Do you have Adobe Reader installed on your computer? If not, that's probably why you can't open the .pdf file. You can get it at http://get.adobe.com/reader
I also watched the lecture, and I have to say it is indeed amazing. It's very newbie friendly, it answered a lot of questions I had, and it solved problems I didn't know existed but that I would have soon hit in the future (Learning by your mistakes is great, but it's even better if you don't have to make them in the first place). It's definitively a great watch!
The now-discontinued Stingray used this arrangement, with one of those funky omniwheels. I wonder if a single one of those, heavy-duty enough, would be a practical alternative.
My suggestion for an Alternate Parallax MadeUSA® Base Design it looks like the like Stingray robot and uses un aluminum heavy-duty omniwheel. It has also a lot of plastic parts from my 3d Printer.
Unfortunately my 3d printer is out of operation for 3 months due to a serious damage and I can’t finish the robot in order to test it. I’m waiting to fix the printer in order to proceed the project. I hope next month I’ll have some progress …
Sorry I am late to this thread. While I agree that a tail drag variant of the platform would have some benefits for some situations, I think it also comes with disadvantages that the current design avoids. Most notably is that of getting stuck against walls and/or in corners, the round design avoids it well.
I have a MadeUSA (upgraded from the original wood to the newer black plastic), an Eddie, and the new Arlo (still under construction slowly). I run them in indoor situations, they haven't had any issues with carpet, hardwood floors with rugs, and bumps from edging between flooring types (doorways between wood and carpeted rooms, etc). I've never had the casters cause the drive wheels to completely lose traction (I've heard people fret over that possibility). Worst case for me is that one drive wheel will lift slightly and the bot will turn a small amount because of it, but it's able to get itself over the bump and correct. The center pivot makes math on the sensors much simpler for rotations. I've been working out code to figure out walls and corners from the multiple sensors, and their uniform placement and motion when turning makes it easier.
I really like the ping/ir sensors going all the way around the edge (My MadeUSA has the 10 pings on it), with the Arlo I am going to have a mixture of Pings and IR like Eddie, except all around instead of just forward. I have plans to modify my Eddie to work without a laptop using a BeagleBone Black or something similar to talk to the Eddie control board, I'll probably add some rear facing sensors at that time. I've even contemplated using two Kinect devices, or building a rig to rotate the one. For Arlo, I will probably stick with just a Prop board (whatever Parallax has planned for it) programmed with C/C++ (propgcc/SimpleIDE, which I really like now).
Anyway, that's my input (at erco's request). If Parallax does decide to do an alternate design, I would most likely get one and try it out. I do have pretty much all of their robot kits (that I am able to get ahold of easily).
Comments
I also watched the lecture, and I have to say it is indeed amazing. It's very newbie friendly, it answered a lot of questions I had, and it solved problems I didn't know existed but that I would have soon hit in the future (Learning by your mistakes is great, but it's even better if you don't have to make them in the first place). It's definitively a great watch!
My suggestion for an Alternate Parallax MadeUSA® Base Design it looks like the like Stingray robot and uses un aluminum heavy-duty omniwheel. It has also a lot of plastic parts from my 3d Printer.
Unfortunately my 3d printer is out of operation for 3 months due to a serious damage and I can’t finish the robot in order to test it. I’m waiting to fix the printer in order to proceed the project. I hope next month I’ll have some progress …
Nikos
I have a MadeUSA (upgraded from the original wood to the newer black plastic), an Eddie, and the new Arlo (still under construction slowly). I run them in indoor situations, they haven't had any issues with carpet, hardwood floors with rugs, and bumps from edging between flooring types (doorways between wood and carpeted rooms, etc). I've never had the casters cause the drive wheels to completely lose traction (I've heard people fret over that possibility). Worst case for me is that one drive wheel will lift slightly and the bot will turn a small amount because of it, but it's able to get itself over the bump and correct. The center pivot makes math on the sensors much simpler for rotations. I've been working out code to figure out walls and corners from the multiple sensors, and their uniform placement and motion when turning makes it easier.
I really like the ping/ir sensors going all the way around the edge (My MadeUSA has the 10 pings on it), with the Arlo I am going to have a mixture of Pings and IR like Eddie, except all around instead of just forward. I have plans to modify my Eddie to work without a laptop using a BeagleBone Black or something similar to talk to the Eddie control board, I'll probably add some rear facing sensors at that time. I've even contemplated using two Kinect devices, or building a rig to rotate the one. For Arlo, I will probably stick with just a Prop board (whatever Parallax has planned for it) programmed with C/C++ (propgcc/SimpleIDE, which I really like now).
Anyway, that's my input (at erco's request). If Parallax does decide to do an alternate design, I would most likely get one and try it out. I do have pretty much all of their robot kits (that I am able to get ahold of easily).