Well, the project is finally done, and we've turned everything in. Look at the final report in the Documents section for more detail on what we did and didn't get done. Here's the short form:

The Gaucho is capable of remote control through direct commands or preprogrammed sequences. The interface used is a web page that can be accessed from any browser. The Differential GPS is usable, but isn't connected to the EBox since there is only one avaliable serial port that is being used for the Gaucho. All of the pieces exist, they just need to be put together yet.

We had our presentation today and it seemed to go pretty well. We did some work last night and were able to take the Gaucho outside to test it for a while. At one point the Gaucho quit responding to us, but that may have been caused by the network bridge losing connectivity. For the most part, the Gaucho's systems appear to be very stable for the purposes of remote control.

We also added code to our web server to allow it to execute a pre-programmed sequence. The sequence we added consists of a 5 second forward movement, a 3 second right turn, and then 5 seconds going forward again. We tested the Gaucho doing this outside, and it worked very well. The 3 second right turn appears to correspond to approximately a 100-110 degree turn. We also used the remote control system to drive the Gaucho down the sidewalk, and were able to drive it the length of Jobst Hall to in front of Bradley Hall and start going back before we had problems, which we think were caused by the wireless bridge losing it's connection. We had to cut power to the system to get the Gaucho to stop moving, so we weren't able to check the status of the system when we got it back to the lab.

After we got back to the lab we ran into a slight problem, we had unplugged parts of the system so that we could take more pictures, and when we went to plug back in power to the Ebox and wireless bridge, one of the batteries got plugged into the power connector for the GPS, which caused a -12VDC signal to be put across the 24VDC-to-5VDC convertor that we were powering the Ebox and wireless bridge with. The end result of this was smoke started to come out of the convertor. We unplugged things quickly, but unfortunately the convertor is destroyed. We did try to use it again to see if it maybe wasn't permanently damaged by the mistake, but an attempt at powering it correctly resulted in pink smoke coming out of the convertor.

Now that the serial port is working, we have been able to build a remote control system for the Gaucho that uses a web page for user interaction. The system has turned out to be a little sluggish, but it appears that this is mostly caused by the web browser being slow to make the requests that trigger different actions. While a web page may be perfectly functional for controlling navigation later on, it probably shouldn't be used for remote control too much.

We are still having issues with the serial dongle, it just doesn't want to be recognized by the Ebox. At this point, it is so close to the end of the semester (presentations start in two days) that we pretty much can't get it to work in time. There is a newer version of the Ebox that has two serial ports on it (the Ebox-2300), which would enable us to get the GPS and Gaucho control working at the same time. However, it is way too late for us to try getting a new Ebox in time to make a difference. Now we need to work on our presentation and final report stuff.

We tried to write a program for the Ebox that takes data from a telnet connection and passes it through to the Gaucho. The was meant to be an initial test and remote control system, but it didn't work - we kept getting an exception whenever we tried to open the serial port. We finally connected a null modem cable from the Ebox to a laptop so that we could see what was being sent out through the serial port. We then found out that the Ebox by default outputs debugging information over the serial port. After trying several things, we eventually found how to disable this debugging information, but the serial port was still erroring when we tried to open it. We discovered that this was caused by a known issue with the version of .NET Compact Framework that platform builder was using to build the image. There has been an update to platform builder release that fixes this issue. After installing the update the serial port started working for us.

We also have been having issues with the serial dongle that we borrowed to give us a second serial port for the GPS receiver. We believe that we included support for USB-serial adapters in our Windows CE image, but when we plug the adapter in, we get a dialog box asking us for the name of a driver to use with the dongle. We need to investigate this further to see if we forgot to include something in our image or if this is because the dongle we have is not compatible with the standard USB-serial driver that Windows CE includes.

So, it finally appears that we are at a point where we can actually build a functional remote-control system. Hopefully this time we will actually be able to get that part of the system working, since so far whenever we have tried to do this, we have encountered another problem that has prevented this from working. By now we ought to have encountered every problem that is possible, so stuff should start to behave for once.

While the Ebox was booting off of a hard drive, it was still interacting with the Visual Studio debugger. This meant that the Ebox would not boot if the debugger was not avaliable. So, we had to disable that in order to get it to boot without such services avaliable on the network. The image is set to use the hive registry on a RAM/ROM file system. It is now booting correctly and we can finally begin development of the control server. We have also spent part of the last two weeks going through and modifying any components of the Gaucho that we have found needing changes made - mostly the power system - and deciding how to mount the wireless bridge, GPS receiver, and Ebox on the Gaucho.

From previous projects and classes we have a .NET library that can listen to GPS and a fairly basic web server that is written in C#.NET. We will hopefully be able to adapt the existing code to be used for this system without too much difficulty. We will still have to write code to output the commands to the Gaucho however. We showed the Gaucho at the Expo Friday, and people seemed to like it.

The Ebox finally will boot off of a hard drive using a CE 6.0 image. The Gaucho is also usable (see previous post), although it could probably use some tweaking yet. We are finally at the point where we can develop the actual control system for the Autonomous Vehicle portion of our project. The control system is going to be written in .NET 2.0, and the primary interface from the user side of it will be a web site.

Yeah, so we finally get to do the real part of our project that we meant to start sometime before winter break.

We got the Gaucho Control System working about a week ago. We can now drive the Gaucho around through serial commands. We had some fun driving it around the department and taking it up the elevator to show to some of the professors upstairs. In order to fix the Gaucho, we ended up having to completely rewrite the code that had been done last year since we were unable to get it to function correctly. The hardware was well built and in functional condition however, which made our job somewhat easier.

We are now at the point of trying to get the Ebox working correctly so that we may use it as the control system for the Gaucho and start driving it around through a web-based interface. However, the Eboxes are refusing to boot off of a hard drive for now, which makes it hard to develop software for them.

Currently, we have implemented a DGPS system, although the range on it probably needs improvement. We have examined the Gaucho and added some components to it in order to better support the parts of our system that will be mounted on the Gaucho. The most obvious part of this is a pole on the back of the Gaucho to support the antenna for the GPS receiver. We are now trying to re-implement the Gaucho control system that had been completed as part of a previous project. However, we are having difficulties getting the previous project's code to compile correctly, and as of now, are unable to control the Gaucho.

If the Gaucho control system turns out to be unusable, we will need to redesign our project in some fashion to perform a task that we can complete.