With the change from the Raspberry Pi stack to the much more powerful Jetson TX1 some changes needed to be made to the overall architecture of Nomad's electronics. Some of it remains the same, such as the sensor cluster passing through the Arduino Mega. Some of the changes are enhancements made necessary by the new processor. The diagram illustrates what the new architecture looks like.
Note: The Edison board has been removed from Nomad and two Raspberry Pi B+s were installed. In time these will be swapped out for Pi 2s, but I had B+s on hand, so that's what went in. The decision to go with the Pis over the Edison had to do with support, or the lack thereof, stemming from issues getting the joystick device to install properly. When attempting to get Nomad to run as BARB did I discovered there joystick was not available on the Edison. After nearly 2 weeks of working with it, scouring the forums, etc., I could find no solution.
I admit the title may seem a little harsh. I tried to find a way to soften it a little bit, but this is the most direct way of saying it. And, frankly, the Edison did fail the Nomad project, not because it's a bad product, it's not. It failed because it's simply not ready for the hobby market. It's not ready to be picked up by your average user looking to do something amazing that pushes the boundary of what hobby electronics, or hooby robotics is. The Edison is great for engineers looking to do something new. But not your average maker. Here's the story...
When the Nomad project got its start it was going to use a stack of electronics including an Intel Edison processor and an Arduino Mega with a sensor shield. For testing and development an XBox Kinect sensor was going to be used. We knew the Kinect wouldn't work out doors due to the intensity of sunlight would wash out the infrared dot cloud. But for initial experimentation it would be a great place to start.
While writing the article that was the product review of the Nomad chassis kit I determined I needed to extrapolate how the chassis would be used. At that point I had been reading about different types of competitions for robotics and was particularly interested in the Robomagellan contests. The Nomad chassis seemed, to me, to be purpose built for this type of application. So the next question was, given the materials I had on the bench at the time, what would an autonomously navigating version of Nomad look like? This is the result.
In December, 2014 I took the first iteration of Nomad to my parents' place in Missouri over the Christmas holiday. While I was up there I was able to complete the build and run it through some tests. These images are from those tests.