Abstract
In designing optimal servocontrollers for mobile robots, a quadratic cost function of states and control inputs is sometimes used [33, 53, 107]. In this design, two unknown symmetric weighting matrices — Q and R — are incorporated normally by designers. Their selection is only weakly connected to performance specifications, and certain trial and error is usually required with an interactive computer program before a satisfactory design results [53]. Regarding this, some guidelines exist in the literature. After selecting suitable weighting matrices, optimal gains are computed based on an algebraic Riccati equation solution that optimizes the fitness function — a rather complex and indirect process. It is also tedious to find stable-optimal gains to make the system stable and provide minimum servotracking error of a control system by off-line trial and error method [83, 127]. To demand stability and minimum trajectory error, one must find a trade-off between them. To alleviate this trade-off, it is possible to tune stable-optimal gains automatically by constructing an appropriate fitness function using evolutionary algorithms [61]. Consequently, it will provide the controller design process with an automatic notion.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2004 Springer-Verlag Berlin Heidelberg
About this chapter
Cite this chapter
Watanabe, K., Hashem, M.M.A. (2004). Evolutionary Design of Robot Controllers. In: Evolutionary Computations. Studies in Fuzziness and Soft Computing, vol 147. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-39883-7_6
Download citation
DOI: https://doi.org/10.1007/978-3-540-39883-7_6
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-05887-5
Online ISBN: 978-3-540-39883-7
eBook Packages: Springer Book Archive