White Paper Discuss: .NET Composite Dashboard Viewer

[seanwitte]

Thread created automatically to discuss a document in the White Papers.

.NET Composite Dashboard Viewer by seanwitte

[lynca]

This program provides first teams with a great GUI camera output along with field waypoint planner. When debugging autonomous, this is invaluable. All teams should consider running this program. The only problem is the .NET requirement. If it can be ported, there would be much more input from other teams.
It is also a great application in C# which seems to be the next generation programming language (mainly in 32-bit PC computers - not embedded).

This challenges IFI to make a more configurable dashboard in which users can control the packet structure to input a variety of sensors and raw camera feeds.

[seanwitte]

Here is the first known bug:

The bits in LED BYTE 1 for the relays are reversed. It was an honest mistake, reading from top to bottom the PWM bits are fwd-rev-fwd-rev, but the relays are rev-fwd-rev-fwd. Anyways:

RLY1_Forward = RLY1_Reverse
RLY1_Reverse = RLY1_Forward
RLY2_Forward = RLY1_Reverse
RLY2_Reverse = RLY2_Forward

This mapping is in the assembly named DashboardTypeLib.dll in the \bin folder. You can download a patch for the dll below. Right-click and select Save Target As. Save the file in the \bin folder of the install directory. Replacing this DLL will correct the error.

Its not obvious I guess, but the way the application is structured you can use any of the CLR languages to write your custom views (C++, C#, VB.NET, or J#). If you run Windows Update regularly you probably already have the CLR installed.

DashboardTypeLib.dll

I also forgot to credit Dave Lavery with the cool field graphic used in the Field control in DashboardControls.dll.

[Francis-134]

So for all of us non-programmers, would this be legal to use in competition as-is, or would it require some modification?

[seanwitte]

Originally Posted by Francis-134

So for all of us non-programmers, would this be legal to use in competition as-is, or would it require some modification?

You can use it as-is, although there isn't much to gain over the IFI dashboard viewer. If you use the CMUCam you can display the current information using the user feedback bytes. The data from the cam struct needs to be copied to the bytes as follows:

User_Byte1 = cam.x;
User_Byte2 = cam.y;
User_Byte3 = cam.x1;
User_Byte4 = cam.y1;
User_Byte5 = cam.x2;
User_Byte6 = cam.y2;

[seanwitte]

Here is a sample screenshot of the dashboard viewer showing the custom controls I wrote to display the robot's position and the CMU Cam tracking data: http://www.chiefdelphi.com/forums/sh...15&postcount=6.

I also have the PWM slider and a simple LED control. The PWM and LED controls can be customized with any color combination you want.

If you would like a custom dashboard view for your 2005 robot I would be willing to write one for a small fee using any combination of these existing controls. The sprite for the robot on the field map can be changed if you provide the image. Send me an email at sean.witte@cox.net if you're interested.

[seanwitte]

If you're using this sofware you might want to download and install the version uploaded this morning. There are a few small improvements as well as a new view for automatically generating an autonomous function.

If you've made changes to the sample views you should save a copy of the source files before you uninstall and reinstall the application.

[seanwitte]

Has anyone tried using the ReplayRC view that is included with this application? It allows you to select a set of PWM outputs to record and generates a C source file to mimic motion captured while under human control. You just hit the capture button, drive around, and hit the stop button. The output is a C source file that you copy into your project and recompile. This has been done before so it wasn't a new idea.

Since the data packets are interlaced its only able to catch every third update of the PWMs. That means it probably will not match the driver's inputs exactly, but should be pretty close. The practice robot was in pieces tonight when I went to test so I'm curious whether anyone has used it and how well it worked.

[lynca]

Fantastic addition, have not yet tried this, but it is definitely on the todo list. I think simulation of a standard robot chassis in something like Player/Stage would be the next logical step toward testing the theory.

I have been working with Linux Robots and higher level navigation for a few months and Player/Stage or OpenSim (deeper) will be a great assistance to teams who are not able to touch a real robot. Mainly because it integrates off the shelf sensors and hardware into a standard package package anyone can pick up an operate. For the best introduction to the player stage world, I suggest running the Pyro LiveCD. Sorry about the off topic,

Sean, you have developed a great resource for the FIRST community, my team has been very impressed!