Home » Books » Robotics » Advances in Robotics: Modeling, Control and Applications » Fault Accommodation In Discrete-Event Robots

Title

Advances in Robotics: Modeling, Control and Applications

Editor

Calin Ciufudean and Lino García

Price

USD$139.00

ISBN

978-1-461108-443

Clicks

33270

Viewed: 1199

Discrete-events robots are complex engineering systems guided by flexible and event adaptive algorithms aimed at application in unstructured, unsafe environments in which faults may cause undesirable losses. The demand on fault tolerance imposed on such class of the systems calls for the use of fault adaptation techniques. There exist two principle ways for adaptation to faults. The first one is self-tuning or fault accommodation. It is related to on-line control law determination that preserves the main performances of the robot in faulty case while the minor performances may degrade. The second way is self-organization which involves the robot reconfiguration to replace the faulty elements with healthy ones. This chapter is concentrated on the fault accommodation problem in discrete-event robots. Up to now, different solutions have been proposed to above problem. All these solutions involved the models of the system under monitoring in the form of linear or nonlinear ordinary differential equations, see e.g. monograph [1], papers [2-7]. Conventional solution of the problem assumes on-line fault detection and estimation to construct the model of faulty system (so-called model tuning) followed by the new control law determination on the base of the tuned model [1-5]. In [6, 7], another approach has been proposed whose feature is the use of full decoupling with respect to fault effects in output space of the system. In contrast to the conventional approach, this approach does not need in fault estimation. Therefore, such approach looks reasonable if on-line fault estimation is problematic. Also, it allows decreasing time expanses for fault accommodation because of excluding the stage of the model tuning. An original solution of the problem considered in present chapter involves finite automation model for the robot specification. Let the work of robot be characterized by fulfilling the final set of the tasks. Each of these tasks can be considered as appropriate automaton state. The automaton inputs which initiate the transitions from one to another tasks solution are generated according to some initial program (which determines the automaton transition function) taking into account the result of previous task solution (the automaton output). Faults in the robot may cause violation of the sequence of tasks under solution that corresponds to distortion of the automaton transition function. Solution of fault accommodation problem in this case is aimed at automaton control such that results in admissible sequence of tasks under fault conditions. The chapter is structured as follows. 1. Introduction. In this section, the problem of fault accommodation in discrete-event robots is discussed. 2. Problem formulation. In this section, the problem of fault accommodation is formulated and the main steps of the problem solution are described. 3. Mathematical techniques. This section gives brief sketch of pair algebra of partitions used for solving the problem of fault accommodation. 4. The auxiliary automaton design. This section gives a procedure of auxiliary automaton design which is a basis to develop the designing procedures. 5. Control law determination. The goal of this section is to develop the designing procedures for control law. 6. The automaton A* design. This section gives a procedure of the map design describing the automaton A*. 7. Example. In this section, an example is presented. 8. Conclusions. This section concludes the chapter. References [1] Blanke M., Kinnaert M., Lunze J., Staroswiecki M. â€œDiagnosis and Fault Tolerant Controlâ€. Springer-Verlag. 2003. [2] Patton, R.J. â€œFault tolerant control: The 1997 situationâ€. In Proc. of IFAC Symposium Safeprocessâ€™97. Hull, UK. pp. 1033-1055. [3] Staroswiecki, M. â€œFault tolerant control: the pseudo-inverse method revisitedâ€. In Proc. of 16th IFAC Congress. Prague, Czech. Republic, 2005. [4] Staroswiecki, M., H. Yang and B. Jiang. â€œProgressive accommodation of aircraft actuator faultsâ€. In Proc. of IFAC Symposium Safeprocessâ€™2006. Beijing, pp. 877-882. [5] Weng Z., R. Patton and P. Cui. â€œActive fault-tolerant control of a double inverted pendulumâ€. In Proc. of IFAC Symposium Safeprocessâ€™2006. Beijing, pp.1591-1596. [6] Shumsky, A., Zhirabok N., Jiang, B. and Ke Zhang. â€œFault accommodation in dynamic systems: fault decoupling based approachâ€. In Proc. of IEEE CDCâ€™2009. Shanghai, PR China. 8464-8469. [7] Shumsky, A., Zhirabok A. â€œUnified approach to the problem of full decoupling via output feedbackâ€ // European Journal of Control, 2010, Vol. 16. â„– 4. P. 313-325. The short list of authorâ€™s publication related to above chapter [8] Shumsky, A., Zhirabok A. â€œMethod of fault accommodation in finite automatonâ€, Automation and Remote Control, 2010, No. 5. P. 122-133. [9] Shumsky A., Zhirabok A. â€œFault accommodation in discrete-event systemsâ€, Proc. 18th Mediterranean Conf. on Contr. & Autom. M

Alexey Zhirabok

*Far Eastern Federal University, Russia*

Alexey Shumsky

*Far Eastern Federal University, Russia*