A Fast Convergence Super‐Twisting Observer Design for an Autonomous Airship

Abstract

Airships are often under the risk of external disturbances, for example, the wind. The accurate external disturbance of the airship are hard to acquire. Typically, the super-twisting observer can track the states for mechanical systems and could be used as an external disturbance tracker. However, the convergence rate of the traditional super-twisting observer is very slow and the observed states often oscillate heavily. We modified the traditional super-twisting observer by adding more variables to the sliding surface, and adding parameters to the newly added variables. Simulation results illustrates that the new proposed observer is able to reduce the convergence time and reduce overshoot greatly, improving the performance of the observer.

Adaptive high‐order sliding mode controller‐observer for MIMO uncertain nonlinear systems

“…Secondly, based on the F-TOSMLO, the controller performances obtained by applying the proposed AHOSMC are compared with those obtained with the adaptive super twisting sliding mode controller ASTSMC proposed in. [38] For the first part of the simulation, the observer gains for both F-SOSMO and F-TOSMLO are selected as: α 2 = 1.9( F + ) 1/3 , α 1 = 1.5( F + ) 1/2 , whereas for F-TOSMLO α 0 = 1.1 F + ,( F + > 0); this choice meets the requirement in [15,27] and guarantees the finite-time stability of (12) and (22), and the linear gains are set through computer simulation as β 1 = 45 and β 2 = 35. For uncertainty reconstruction in F-SOSMO, a low pass filter is used with sampling time Λ = 10 −4 s and τ z = Λ 1/2 s. The comparison between the velocities estimation for F-SOSMO and F-TOSMLO is shown in Figure 2.…”

“…Among the most effective and appealing HOSMs in real applications, the so‐called super‐twisting algorithm (STA) introduced in,[21] is very useful for both control and observation. [17, 22] Contrary to other HOSM algorithms and instead of a classical sliding mode, STA is designed for systems with relative degree one with respect to the sliding variable σ . Hence, only information about σ is required for ST control design.…”

Trajectory Tracking Control for a Stratospheric Airship Subject to Constraints and Unknown Disturbances

“…A radial basis function neural network was applied to compensate for the unknown wind field in [31]. The author of [32] proposed a novel super-twisting disturbance observer to improve the convergent rate of disturbance tracking. Compared with the techniques above, the extended state observer regards the lumped disturbance as new state to compensate and has no need for prior information about the bounds of disturbance.…”