Control and stability of a torsional elastic robot arm (Q1972560)

The analysis of vibrations of a cantilevered beam with a mass attached to the free end has been subject of several studies. Among others is the recent article of \textit{K.} and \textit{Z. Liu} [J. Comput. Appl. Math. 114, 1-10 (2000; Zbl 0969.74033)], obtaining some results which parallel those of the present authors. However, the latter specialize their work to control and stability of a robotic arm. The authors model its transverse vibrations utilizing the Euler-Bernoulli beam clamped at \(x= 0\), with internal damping of Voigt type: \[ {\rho\partial^2 y\over\partial t^2}+ 2\delta EI {\partial^5 y\over\partial t} \partial x^4+ EI {\partial^4 y\over\partial x^4}= -x {\rho\delta^2\theta\over\partial t^2}, \] with clamped boundary conditions \(y(t,)= \partial y/\partial x(t,0)= 0\). The torsional vibrations are governed by the equation: \[ {\rho\partial^2}\phi/\partial t^2- 2\delta GJ {\partial^3 \phi\over\partial t},\partial x^2+ (GJ/k^2){\partial^2\phi\over\partial t^2}= 0, \] with boundary condition: \(\phi(t,0)= 0\). The authors compute the kinetic energy of the end mass and derive coupled bending and torsional vibration equations as the Euler-Lagrange equations for this system. They choose the appropriate inner product and reduce the problem to an evolution equation \[ {d{\mathbf u}\over dt}= A{\mathbf u}+ F(t,{\mathbf u}). \] The corresponding homogeneous evolution equation is \(d{\mathbf u}/dt= A{\mathbf u}(t)\). They prove that the operator \(A\) is densely defined, positive, and self-adjoint, and that the inverse exists and \(A^{-1}\) is compact. While Liu and Liu proved that the inverse of the Euler-Bernoulli operator with Voigt damping is compact using a semigroup theory approach, the present authors prove the result for the same operator but incorporating coupled torsional vibrations. The authors of this article also use the semigroup generated by \(A\) to prove existence and uniqueness of solutions of the coupled vibrating system. They show that existence of weak solutions implies existence of strong solutions which are asymptotically stable. They also show that their system is controllable, meaning that the robot arm will arrive at any designated position, while the coupled bending and torsional vibrations are exponentially suppressed. The article is well-written and derives important results. The reviewer saddly comments that Voigt damping is not supported by tests of most engineering structural materials. However realistic internal damping of materials such as medium carbon steel, and even more so of light-weight alloys seems to be modelled by differential operators of fractional orders, and so far for only a few researchers had the courage of offer some partial results for very simple systems.