Differences

This shows you the differences between two versions of the page.

--- eeros_architecture:control_system:signals [2015/10/14 13:44] – graf
+++ eeros_architecture:control_system:signals [2021/03/03 08:29] (current) – [Reading from Unconnected Inputs] ursgraf
@@ Line 3: / Line 3: @@
 A signal represents a succession of varying physical quantities over time as used in signal processing and control. It has certain properties, such as a value and a timestamp, indicating the exact time corresponding to the value of the represented physical quantity. This is usually when the value was captured or generated.
-{{ signaluml.png?150 | Signal type with its properties}}
+===== Properties =====
+Each element of a signal is characterized by a value, a high precision timestamp, a name and additional properties, depending on the type of signal, which will provide semantic information. Some examples are dimension (e.g. length, mass, current, etc.), SI-unit (e.g. m, m/s, kg, Nm, A, etc.), reference coordinate system or activation state (enabled/disabled). Other properties could include information about the quality of the signal like noise or quantification which could be used for Kalman filtering.
+[{{ signaluml.png?150 | //Signal type with its properties//}}]
 It's important to note that a signal can carry a single value as well as a vector or matrix of values. For this purpose the field //value// is declared as of type T.
+Signals are used to connect different blocks in the control system and share information between the control system, the sequencer and the safety system.
-The example below shows a gain block being defined, where its input and output are set to a three dimensional vector. Its gain is therefore a 3x3-matrix.
+Supported types for signals are
-<code c>
+  * Arithmetic (int, double ...)
-	Gain<Vector3, Matrix<3,3>> gain(c);
+  * Logic (bool)
-</code>
+  * Vector (Vector2, Vector3, Vector4)
+  * Matrix
+Each signal can be assigned a name and a unit. All dimensions of a signal share the same name and unit.\\
+As an example we look at a block which does summation. Two signals, each of dimension 3, will be added together.
+<code cpp>
+  Sum<2,Vector3> sum;
+  sum.setName("adds feedback");
+  sum.getOut().getSignal().setName("control signal");
+</code>
+This is shown in the following diagram:
+[{{ sumblock.png?300 | //Summation block with threedimensional signals//}}]
-Each element of a signal is characterized by a value, a high precision timestamp, a name and additional properties, depending on the type of signal, which will prrovide semantic information. Some examples are dimension (e.g. length, mass, current, etc.), SI-unit (e.g. m, m/s, kg, Nm, A, etc.), reference coordinate system or activation state (enabled/disabled). Other properties could include information about the quality of the signal like noise or quantification which could be used for Kalman filtering.
+The functions //getIn()// and //getOut()// return an input or output. To access the signals on those in- or outputs, use //getSignal()//.
+===== Timestamp =====
+Every signal carries a timestamp. The timestamp of a signal is usually set in input blocks. Block which read from an external source and deliver this value on a output signal will set its timestamp to the time of the acquisition of the value. \\
+Most blocks will not alter this timestamp. However, blocks which depend on the sampling time such as integrators or differentiators will change the timestamp according to their inner algorithm.
+===== Reading from Unconnected Inputs =====
+If you try to read from an input, which is not connected to an output, a //NotConnectedFault// is thrown. Please make sure that such faults are handled correctly. For an example refer to [[getting_started:tutorials:controlsystem|]].
-Signals are used to connect different blocks in the control system and share information between the control system, the sequencer and the safety system.
-In order to access the individual dimensions of a certain signal, use functions like
-<code c>
-	sum.getOut().getValue();  // returns dimension 0 of output
-	sum.getOut().getValue(0);  // returns dimension 0 of output
-	sum.getOut().getValue(2);  // returns dimension 2 of output
-	sum.getIn(1).getValue(2);  // returns dimension 2 of input 1
-</code>