A Telemetry Browser built with Java Components

Size: px

Start display at page:

Download "A Telemetry Browser built with Java Components"

Brendan Patrick
5 years ago
Views:

A Telemetry Browser built with Java Components Erwann POUPART, CNES, DTS/MID/VM/TD, BPI 1501 18 Avenue Edouard-Belin 31401 TOULOUSE CEDEX 4 Tel : +33 (0)5 61 28 13 49 Fax : +33 (0)5 61 20 19 20

Introduction The aim of this paper is to present how off-the-shelf components are assembled together to realize a TelemetryBrowser able to display at the same time multiple parameters coming from

1 A Telemetry Browser built with Java Components Erwann POUPART, CNES, DTS/MID/VM/TD, BPI Avenue Edouard-Belin TOULOUSE CEDEX 4 Tel : +33 (0) Fax : +33 (0) erwann.poupart@cnes.fr 1. Introduction The aim of this paper is to present how off-the-shelf components are assembled together to realize a TelemetryBrowser able to display at the same time multiple parameters coming from different telemetry formats. This TelemetryBrowser, which is also a CORBA server, can be interconnected with another through Internet tolerating line cut-offs. COTS used are : JCLASS Java components from SITRAKA i for telemetry visualization, MACS ii using a MAPINFO iii runtime for geographic localization, JavaORB for distribution over Internet and connection with other Telemetry Browsers (applications or applets). It was used to survey CNES balloon scientific campaigns, such as the ARCHEOPS campaign in December 2001 at Kiruna (Sweden). This TelemetryBrowser, working as a CORBA server, was connected to a direct link with the ARCHEOPS balloon. Multiple other TelemetryBrowsers, working as CORBA clients, were connected to the first one through Internet. One of them, using a CORBA applet, was located at CNES Toulouse (France), and worked for three hours with a telemetry rate of one frame per second (1Kbyte/sec). 1

2. Software Architecture A TelemetryReader which is a parameter producer communicates with Dynamic Tables (using JCLASS LiveTable component), Dynamic Charts (using JCLASS Chart component) and Dynamic

2 2. Software Architecture A TelemetryReader which is a parameter producer communicates with Dynamic Tables (using JCLASS LiveTable component), Dynamic Charts (using JCLASS Chart component) and Dynamic cartographic Clients (using MAPINFO) which are parameter consumers. Communication between those components operates through listener lists (see scheme below). 2.1 Telemetry is decomposed into TelemetryParameters : A TelemetryReader object reads bytes from a data source which can be a serial port, a file or network (FTP, HTTP or IIOP). Then, it transmits bytes to a TelemetryServer object where telemetry is decomposed into TelemetryParameters which are defined as described below : Name, Units, timeparametername, parameterlisteners, boolean istimeparameter(), void addparameterlistener(parameterlistener l), void removeparameterlistener(parameterlistener l), String decomute(byte[], position, occurence), void setvalue(parametervalue value). TelemetryParameter name must be unique, it's the concatenation of telemetry identifier and parameter effective name. 2

3 For each parameter, there is only one date, which is also a TelemetryParameter. Parameters can be under or over commuted. Occurrence information is a counter that increases for each frame and for each over commuted parameter if there is one in the frame. For example, if for frame number i you have parameter P0 overcommuted 4 times, a value will be sent at occurrence i, i+1, i+2 and i+3. And the next frame will start at occurrence i+4. Time parameter which contains date information has a value for each occurrence. It's possible to have multiple time parameters in one frame. In this case, parameters don't have all the same time parameter. Parameters are also structured into groups, so that clients can ask to the TelemetryServer object for those groups to display only some of them. For example, in Archeops telemetry, there are two groups of parameters for each frame, one "Localization" group with GPS date, and one "Automating" group with another date for the separation automate. Each parameter is able to "decommute itself", receiving a byte array at current position, it is able to calculate its value and optionally to qualify it (no qualification, nominal, abnormal, critical). Those operations are made in decomute(byte[], position, occurence) function. It creates a TelemetryParameter value, which is a generic ParameterValue composed of : occurence, value, nominalvalue. value field is a generic one using type Any of Corba. The advantage of this type is that clients can get it (float, int, data structures, etc.) dynamically during execution time, and process it as convenient. nominalvalue field is the qualified value of the precedent one. There is a codification (-1 : no qualification, 0 : normal, 1 : abnormal and 2 : critical). Then, this ParameterValue, created by the TelemetryParameter during decommutation operation, is sent to each parameter listener using setvalue(parametervalue value) function. It uses JavaBound properties and firepropertychange(string parametername, null, ParameterValue value) function which calls propertychange(propertychangeevent e) on each listener (parameter consumer) to transmit this new parameter value. 2.2 Connection between parameter producers and parameter consumers : This connection is made very easily with listener lists updated using addpropertychangelistener(propertychangelistener l) and removepropertychangelistener(propertychangelistener l) functions. Parameter consumers just have to implement propertychange(propertychangeevent e) function and process this information as convenient. 3

2.2.1 Connection initialization between parameter producers and parameter consumers : This software can process different telemetry formats from the same input source.

4 2.2.1 Connection initialization between parameter producers and parameter consumers : This software can process different telemetry formats from the same input source. Formats are identified during execution, and if a new one appears then a new TelemetryFormat event is generated, sent to TelemetryBrowser application which starts new parameter consumers (see scheme below). Parameter consumers, like dynamic tables, start a dialog with TelemetryServer object to get parameters information and be ready to receive parameter values. 4

Finally, JavaORB is the ORB used to interconnect multiple TelemetryBrowsers (application or applets) through Internet. Using ORB included in the Java Plugin reduces significantly applet's size.

5 Finally, JavaORB is the ORB used to interconnect multiple TelemetryBrowsers (application or applets) through Internet. Using ORB included in the Java Plugin reduces significantly applet's size. So, each TelemetryBrowser is a Corba object able to redistribute telemetry raw data. To realize this operation, a TelemetryDaemon object was developed to manage multiple connections and line cut-off. It buffers telemetry raw data coming from TelemetryReader for each remote TelemetryBrowser connected through addpullclient(string clientname) function. And each remote TelemetryBrowser asks at its own speed for the last telemetry raw data using getlastrawlines(byte[] rawdata, int len, string clientname) function. 5

6 3. Concrete application reusing JCLASS LiveTable, JCLASS Chart components and MAPINFO application. JCLASS Table and JCLASS Chart components from SITRAKA had to be specialized so that they can add, remove and modify dynamically new Telemetry Parameters into their data models. 3.1 JCLASS LiveTable short description : JClass LiveTable is a high quality grid/table JavaBean for data-driven display. Highly Configurable JClass LiveTable provides extensive properties and methods with a wide variety of customization options. It can easily be configured to act as a spreadsheet, data-entry form, multicolumn list or data grid. With JClass LiveTable, the data you display can come from a database, a real-time system, a comma-delimited file, or directly from the Web page that launched the applet. Interactive features such as cell resizing, on-the-fly sorting and searching are easy to use. And since the properties of JClass LiveTable can be changed at run-time you get the ultimate in flexibility and reuse. Superior, Scalable Performance JClass LiveTable is optimized to provide superior performance as you scale your operating conditions. It uses a unique queued scrolling system that allows real-time scrolling with smooth repainting. JClass LiveTable supports huge virtual tables so you can choose to deliver only the information the user needs. Fast cell-level redraw gives you maximum efficiency, and with JClass LiveTable you have the ability to turn double-buffering on or off as needed. Any Data from Anywhere With JClass LiveTable, your tables can be populated with data from a file, a URL via HTTP, an XML source, a relational database via JDBC, supplied programmatically or entered by an end user. Specify cell values as regular text strings or as any arbitrary Java object. It's even easy to use real-time data with a custom data source that notifies the table whenever new data is available. 3.2 Connection with JCLASS LiveTable component : JCVectorDataSource which is the main data source of JCTable was used. It stores label and cell values in a Vector of Vectors and provides a range of methods for getting and setting the cell values. Two main columns were created, one for parameter names, and one for parameter values. Adding a parameter, means adding a new row to the data model. Modifying a parameter means modifying the cell at the second column and at the right row. Background cell color is modified according to nominalvalue (no qualification : white, nominal : green, abnormal : orange, critical : red). 6

7 Removing a parameter, means removing a row from the data model. Afterwhat JCTable class was used very easily. A thread was used to make independence between user interaction on the table (parameters can be selected to plot them in a JCLASS Chart, see below) and data model modification made by parameter producers. 7

3.3 JCLASS Chart short description : Powerful Components and Easy-to-Use Beans With JClass Chart, you can develop sophisticated charting applications with components, or develop fast, effective chart

8 3.3 JCLASS Chart short description : Powerful Components and Easy-to-Use Beans With JClass Chart, you can develop sophisticated charting applications with components, or develop fast, effective chart solutions with the MultiChart and SimpleChart JavaBeans. Business and Scientific Charts Take your pick of multiple chart types with JClass Chart, including Polar, Radar, Area Radar, Bar, Stacking Bar, Pie, Hi/Lo, Hi/Lo Open/Close, Area & Stacking Area, Scatter Plot, Plot, Line, Candle. Flexible Data Sources JClass Chart supports a vast array of data sources including files, sockets, real-time feeds and even shares data with JClass LiveTable. With its MVC (Model-View-Controller) architecture, JClass Chart makes it easy to bring data in from any source. Use the pre-built data sources included with JClass Chart to read in data from memory, a file, a URL, an XML source, or any arbitrary InputStream. And if you want total flexibility, you can even build custom data sources to retrieve data from remote servers or real-time information streams. Fully Interactive JClass Chart is ideal for decision support applications because it has built-in support for common chart interactions, like panning, zooming, rotating and picking. It's easy to enable or disable these actions for your users and you have the flexibility to bind to any mouse button and modifier key combination. With JClass Chart's map and pick functions you can enable applications to provide feedback when the user clicks the mouse anywhere on a graph. Transpose, drill down, edit and scale with any chart type to give your users the ultimate in interactive capabilities. Easily Customizable JClass Chart gives you great looking defaults for your charts. Plus, it's easy to customize the appearance of your chart axes, headers, footers, legends and data labels. Advanced features like reversed logarithmic and multiple Y-axes are also supported and you have lots of formatting options, including scientific notation, percents, and currencies. Choose between complete control over text and label positioning or letting JClass Chart's dynamic defaults take over. The built-in JavaBean customizer allows JClass Chart objects to be manipulated at design time or run-time and lets you modify and save charts using the chart applet HTML format. 8

9 3.4 Connection with JCLASS Chart component : Data source is specified in terms of data series (data historization). Chart asks for the x or y data for a particular series, and the data source implementing this class is expected to return an array of double values (double[]). Adding a parameter, means adding a new Yseries to the data source and if necessary a new Xseries for the corresponding time parameter. Parameters can come from different telemetry sources, so it's possible to plot in the same chart altitude parameter coming from two or more balloons for example. Modifying a parameter value means modifying Yseries at occurrence index with the new value. If it's a time parameter, then it's the Xseries that is modified. Removing a parameter, means removing a series from the data source. Then JCChart class was used very easily. A thread was used to make independence between user interaction on the chart controls and data model modification made by parameter producers. 3.5 Connection with MAPINFO application : A position consumer (geographic localization Object) was developed to interface with MAPINFO application. It listens to localization parameters (time, latitude, longitude, and altitude) coming from each telemetry, and reformat this information into MAPINFO format for geographic localization tracking. 9

10 4. Conclusion This flexible architecture allows to add or replace easily new parameter consumers. They don't have to know about telemetry formats. At the other end, it's also easy to add or modify a telemetry format on the TelemetryServer. This shows that it is possible to build an application reusing COTS even if it is necessary to develop some code to make them communicate. For release number 3.0 in preparation, several new features will be added : A new component called FLAVOR iv, developed at Columbia University, will be added on the TelemetryServer to generate automatically decommutation classes from the binary description of any telemetry format. It will be also possible to connect clients to multiple telemetry sources. TelemetryReader and TelemetryServer will become independent Corba objects. With this architecture, it will be possible to spread over the network raw data and parameter producers and consumers wherever you want according to needs and constraints. Security aspects will be incorporated to go through firewalls. i SITRAKA is a provider of reusable components ( ii MACS is a CLS/ARGOS product using a MAPINFO runtime ( iii MAPINFO runtime is a MAPINFO corporation product ( iv Flavor is a language that allows the description of coded audio, image and other data bitstreams, and the automatic generation of C++ and Java code to read/write them ( and 10

Geographical mapping of data

BioNumerics Tutorial: Geographical mapping of data 1 Aim In many research projects, especially epidemiological, biological data is closely linked to geographical data. Geographical information provided