Notifications
Clear all
Robot Projects
7
Posts
3
Users
3
Likes
2,139
Views
Topic starter
2019-11-30 9:47 pm
Although a ROS based robot would be too complex for me personally to build for those who have such ambitions and abilities I wondered if you have read this series of so far 8 articles starting with this one.
https://www.codeproject.com/Articles/1249436/Rodney-A-long-time-coming-autonomous-robot-Part-1
Duce robot reacted
2019-12-01 9:41 am
You can still install ros and build up to it gradually. I can't really use it in my project it isn't practical for it but I still would like to learn it.it seems interesting.
Topic starter
2019-12-04 12:03 am
My interest in robotics began a long time ago before ROS existed and as such it is clear in my mind how I would go about programming such a machine from the bottom up including vision. I have spent some time looking into ROS and for me as a hobby programmer it doesn't seem to offer anything worthwhile when I consider all the time and effort that would be required to learn to use it.
My initial interest in this forum was for exchanging ideas about building a robot but after a flurry of interest it seems to have mostly died down. It has been 6 months and nothing much has happened recently with the DB1 project. Still I will plod along and see if I can get an autonomous robot base working. After 3 weeks break I am working on it again. I dislike all this need to plug and unplug the robot base from the computer during development of the Arduino code and look forward to doing the main AI on a pc (or RPi) instead.
Duce robot reacted
2019-12-04 2:57 am
@casey
I hear you that's why I put five computers on the robot a bit excessive I admit but the panda with the built in Leonardo does help quite a bit plus it has pronterface and silc3r for the printer but yes I agree that for me an autonomous robot is a bit in the future the learning curve for this was a sharp one I think I'm also about ready for a three week break.??
2019-12-04 3:30 am
My interest in robotics began a long time ago before ROS existed and as such it is clear in my mind how I would go about programming such a machine from the bottom up including vision. I have spent some time looking into ROS and for me as a hobby programmer it doesn't seem to offer anything worthwhile when I consider all the time and effort that would be required to learn to use it.
That's my view of ROS as well. Not only this but I view it as "bloatware". The reason being that ROS includes a gazillion capabilities that I'll most likely never even use. Many of which are hardware specific for hardware that I'll never buy or own. I'm even willing to bet that a lot of things I might want to do ROS wouldn't even be compatible with anyway. I would just be told to buy different hardware and I'm not interested in going there.
Finally, many of the features that ROS supposedly does are things that I am learning from the ground up anyway. Like the course I'm currently taking with Paul McWhorter on how to write my own tracking and object recognition routines using OpenCV on the Jetson Nano. ROS is most likely designed for people who just want to use existing libraries to do things that have already been done before. And that's fine for that purpose, but I'm personally more interested in developing my own ideas from scratch anyway. So for me ROS is just not all that inviting.
Besides, my biggest hold up is building the physical aspects of the robot, not the programming aspect. Programming isn't going to help much if you don't have a physical robot to program.
My initial interest in this forum was for exchanging ideas about building a robot but after a flurry of interest it seems to have mostly died down. It has been 6 months and nothing much has happened recently with the DB1 project.
Same here. I became a Dronebot Workshop Robotics Engineer, and I see the cost of that decision being drained from my bank account every month. My main interest was in the DB1 project, even though I'm not attempting to build an exact replica. None the less I've been looking forward to seeing the wiring and coding for the motor control, with acceleration techniques based on encoder feedback, etc.
So far it has been an extremely LONG WAIT with no indication that it's ever going to even happen at all. I'm about ready to just go ahead and design my own system and forget about the DB1 entirely. We're almost into Christmas and the DB1 series videos are non-existent. We were also promised schematics and code on GitHub, but I haven't seen that happen either.
Still I will plod along and see if I can get an autonomous robot base working. After 3 weeks break I am working on it again. I dislike all this need to plug and unplug the robot base from the computer during development of the Arduino code and look forward to doing the main AI on a pc (or RPi) instead.
Yes, in a way I've already written my own version of ROS (it's certainly a Robot Operating System to be sure), I wrote it in C# on Windows so I can communicate and program my robots from my notebook computer. I even have it set up so I get all the navigation data and floor plan maps displayed on the screen of my notebook even though the robot is creating them. So in a very real sense I've written my own version of ROS.
Lately I've been working on a non-robotic project to automate my woodstove boiler. But hopefully that will soon be finished and I can finally get back to concentrating on the robots again.
DroneBot Workshop Robotics Engineer
James
Topic starter
2019-12-04 9:46 pm
@casey
I hear you that's why I put five computers on the robot ...
I would be happy with ONE computer with direct i/o access the way it was with the old DOS computers and their parallel ports as all the code can be written in one computer language as a single unit. It is bad enough having to code and test the PC code and the Arduino code separately. Computer languages are like the tower of Babel, umpteen idiosyncratic ways to write the same piece of code. There are complicated ways of dealing with this such as encapsulating a set of functions in a dynamic link library providing the mother language has the capability to make use of such functions. For example I use escapi.dll (webcam functions) and K8055.dll (i/o interface functions to K8055 board) in my FreeBasic programs. I think this is a feature of ROS where some master program can use functions written in other languages. At this stage am writing code to interface to a preprogrammed Arduino board with associated i/o circuitry much like the K8055 board but more powerful.
https://forum.dronebotworkshop.com/user-robot-projects/k8055-robot/
Duce robot reacted
Topic starter
2019-12-05 7:21 am
Besides, my biggest hold up is building the physical aspects of the robot, not the programming aspect. Programming isn't going to help much if you don't have a physical robot to program.
And the cheapest robot base capable of just about any task would be a secondhand wheelchair. There is no way you could build anything as good out of parts for the same price. And how much would you be prepared to pay for a good solid robotic base? The cheapest I can find so far in Australia is $600 but I am guessing they would be cheaper in America. This would make a stable powerful robotic base that could carry wood, mow lawns, handle a powerful vacuum cleaner and so on. It can also carry as much computing power as you need. From what I have read you seem to have all the computing power and sensory circuits already paid for? You already have the toy robots to test ideas out? I am using a weak old secondhand vacuum cleaner base to test out the software and sensors but at any time I can lift out the "brain" and "spinal cord" with its attached sensors and redirect the motor commands to the wheelchairs controller instead of the L298 🙂
https://www.allaboutcircuits.com/projects/building-an-rc-robot-using-a-wheelchair-base/