Robot driving fully autonomously using Marvelmind Indoor "GPS"
ฝัง
- เผยแพร่เมื่อ 30 ก.ย. 2024
- marvelmind.com/
Fully autonomous robot is driving by its own relying on:
(1) Marvelmind Indoor "GPS"
(2) On-board odometry and inertial units
The robot receives coordinates for the key points to visit from the user (table on the right) and then creates and follows the path by constantly correcting its position against the path. Coordinates are formed automatically in the Dashboard by simple clicking on the map.
Distances between beacons is up to 30 meters.
It is possible to cover with precise "GPS" the full campus by installing more beacons every 20-30 meters.
The same map support up to 250 robots.
I have project same like that. What should i buy for the component. And are the beacon work outdoor?
Hi sir, I have a project similar to yours. Is the car or vehicle that I made can follow the gps that has been made ?. Like for example I told my vehicle to go to a place
Yes, we have several outdoor (IPxx, IP56, IP67) variants as well as Industrial versions for most demanding customers: marvelmind.com/products/
WE need thé process, please.
What exactly do you need? ;-)
Why is this thing swiveling so much?
Thank you :-)
If something is unclear or we made a mistake, please, let us know. We will update and improve.
Hi, I'm wondering how many beacons it took to cover the area in this video. I see it advertised that there can be up to 50 meters between beacons. If that is the case, then to cover an area of about an acre outdoors I would only need four beacons if positioned properly? Or do thy need to positioned at the corners of a perimeter? Also is there any kind of environmental protection for use outdoors? Thank you.
Hello,
There is a table of distances on the left top. The area was covered by 4 stationary beacons.
Yes, in ideal conditions the distance between stationary beacons can be up to 50 meters. In real life environment we recommend to use up to 30 meters.
Area of 1000m2 (~30x30m) can be covered by 4 stationary beacons. Placing in the corners is just fine.
Current implementation is absolutely not protected against rain, moisture, etc. In a couple of months we plan to introduce an IP67 version. But it is still in R&D phase, so delays may be possible.
BR,
Maxim
Can i have a source code for this program?
We have fully open and well-described interfaces: marvelmind.com/download/. But the source code is not open-sourced and is not planned to be, unfortunately
can you share the link of making this robot?
We are offering Indoor "GPS" for mobile robots. Not yet the complete robots. This demo is showing what could be done:
- The robot was build by us
- It is using our own Starter Set HW v4.9 with IMU: marvelmind.com/products/
@@MarvelmindRobotics we need a tutorial! This is amazing!
What robot can I buy and add this too?
Hello, would it be possible to make the equivalent of this system using arduino and some modules? What would I need to do to make such robot? It doesn't have to be as accurate as the one in the video, I just need it to be able to follow a map (like the one on the monitor), and drive following the shape of the map. Thank you
Yes, building robots is a great fun.
- marvelmind.com/solution/robots/ - our hints and guidance on how to build autonomous robots with the focus on positioning
Arduino is good enough as robot's brain. However, for positioning, you need something that Arduino doesn't have.
Actually, you shall choose between:
- Marvelmind - (ultrasound+radio)-based indoor positioning system - cm-level positioning accuracy. Good for Location + Direction
- UWB - decimeter(s)-level accuracy - somewhat a borderline for general driving and is already not good enough for Location + Direction
- Optical, for example, discontinued Intel RealSense
- LIDARs - expensive, bulky, and full of limitations, because it requires predefined and unchanging environment
You may consider:
- Odo+Optical
Our system is the easiest to use and integrate and the most accurate of them all. It is particularly good for autonomous robots because it is specifically designed for it.
See more:
- th-cam.com/video/hW9kYgiD4oE/w-d-xo.htmlsi=UM2a9xeJ-URcx1P7
- th-cam.com/video/Ql0YpMh9wX8/w-d-xo.htmlsi=aaHZNgRysS2_mBiD - use as a guidance
- th-cam.com/video/zg3oW_U_jdY/w-d-xo.htmlsi=dUW9FNZmJKxD-3r1 - compare different indoor positioning systems
Can u guide me how you made this awesome car
We are offering Indoor "GPS" for mobile robots. Not yet the complete robots. This demo is showing what could be done:
- The robot was build by us
- It is using our own Starter Set HW v4.9 with IMU for autonomous driving: marvelmind.com/products/
Hello, are you using any other software rather than Dashboard to do this project?
Does the Dashboard software help me move the robot to its specific destination?
Marvelmind Indoor "GPS" is like your real GPS. Does it help you move your robot from point A to point B? - not really. Autopilots do, for example, ArduPilot or PX4, if we talk about the PixHawk ecosystem.
However, our system is a "GPS+" kind of system:
- It provides GPS-like coordinates - even in the same NMEA format - native GPS format. We also support other formats, for example, uBlox format or Marvelmind internal format
- Marvelmind Indoor "GPS" system also provides the functionality of setting up the waypoints: you can send them from the Dashboard to your robot or drone: th-cam.com/video/vrn0N97A-8g/w-d-xo.html
- We have geo-fencing zones: th-cam.com/video/hV5bxmm4tNM/w-d-xo.html
- It is possible to send user data from the robot via mobile beacon to the modem and then to your system and back: marvelmind.com/pics/marvelmind_robotics_location_update_rate_and_user_data_transfer_rate_table.pdf
And so forth. There are many "GPS+" functionalities.
However, our Indoor "GPS" is not an autopilot. It doesn't control your robot or your motors. But we provide the most essential information for the autopilots: current location and the waypoints.
It is possible to use the API and control the system and not to use the Dashboard. You can build and use your own nice and lovely GUI instead, if you wish.
GPS Auto meters I like
Thank you
Are you guys using Paired Hedgehogs or just one?
The video was shot some years ago. We haven't invented the Paired Beacons configuration there yet. It was using a single hedge.
It is possible to use a single hedge, but it is significantly more complex than with the Paired Beacon configuration and it cannot work in static by definition if there are no other sources of information. It does work in dynamic but significantly more complex. Thus, we always recommend to use and use ourselves, of course - th-cam.com/video/xZ-jnoRNDE0/w-d-xo.htmlsi=jTHwgV0lvBJXSeCd and th-cam.com/video/UkkPnd6_0NI/w-d-xo.htmlsi=KH06X1-Mw7CrD0aB - the Paired Beacons configuration if the size of the robot allows. It starts working reasonable well with 30-40 cm base. More than then - very comfortably.
Is it possible to have the same accuracy on more then 100 000 m2?
Yes, it is possible. Just more and more submaps. We support today out of the box up to 250 submaps. Each submap can be up to 500-1000m2 of open space
The pulsing is do to location checks?
I am not sure what pulsing you are referring to ... please, elaborate on the question
Marvelmind Robotics the stop go of the drive or what I am calling pulsing as opposed to a smooth continuous drive
Why is the robot so slow and jerking constantly? Is the location update rate set at a low rate?
We are reasonably good in electronics and SW. But our HW/mechanics skills are so-so :-)
So, we just took an RC car, threw away its electronics and added Marvelmind instead. Since, the car's mechanics wasn't really suitable for turning on a small angle, especially on that surface, the car had to adjust too often and to overshoot all the time.
We have another mechanics. It is far more suitable for turning: th-cam.com/video/TBTw1njQ7QY/w-d-xo.html - as you see, it is far more steady. Purely, because of mechanics.
Let me ask you how the robot can navigate and how it gets the navigation instructions from where to go
- Robot gets a path as waypoints sent via our system.
- Robot also know its current location based on our Indoor "GPS".
- That was our system's part
- After that, robot makes decisions where to drive and where to turn based on its SW. This is not a part of Indoor "GPS", but a part of robot's SW
which gps module do you have used? whats the accuracy of it?
We use Indoor "GPS" by Marvelmind Robotics. It has precision of +-2cm or better. It is used as GPS and streams out data as GPS, but it has nothing to do with GPS as a technology. Thus, it doesn't need clear view to satellites and can work indoor or even underground.
The technology relies on Marvelmind beacons that are placed every 30 meters or closer.
thank you so much for reply. can you please suggest any gps module which have higher accuracy? we want to use it outside the room.
There are quite many RTK GPS modules on the market. They give cm-level precision
for RTK GPS we have to use base station. so it will give accurate location for particular defined area only.
Correct. There are even more restrictions with RTK GPS. However, depending on the details of environment, I would go either with Precise (+-2cm) Indoor "GPS" from Marvelmind or some of the RTK GPS solutions. I don't know other better options.
If you are OK with 10-30cm precision and don't need 3D (no Z), I would also consider UWB solutions.
I still do not know how to be able to transfer data from this product to the robot / robot can adjust the moves position. you can help me document about it?
The easiest way is to use examples: www.marvelmind.com/#Interfaces
2b) C library:
Here is the C code + guiding document
Supported operating systems:Microsoft WindowsGNU/Linux (including Raspberry Pi)Mac OS X
2a) Python library:
Here is the Python code for Linux-based devices integration via virtual UART via USBLibrary is easy to use. Also, here is detailed description and example
Code is fairly universal; Tested on:Ubuntu 16.04Raspbian JessieMac (OS X 10.10)Python v2.7CRC calculation is done in "pythonic way", i.e. library crcmod is usedThere are handlers of different exceptions, for example, non-standard states of COM/USB
The Python library provides quick and easy ability for integration of
ultrasound navigation system with various linux-based devices. Among
them are Raspberry Pi and other single-board PCs. This code is free to
use in your applications (please include a reference link to this web
page).
In this you will be able to get the coordinates from our system. Then your robot shall take them and use accordingly.
Hi,
Can u help you help with the same project that you have made it?
i woukd like have the list of the components and how to assemble it step by step
Sooner or later, we will have enough time and make similar robot commercially available. Send your pre-orders and wishes to info@marvelmind.com. Special discounts for pre-orders :-)
Can we do it with ardunio and gps
Sure! Moreover, we have sample codes for Arduino and we stream out data in native GPS format - NMEA0183, i.e. your external system will be believe that it is connected to a real GPS, but a very precise one and the one that works indoor: marvelmind.com/download/
Do you think the indoor "gps" will work as precisely and accurately in water environment ? When the model is moving in water ?
any chance of reflection of ultrasonic waves used by your positioning system being reflected from water surface ? Will that affect its accuracy ?
if the water surface is turbulent , waves in water what would be the effect ?
Current version of beacons' HW doesn't support underwater configurations. We are discussing a such a product variant too, but nothing practical yet
Yes, it will now: th-cam.com/video/5DGJeRrO3r4/w-d-xo.html. We have several water related cases: boats, ships, automatic docking, etc.
Have you used any obstacle detection method to drive the robot on due path ?
This demo focused purely on driving based on waypoints and constant comparison of the calculated path and measured with the help of Indoor "GPS" current location. The robot was blind and didn't have any obstacle avoidance activated.
Can u put a video on how u did these project
Well, the configuration was:
- 4 stationary beacons
- 1 mobile beacon
- Typical configuration: marvelmind.com/product/starter-set-hw-v4-9-imu/
And our own robot with Kalman and odometer.
Hey sir i want to talk with you
With pleasure: marvelmind.com/#contact
How do you make your robot follow a path? I mean always to go by road and not go off road?
We use Precise (+-2cm) Indoor "GPS" by Marvelmind Robotics (www.marvelmind.com/):
- There are stationary beacons every 30 meters or so
- There is a mobile beacon on the robot
- In the Dashboard - the monitoring and control system of the Indoor "GPS" - you click the map and set the way points
- The robot receives the way points locaton and compares its (robot's) location against the way points and corrects its movement and direction to stay on the path. The key information for that - robot's current location. And robot gets it from the Indoor "GPS": www.marvelmind.com/
is this car controlled by the marvelmind beacon?
Yes, it was: marvelmind.com/products/
Hi! I'm an undergraduate,
could you please give your instructions for my mini project, I need to build fully autonomous driven system for the robot using GPS.
what are the hard wares and soft wares you use for this robot?.hope your response ASAP.
Hello,
There are several key navigation elements in our robot:
1) Indoor "GPS" by Marvelmind: www.marvelmind.com/ - it provides key location data against which the robot makes the decision where to go and how to correct its position against the required path
2) We employ odometer made also by us
3) We also rely on inertial measurement unit - gyro + accellerometer - also by us
4) And the key is the SW that combines all these data and use the best source at any given time.
BR,
Maxim
thank you maxim...