1 behavioral systemsThe role of self-organization for the synthesis and the understanding of
1.1 Introduction
The basic idea behind evolutionary robotics goes as follows (see figure 1.1). An initial
population of different artificial chromosomes, each encoding the control system (and
sometimes the morphology) of a robot, are randomly created and put in the environment.
Each robot (physical or simulated) is then let free to act (move, look around, manipulate)
according to a genetically specified controller while its performance on various tasks is
automatically evaluated. The fittest robots are allowed to reproduce (sexually or asexually)
by generating copies of their genotypes with the addition of changes introduced by some
genetic operators (e.g., mutations, crossover, duplication). This process is repeated for a
number of generations until an individual is born which satisfies the performance criterion
(fitness function) set by the experimenter.
Evolutionary robotics shares many of characteristics with other approaches, such as
behavior-based robotics, robot learning, and artificial life.
Behavior-Based Robotics
The behavior-based robotics approach is based upon the idea of providing the robot with
a collection of simple basic behaviors. The global behavior of the robot emerges through
the interaction between those basic behaviors and the environment in which the robot finds
itself (Brooks 1986, 1999; Arkin 1998). Basic behaviors are implemented in separate sub-
parts of the control system and a coordination mechanism is responsible for determining
the relative strength of each behavior in a particular moment. Coordination may be accom-
plished by means of competitive or cooperative methods. In competitive methods only one
behavior affects the motor output of the robot in a particular moment (see, for example, the
subsumption based method proposed by Brooks 1986). In cooperative methods different
behaviors may contribute to a single motor action although with different strength (see, for
example, the method based on behavioral fusion via vector summation [Arkin 1989]).
In this approach, as in evolutionary robotics, the environment plays a central role by
determining the role of each basic behavior at any given time. Moreover, these systems are
usually designed through a trial and error process in which the designer modifies the current
behaviors and progressively increase the number of basic behaviors while testing the
resulting global behavior in the environment. However, evolutionary robotics, by relying on
an automatic evaluation process, usually makes a larger use of the trial and error process
described above. Moreover, while in the behavior-based approach the breakdown of the
desired behavior into simpler basic behaviors is accomplished intuitively by the designer,
1 behavioral systemsThe role of self-organization for the synthesis and the understanding of
1.1 Introduction
The basic idea behind evolutionary robotics goes as follows (see figure 1.1). An initial
population of different artificial chromosomes, each encoding the control system (and
sometimes the morphology) of a robot, are randomly created and put in the environment.
Each robot (physical or simulated) is then let free to act (move, look around, manipulate)
according to a genetically specified controller while its performance on various tasks is
automatically evaluated. The fittest robots are allowed to reproduce (sexually or asexually)
by generating copies of their genotypes with the addition of changes introduced by some
genetic operators (e.g., mutations, crossover, duplication). This process is repeated for a
number of generations until an individual is born which satisfies the performance criterion
(fitness function) set by the experimenter.
Evolutionary robotics shares many of characteristics with other approaches, such as
behavior-based robotics, robot learning, and artificial life.
Behavior-Based Robotics
The behavior-based robotics approach is based upon the idea of providing the robot with
a collection of simple basic behaviors. The global behavior of the robot emerges through
the interaction between those basic behaviors and the environment in which the robot finds
itself (Brooks 1986, 1999; Arkin 1998). Basic behaviors are implemented in separate sub-
parts of the control system and a coordination mechanism is responsible for determining
the relative strength of each behavior in a particular moment. Coordination may be accom-
plished by means of competitive or cooperative methods. In competitive methods only one
behavior affects the motor output of the robot in a particular moment (see, for example, the
subsumption based method proposed by Brooks 1986). In cooperative methods different
behaviors may contribute to a single motor action although with different strength (see, for
example, the method based on behavioral fusion via vector summation [Arkin 1989]).
In this approach, as in evolutionary robotics, the environment plays a central role by
determining the role of each basic behavior at any given time. Moreover, these systems are
usually designed through a trial and error process in which the designer modifies the current
behaviors and progressively increase the number of basic behaviors while testing the
resulting global behavior in the environment. However, evolutionary robotics, by relying on
an automatic evaluation process, usually makes a larger use of the trial and error process
described above. Moreover, while in the behavior-based approach the breakdown of the
desired behavior into simpler basic behaviors is accomplished intuitively by the designer,
