Real-World Applications - Immunological Computation: Theory and Applications - page 185

Wantanabe et al. (1998) developed a work on decentralized adaptive-control
mechanism for the walking behavior of a six-legged robot. By using the idea of B
cells and immune networks, where a B cell is considered to be a leg in the robot and
the immune network a mechanism by which legs communicate with each other,
they proposed a system that can learn how to control the walking motion of the
robot. Each B cell records the level of antibody concentration. h e B cell with the
highest concentration is the leg that swings forward.
Singh and h ayer (2001) proposed Immunology-derived Distributed Autono-
mous Robotics Architecture (IDARA)—a self-optimizing and dynamic robot con-
trol architecture. IDARA distributes tasks so that routine actions are refi ned and
followed by more specifi c responses. h ey use a multitiered response layer equating
it to the response pattern of the innate immune system. h erefore, initial agents can
fail and pass on because they fail to successive levels of control, then a more suitable
agent is evolved. Basically, the idea is that many agents respond to one problem and
the failure of one helps improve the next agent—the idea of innate immunity before
the optimum agent can be evolved. Traditional communication in distributed sys-
tems has a signifi cantly high cost. h ey use computer simulation of a “self-healing”
mobile minefi eld having up to 7500 mines and 2750 robots and use a multitiered
response layer equating it to the response pattern of the immune system.
Lau and Ko (2007) proposed a robotic search and rescue system based on an
immune-control framework, called general suppression control framework (GSCF).
GSCF is based on the suppression mechanism of immune cells. A decentralized
system based on GSCF to assist a search and rescue a robot system to communicate
and navigate in unstructured disaster-aff ected areas is developed. h e robot system
consists of two robots and one operator console, but can be extended to a higher
number of robots. GSCF is a modular system that consists of fi ve major compo-
nents: a nity evaluator, cell diff erentiator, cell reactor, suppression modulator, and
the local environment. h us, autonomous T cells that continuously react to the
changing environment and aff ect other cells in the environment are modeled.
Lee et al. (2007) used clonal selection algorithms in controlling autonomous
underwater vehicles. Particularly, clonal selection is used to tune control param-
eters— K p , K D , and K I — of proportional integral derivative (PID) controllers. h e
proposed approach was compared with the Ziegler-Nichols (Z-N) technique with
respect to the settling time, overshoot and an a nity in submerging underwater,
and turning the yaw angle through simulation. h e immune approach is more e -
cient than the Z-N technique in submerging and turning the yaw angle.
7.5 Application in Fault Detection and Diagnosis
h e fi eld of fault diagnosis needs to accurately predict or recover from faults occur-
ring in plants, machines such as refrigeration systems, communications such as
telephone systems, and transportations such as aircrafts. Bersini and Varela (1990)