Biomimetic and Biologically Inspired Control - Biomimetics: Biologically Inspired Technologies - page 423

In this chapter, we have introduced some of the recent researches from the point of view of
biomimetic control. Especially we have concentrated our discussions on how to organize the system
redundancy, the optimal motion formation, and the environmental adaptive control.
There are also many other issues, such as autonomous decentralized system control and
hybrid system control, that is not mentioned here but are basically important in biomimetic
control.
An autonomous decentralized system is a system in which the functional order of the entire
system is generated by cooperative interactions among its subsystems (Yuasa and Ito, 1990). Each
subsystem has the autonomy to control a part of states of the system. It is well known that biological
systems possess this autonomous decentralized characteristic. For example, animal movements are
generated by the cooperation of many motor neurons. These motor neurons control muscular fibers
to generate different movements. Rhythmic movements, such as walking, flying and swimming, are
generated by the control from the mutually coupled endogenous neural oscillators. The rhythmic
moving patterns can be changed with respect to the environmental conditions as well as various
objectives. The walking movement of a cat is a typical example. Here, the periodic motions of each
limb cooperate with each other to generate stable gait patterns. As the cat moves faster, the gait
pattern changes from ''walk'' to ''trot,'' and finally to ''gallop.'' Although the gait changes
discontinuously, the speed of the body movement varies continuously. This pattern switching
behavior can be formulated based on the bifurcation theory of nonlinear dynamic systems. In
addition, the graph structure of the interaction network between all subsystems and the time delay
of the interactions are important to determine the over system's performance such as synchroniza-
tion (Wu and Chua, 2002; Amano et al., 2004).
On the other hand, if the system contains both continuous time-driven and discrete event-driven
dynamics, it is called a hybrid system. Human brain has a high ability to mix discrete logical
thinking with dynamic body movements control simultaneously. It realizes the diversity of the
movements with respect to the environmental conditions as well as task requirements. The
biological research on basal ganglia suggests that such kind of a skillful optimal motion pattern
scheduling function is generated through the interaction between the basal ganglia and the
high level motor cortex (Hikosaka et al., 1996). Nowadays, hybrid system arises in a large
number of application areas. However, the problem of the hybrid system is inherently difficult
because of its combinatorial nature. A straightforward application of the available frameworks
faces the limitation of computational complexity and lacks the theoretical prediction of system
properties. Hybrid systems can be formulated by many kinds of models, such as piecewise
affine (PWA) system, linear complementary (LC) system, and mixed logical dynamical (MLD),
system, etc. The equivalence between each model formulation was studied. For the MLD system,
several powerful mixed integer quadric programming (MIQP) algorithms have been proposed to
solve the on-line optimization procedures (Bemporad, 1999). In robotic applications, recently,
modeling and control of dexterous multi-fingered hand operations and multi-legged dynamic
walking movements are studied from the hybrid system point of view (Yin et al., 2003). But still
few theoretic results have been achieved, and there remain many challenging problems in biomi-
metic control research. It is expected that theoretical study of hybrid system control may also lead
to a better understanding of biological motor control functions at the high level of brain motor
cortex and basal ganglia.
ACKNOWLEDGMENT
The authors are pleased to thank Professor Neville Hogan, Dr Yoseph Bar-Cohen, Dr Mikhail
Svinin, and all reviewers for their very contributive comments and help.