![](bg4.png)
3.1 MAS Structure Description
The higher level of control system is represented by
a personal computer. In the PC the signal from
differentially GPS positioning system may be
entered, which represents the relative coordinate
system of environment – allow the precise of the
position of the race car on the circuit. At the output
is connected GPRS communication modem which
transmits commands for race car.
The algorithm of agent’s cooperation was
proposed with the control agent on a higher level.
The control agent determines the required behavior
of the whole control system as the response to the
dynamic behavior of car and to the one‘s own global
strategy in the task and knowledge about the last
situations, which are saved in the database. The
agent on a higher level controls the other agents
(Srovnal, V., Horák, B. and Bernatik, R., 2004).
The separate task is the transformation which
converts the digital data position into the object
coordinates (car position on the circuit) which are
saved in the database of the circuit. This database is
common for all agents in the control system. Each
agent sees actual the whole data and is capable of
controlling its behavior in a qualified way. The basic
characteristic of a control algorithm of a subordinate
agent is the independence on the number of decision
making agents for car on the circuit.
Agent system has a common goal, to control of
the car during race with optimizing - minimizing of
fuel consumption and control of critical speed. For
successful assertion of one’s own race strategy the
extraction and knowledge of changeable
environment and learning capabilities are very
important.
Main architecture of such hybrid agent system is
characterized via:
• Layered control. Agent is described by number
layers of abstraction and complexity.
• Layered knowledge base.
• Bottom-up activating
• Top-down execution.
Agent is connected with environment through
interface with sensors, actuators and communication
module. Control is allowed through layers at three
levels: reactive layer, layer of local planning, and
layer of cooperative planning. They are use
information from knowledge bases (“world” model,
“mental” model and “social” model), (Garani, G.
and Adam, G., 2006).
Reactive layer is responsible for adequate
reactions at the stimulations from environment that
require immediate reaction and execution of called
procedures from local planning layer. Fundamental
characterization of such layer is:
• Use of effective algorithm of compare with
patterns of behavior. Serve to pick-out of the
actual situations.
• Situation description for timely actual reactions
at received stimulus.
• Hard-wired links. Recognized situations are fix-
connected with targets for reactive behavior.
Immediate execution of program actions.
• Solution request of situations not–corresponding
with couples situation-action are transmitted in
local planning layer.
• Execution liability is coming from local
planning layer activate procedures of reactive
layer patterns of behavior.
Some situations can be not solved by execution
of template action like an answer to stimulation from
environment only, but they require certain level of
deliberation. A function of plans creation for solving
of the targets performs the layer of local planning.
Local planning layer have such fundamental data
structures:
• Targets – state sets. Sets are characterized by
attributes that are fulfilled at reaching targets.
• Planning – planning from second principles.
Sets of plans are defined before in data structure
– plans library. Mapping of target sets to plans
library is existed. For each target is possible to
assign the plan for its reaching.
• Plans library – contain the plans for reaching of
agent targets.
• Scheduling – secure the timely limited plans
stratification. Be created the plan schedules like
the step sequences, to execute.
3.2 Cooperative Planning Layer
A basic control cycle of cooperative planning layer
is creation, interpretation, decision making and
execution of local plans.
In first phase the reports from nearby layers are
processed. Reactive layer sends requests to solve
new task or status of executed behavior templates.
Schedules of active plans are actualized.
Subsequently the status from reactive layer executed
procedures is checked.
In case of successful procedures finalization the
plan is erased from accumulator. Reports from
highest layer are related to creation or cancellation
of commitment for the plan execution at local base
or plan evaluation. In case of plan execution request
or his cancellation the accumulator of active plans is
actualized.
The plan availability is a result of difference of
his relative value for the agent and his costs for
execution. The plan value is derived from target
HYDROGEN POWERED CAR CONTROL SYSTEM
65