automata (finite state machine)
This block gives the possibility to construct hybrid automata, i.e., a
hybrid system whose discrete part is defined via modes and transitions
between modes, and the continuous part is defined via DAE
(differential algebraic equations).
The automaton block provides a switching mechanism between subsystems
corresponding to control modes of an automaton. Subsystems are
constructed in such a way that they have the state vector as input (
coming from the automaton block) and compute the flow and jump
functions (zero-crossing) and pass them back to the automaton
block. The state variables are defined in the automaton block and the
subsystems are static functions.
Suppose that a hybrid automaton consists of M control
modes. The continuous-time dynamics in mode i is defined with DAE :
where i is in [1...M] and the dimension
of x is N (N >= 0) for any i in [1...M].
Suppose that in control mode i, there are Z_i jump conditions
indicating jumps toward other modes. The jump conditions are defined by functions :
where j is in [1...Z_i].
When a jump function changes sign and becomes positive, a mode
transition will happen. When k-th transition function becomes
positive, a transition to mode k happens and state vector x
is reset to :
For i.e.,
In order to develop an automaton containing a mode with multiple
reset functions, the value of the current and previous active modes
should be used. These values are available at the first output port
of the block.
The automaton block has the following input/output ports.
- Output 1: The first output port is a vector of size two
consisting of the current and the previous active control modes,
i.e.,
- Output 2: The second output port is a vector of size 2N providing the state vector and its first time derivative, i.e.,
.
- Inputs: The automaton block has M vector input ports
corresponding to M modes or subsystems of the automaton. Each
input defines the dynamic behavior in the control each mode as well
as the reset functions and the transition functions. The input port
i which is the output of the i-th subsystem is a vector of
size 2N+Z_i. Each input is composed of the following vector.
- The first N elements of the Input_i are the continuous-time
dynamics. The dynamics of the system in the control mode i is
described by a smooth index-1 DAE, i.e.,
- The next N elements of Input_i are the values used to
reset the continuous-time state when a transition to
control mode i is activated.
- The next Z_i elements of Input_i are the jump or
zero-crossing functions. If the j-th zero-crossing function of
mode i crosses zero with negative to positive direction, a
transition to j-th destination mode happens.
- Event Output: This is an event output port, which is
activated whenever a mode transition happens. This event is useful
when an event is needed to activate or initialize a part of the
subsystem not included in the internal dynamics of the automaton
block.
In the interface window, the number of control modes, the initial
control mode and the initial value of continuous-time state at the
beginning of the simulation should be given.
Find more documentation and demos about the Automaton block oat
www.scicos.org. Interested users are referred to the paper "Modeling
Hybrid Automata in Scicos", Masoud Najafi, Ramine Nikoukhah, 2007 IEEE
Multi-conference on Systems and Control, Singapore.
- Number of (finite-state) Modes
Number of modes in the automation.
Properties : Type 'vec' of size [1,1].
- Initial Mode
Initial active mode at the beginning of the simulation.
Properties : Type 'vec' of size [1,1].
- Number of continuous-time states
Number of continuous-time states at modes. Note that the number of
continuous-time states is the same in all modes.
Properties : Type 'vec' of size [-1,1].
- Continuous-time states initial values
Initial value of continuous-time states at the beginning of the
simulation.
Properties : Type 'vec' of size [-1,1].
- Xproperties of continuous-time states in each Mode
In this field the state types in mode are given. A state in an index 1
DAE can be either differential state or algebraic state. Xproperties vector is coded in an M*N matrix, where M is the
number of modes and N is the number of states. This matrices
indicates whether a continuous-time state is algebraic or
differential in each control mode. If in the i-th mode, j-th state
is differential, the (i,j)-th element of the Xproperty matrix
should set to "+1", otherwise it should set to "-1". Xproperty can
be given as a 1*N vector if type of states remain the
same in all modes.
Properties : Type 'mat' of size [-1,-1].
- Jump from Mode 1:[..;M_final(Guard=In(1).i);..]
The Jump fields express the mode transition information. Suppose
that all control modes are labeled from 1 to M. Then, in the field
corresponding to control mode i, destination modes of mode i
are defined in a vector. j-th element of this vector gives the
destination mode when j-th jump function :
becomes positive. For example, if in the field of the mode 2, the
user defines [1;3;4], it means that in mode 2, there are three
active jump functions. When, for example, the third jump function
becomes positive, a mode transition to mode 4 will be activated.
Properties : Type 'vec' of size [-1,1].
- always active: yes
- direct-feedthrough: no
- zero-crossing: yes
- mode: no
- regular inputs:
- port 1 : size [3,1] / type 1
- port 2 : size [3,1] / type 1
- regular outputs:
- port 1 : size [2,1] / type 1
- port 2 : size [2,1] / type 1
- number/sizes of activation inputs: 0
- number/sizes of activation outputs: 1
- continuous-time state: yes
- discrete-time state: no
- object discrete-time state: no
- name of computational function: automat
Masoud Najafi INRIA