Time-varying sliding modes for second and third order systems (Q2518264)

This monograph discusses the design of switching lines (or surfaces) for sliding-mode control of systems of the type \(\dot x_1=x_2\), \(\dot x_2=f(x,t)+b(x,t)u+\Delta(x,t)\) (and, similarly, third-order systems). In the right-hand side \(f\) and \(b\) are assumed to be known (and bounded), and \(\Delta\) represents the bounded uncertainty. The output of the system is assumed to be the whole state such that the goal of the control is to drive the \textit{tracking error} \((e_1,e_2)=(x_{d,1}-x_1, \dot x_{d,1}- x_2)\) to zero. First, the authors give a short introduction of the fundamental ideas behind sliding mode control: for example, for third-order systems, if one chooses a switching line of the type \(s(e)=e_3+c_2e_2+c_1e_1+A+Bt\) such that the polynomial \(\lambda^2+c_2\lambda+c_1\) is Hurwitz then the control \(u\) only has to guarantee that the sliding condition for the surface \(s\) is satisfied (which is easy under the conditions listed above). The main part of the monograph gives the detailed derivation of optimal choices of the parameters \(c_1\), \(c_2\) and \(B\). Optimality is understood with respect to the Integral of the Tracking Error (IAE) or the Integral of the Tracking Error multiplied by time (ITAE). The authors also derive optimal choices (with respect to IAE and ITAE) under various contraints: bounds on the input signal, on the velocity or acceleration, or combinations of these bounds. All derivations are carried out explicitly, arriving at analytical formulas and graphs, which may make the book most useful for engineers and students in the control design community who want to see the derivation behind classical control design prinicples.