Information Technology Reference
In-Depth Information
ing performance which is presented in simulation results. This paper is organized
as follows. Section 2 reviews intelligent control based on brain emotional learning.
Section 3 introduces its computation model. In Section 4 the structure of our imple-
mented controller is discussed. Section 5 is about stress generation. Simulation and
experimental results are given in Section 6. Finally, conclusion and future works are
presented in section 7.
2
Control Based on Brain Emotional Learning
Decision making process in human brains is not limited to computation and logic
which take place in upper portion of brain, also emotions that their source is in cere-
bellum and middle portion of brain are involved in decision making. Indeed solving
a decision making problem considering the complexity of the solution based on
computation and logic besides the existence of uncertainty sometimes is too intri-
cate. Accordingly, before the problem is processed in cognitive stage, it means con-
sidering total representation of external stimulus, the process of problem would be
performed with simpler representation of stimulus by emotions and relatively good
answer would be obtained so fast [4]. That is how emotional process accelerates de-
cision making procedure. There were many attempts for identification of emotional
decision making process In 80s which led to emotional system to be introduced
as an expert system [5]. In this approach images that involve the representation of
stimulus and the response of the expert system to it will be labeled with good or
bad and during decision making those images which are labeled bad will be omitted
and decision making will takes place between the rest of images [5][27]. In recent
approaches, presenting the computational model of those brain segments which are
responsible for emotional process is been considered. In methods based on com-
putational models, emotions are signals representing outer environment. In psycho-
logical researches emotion is introduced as desirability degree factor [6]. There are
the same approaches in Control. In proposed controller in [12] those items that de-
signer is sensitive to them are considered as stimulations which cause stress, and
control system must work in the way that reduces the stress. Therefore, an agent
based neuro-fuzzy controller is designed in [7] in which the parameters are adjusted
by learning. Moreover, the fuzzy controller proposed in [21] employs 3 negative
emotions: fear, anxiety, and pain for learning a mobile robot. In Brain Emotional
Learning Based Intelligent Controller (BELBIC) [15] which is remarked in this pa-
per, stress would be produced as a negative factor and the parameters of controller
which has a network structure would be adjusted based upon it. This controller is
founded based on AMYGDALA computational model [19].
3
Computation Model of Brain Emotional Learning
Amygdala is a part of brain which is responsible for emotional processes and is
connected to sensory layer, Thalamus, and Orbitofrontal portion (Fig. 1). Amygdala
and Orbitofrontal have network architecture in their computational models in which
Search WWH ::
Custom Search