Mobile Robot System
20 Jul 2021 05:26 #215377
by dhruv
Replied by dhruv on topic Mobile Robot System
It's not so much about the mecanum or omni wheels but the type of roller i.e. PE or PU rollers
And the biggest cut will be a straight cut of 15m. The shapes to be cut are going to be basic rectangles and semicircles for most parts.
And the biggest cut will be a straight cut of 15m. The shapes to be cut are going to be basic rectangles and semicircles for most parts.
Please Log in or Create an account to join the conversation.
20 Jul 2021 13:16 - 20 Jul 2021 13:16 #215411
by Grotius
Replied by Grotius on topic Mobile Robot System
Hi,
I had thought about using two such laser distance measurement devices with reflective surfaces at the boundary of the work area so the readings will be further more accurate. There are two problems though. In continuous measurement mode, the measurements are made every half second on an average. So, the frequency seems insufficient. And the second part is interfacing the data into the system. Can we read the distance data into our system from their proprietary devices?
The original device will do it at a interval of 0.5sec, following the spec's.
If you strip apart the device, you can connect to the chip or laser interface and read out data at 1ms interval. Reading the data into hal should be no problem. There are plenty solutions how to do this.
To mention. The idea is to use a robot on wheels. A cnc portal is a good alternative. But when the question is about using a
robot on wheels. I give the advise what would work using a robot and what would be a cost efficient solution.
I had thought about using two such laser distance measurement devices with reflective surfaces at the boundary of the work area so the readings will be further more accurate. There are two problems though. In continuous measurement mode, the measurements are made every half second on an average. So, the frequency seems insufficient. And the second part is interfacing the data into the system. Can we read the distance data into our system from their proprietary devices?
The original device will do it at a interval of 0.5sec, following the spec's.
If you strip apart the device, you can connect to the chip or laser interface and read out data at 1ms interval. Reading the data into hal should be no problem. There are plenty solutions how to do this.
To mention. The idea is to use a robot on wheels. A cnc portal is a good alternative. But when the question is about using a
robot on wheels. I give the advise what would work using a robot and what would be a cost efficient solution.
Last edit: 20 Jul 2021 13:16 by Grotius.
The following user(s) said Thank You: dhruv
Please Log in or Create an account to join the conversation.
29 Jul 2021 21:05 #216270
by andypugh
Replied by andypugh on topic Mobile Robot System
I think that I would be tempted to try dead-reckoning using the IMU.
Just integrate the accelerations and see what repeatability you get.
If you can get that position back in to HAL as X and Y then you can use a PID loop to drive the motors.
This would not need encoders, and would be insensitive to wheels slipping, jolting over cuts and so on.
So, first, just try it. Move the bare IMU around the room, to a known point, and back to the start. Look at the errors in the calculated position. if that is good enough then an awful lot of the complexity of the project disappears.
Just integrate the accelerations and see what repeatability you get.
If you can get that position back in to HAL as X and Y then you can use a PID loop to drive the motors.
This would not need encoders, and would be insensitive to wheels slipping, jolting over cuts and so on.
So, first, just try it. Move the bare IMU around the room, to a known point, and back to the start. Look at the errors in the calculated position. if that is good enough then an awful lot of the complexity of the project disappears.
Please Log in or Create an account to join the conversation.
30 Jul 2021 00:08 #216285
by cakeslob
Replied by cakeslob on topic Mobile Robot System
seems far off what everyone else is saying, but you could use 2 draw wire encoders, they sell them in 15000mm -45000mm lengths im seeing. fix the4 ends at 2 corners of the part and use trigonometry to get position.
www.siko-global.com/en-ca/products/linea...re-actuated-encoders
www.kuebler.com/en/products/measurement/...s/draw-wire-encoders
www.siko-global.com/en-ca/products/linea...re-actuated-encoders
www.kuebler.com/en/products/measurement/...s/draw-wire-encoders
Please Log in or Create an account to join the conversation.
30 Jul 2021 06:26 #216317
by Bari
Replied by Bari on topic Mobile Robot System
www.kuebler.com/en/products/measurement/...product-details/D135
max. linearity ±0,02 % x 15,000mm = +-3mm
1220 € a budget killer
max. linearity ±0,02 % x 15,000mm = +-3mm
1220 € a budget killer
Please Log in or Create an account to join the conversation.
01 Aug 2021 03:06 - 01 Aug 2021 03:07 #216579
by andypugh
Replied by andypugh on topic Mobile Robot System
Luckily this seems to be an academic project, so something that works, but poorly, with a detailed discussion of how a better solution is possible but out-of-budget will probably still get good marks.
1mm in 15000mm is way outside camera resolutiion, so webcam feedback probably isn't an option.
1mm in 15000mm is way outside camera resolutiion, so webcam feedback probably isn't an option.
Last edit: 01 Aug 2021 03:07 by andypugh.
Please Log in or Create an account to join the conversation.
01 Aug 2021 03:27 #216580
by Bari
Replied by Bari on topic Mobile Robot System
You don't use a single camera over the 15000mm. Multiple fixed cameras, 2x per sq meter for 1080p or cameras on the bot scanning a grid above or below.
Please Log in or Create an account to join the conversation.
19 Sep 2021 17:15 - 19 Sep 2021 17:16 #221121
by Bari
Replied by Bari on topic Mobile Robot System
@
dhruv
How did you end up building it and how did it perform?
How did you end up building it and how did it perform?
Last edit: 19 Sep 2021 17:16 by Bari.
Please Log in or Create an account to join the conversation.
Time to create page: 0.083 seconds