Thinking to start using ROS....?
Yeah well offcourse the more knowledge of on of those programming languages the easier it will beprobably. But by doing it one will learn. So I have been using the simmulation capabilities of ROS the last couple of weeks and it is pretty fun to use ROS IMHO. It takes some time to get used to it but it is possible even with some basic knowledge of C++ or Python as I have.
That is good to read.
I don't know if there is any other active members here also using ROS?
We all have to run our own race and if you are having fun that is all that counts.
Keep in touch with your progress it might inspire others to give it a go.
Yes I agree everybody has to do its own way and as long it's fun it is good.
I was just thinking.... I'm trying to follow a course about ROS and maybe for the sake of fun you'd like to see what I have done so far with ROS. Sorry about my terrible english in the video but here I try to give the robot new coordinates and it should move to those.
This was a fun assignment to do and I learned a lot from it. I hope the video is not to lng. I had to make a video because it was a assignment and I share the same video here.
Your English was easy for me to understand.
Once you think you have become proficient at using ROS and the simulator it would probably worthwhile getting yourself a real physical robot to see any practical details required to get ROS working with real hardware.
James Bruton has made ROS enabled robots.
One of his last projects was a really useful robot.
Haven't seen it do anything really useful yet 🙂
And also he made a simpler ROS enabled robot.
Well I have ordered a NVIDA Jetson with cam and WiFi card. So what I wan to accomplish is being able to control the bot with my joystick (with the simulation I have been able to use the turtle with the joystick allready). Use the cam and stream it to my desktop and ontrol the bot from there. I have als o a bunch of servo's so to extend the project it would be cool to use the HAT button/switch of the joystick to controll the position of the cam.
If I can make that all work I will be really happy and have made a greate accomplishment ....for myself at least 😉
so that is what I want to do now. Regarding the controls for driving the motors I have allready the TG6612FNG and that should work together wiht the Jetson. If not I will need to order another motor driver board anyhow to get better results a better motor driver board would help I have seen other suggestions hough so maybe I will change that in the near future
When it comes to ROS, the major benefits of this framework is the decoupling of functionality between different modules. The modular design allows for, what in software development is knows as, Separation of Concern, and while it doesn't really make sense for simple designs, when you get to multi layer processing, multiple sensors and functionality, it can help to avoid your code becoming a massive ball of intertwined functions.
ROS also allows for clustering and large scale networking, where you can have some functionality running on the bot, and some other functionality running on a server. For example: your bot could have a camera, it would publish the images onto the network, and a server processes these images and sends the results back to the bot. Is such a solution good for obstacle avoidance? No, because latency. Is such a solution good for image recognition and navigation planning? Yes. This opens up possibilities for swarm control, where swarms are managed through a central node/server and the bots handle the collision avoidance and hardware driving.
There's obviously a cost to this decoupling, including learning curve and coding time. So in deciding whether to go for ROS, you need to think about the cost benefit. Are you going to build a robot which you'll keep expanding on and adding features as you go along? If yes, then it might be worth investing time into learning and implementing ROS. Are you building a simple robot which just moves according to your input from a PS controller? Probably not worth the investment.
If what you're after is knowledge and experience, then the cost is irrelevant. 🙂
Actually I'm going for both and I can explain that.
- This is hobby and I want to learn and am intereseted in programming, electronics and always love to learn
- Yes im building a robot whcih first should be able to take some simple controls from a joystick but the idea is to make it a bit smarter. I want to do this step by step.
Regarding costs I don't see a huge point, not that I have unlimited funds but ROS is free and the price of a ESP or Arduino together with a RPI is not that high ( at least for me). But I think if one goes for high quality sensors for example yes costs can get really high. A good LIDAR may cost you easily 1000 USD.
I think it is rather the time because ROS has a steep learning curve that make people run away from it. Cause it really takes some time to get used to it. But I personally like it and I don't bother it will take me a year to get my project running as I want it. But that is offcourse for everybody different.
I am learning at the moent ROS and get into so many issue but I really love it to learn new things so when I tackle a issue it gives me the satisfaction that I want from a hobby so all good
But everybody is different in this point of view. Anyhow thanks for your reply much appreciated!
When I was referring to cost, I was referring to the time investment you'd need to put into it rather than the monetary side of things. As you correctly stated, ROS is a free platform and thus monetary wise, there isn't any cost. But time wise, (which is a finite, non-renewal resource), there's a bit of investment to make. But if your intent is to learn, then the cost is irrelevant as knowledge is priceless.
I'm planning on using ROS as well for my project, so I've been reading about it and trying it out for several months, and I can attest to the initial learning curve there is to it, and I've only covered what I consider the basics thus far.
If I was in your situation, with your use case, then I'd definitely go for ROS - you seem to have a good plan laid out with incremental enhancements which fits perfectly with a platform such as ROS.
Good luck! I'll be on the look out for updates from your project 🙂
I have been bussy now or about two months with ros and following tutorials but still making newbie mistakes so it will take a while. But it is fun, when I started I did not know anything about Linux, Python and a little bit about C++. So it was pretty hard to get started but slowly I start to get the grasp of it.
Than I thought ok now I jump back into the hardware and programmed a Arduino with some new hardware in order to get my motors running good. Well this took me a while also but finally got a decent PID control for the motors. Than I though now I should publish the output of the motor speeds to my ROS Master and offcourse I got errors hahahaha a new adventure....But all fun it is hobby
Finally a new update regarding my adventure with robotics and ROS. I has taken a while but besides lot of other personal stuff such as work and family it took me a while to get things running up and working as I would like to. Here a video which I would like to share with the hope to inspire others to make a robot for themselfs as well.
I hope there are people enjoying to watch this
Thanks for the video, its good to see a bot on the move. With all those flashing lights I will think of it as a disco bot and expect it to do a jive in your next video. ( maybe a twist -all together now - 'Lets twist again last we did last summer, Lets twist again like we did ..... many many moons ago 😀 )
I suspect you are confusing two different musical eras.
Jive and Twist predate The Beatles coming to fame in 1962, albeit the "original" Fab Four did covers, including Twist and Shout in 1963. Let's Twist Again is 1961.
Disco era is 1970s, with Saturday Night Fever of 1977, based on experiences of most of the 1970s.
This mixing of eras could be very confusing to a robot too young to remember their musical origins. 😀 😀 😀
Best wishes and happy nostalgic memories to all! Dave
Although I know these musical eras it also has to do with preference but the Beatles I like but I prefer more a bit wilder stuff from the 60's and 70's. Things like Pink Floyd, Fleetwood Mac (the old formation), Uriah Heep and so forth.
Regarding the lights it was a small fun detail that I liked to use same as ships and planes have. I'm pretty happy with this result as some forum members remember maybe I had quite a struggle with speed detection and PID usage, all is implemented and the next adventure will now be odometry and camera comtrol with the hat-switch on the joystick.
Almost forgot, did you see the fuse I placed? That was a nice idea from you that I applied 😉