler-Lagrange equations (trimming points) is required [5].…”

ArduPilot-Based Adaptive Autopilot: Architecture and Software-in-the-Loop Experiments

“…Many approaches to UAV autopilot deal with rotor UAVs (cf., [16]- [20]). At the same time, the flight control community has turned its attention toward fixed-wing UAVs, which often require approaches similar to largescale airplanes.…”

Artificial-Delay Adaptive Control for Underactuated Euler–Lagrange Robotics

“…Because of the reduced number of actuators, under-actuation in robotic systems has advantages in terms of lightweight, energy consumption and system simplicity. Under-actuated robotics encompass tower [23], [24] and offshore cranes [25], [26], robotic aerial [27], marine [28], [29] and ground vehicles [30], robotic hands [31], and many more. Control of underactuated systems is more challenging than the corresponding fully-actuated version.…”