Autonomous GPS-driven rover - Post #1 - FEASIBILITY

Hi everyone.
I am new to the forum.
While not being an expert in electronics or programming, I have some school background regarding those subjects ... and still a lot to learn!
I am bringing up again the issue of a grass cutting rover with a twist.
I have searched several different options here or on other makers' websites without finding a way to integrate the different solutions.

This part of the forum is - inter alia - about feasibility, so I am asking a little help from the experts before moving to the design phase.

The rover I would like to build should be capable of:

  • starting from a base station that is used as a shelter from weather;
  • navigating across the garden using a series of pre-determined GPS waypoints;
  • turning on the cutting blade motor upon reaching the first waypoint;
  • going back to the base station for recharging (if some of the waypoints have still to be reached).
  • going back to the base station for recharging after the last of the waypoints has been reached and turning off the cutting blade motor.

I want to develop this project in steps, and the first one is feasibility.
While in the design phase, I will post specific issues to the relevant forum topics and cross reference the previous ones, so that readers can follow the thread if they want.

While I understand Arduino can manage several operations at once, I have some doubts about what Arduino board to choose for this task.
Moreover, I would also install some ultrasound sensors - which would trigger a change in heading - to avoid collision with obstacles like plants or trees. I know it seems redundant given the waypoints, but I would also account for GPS tolerance.
Attached is a block diagram of the shields I think I could use and a tentative function each of them should fulfil [in brackets].

QUESTIONS
What Arduino board do you think best fits the purpose?
Is the attached block diagram correct? Would you change/add anything?

The project seems feasible to me, but any suggestion, comment, remark of critique is appreciated.
Thanks

Block diagram.pdf (36.3 KB)

Hi,
I'm sure others will have ideas, but to start:

I want to develop this project in steps, and the first one is feasibility.

I agree. "Stepwise Refinement". Get each part working separately, THEN start to integrate them.

Hardware questions:

  • "Shields" stack (some do not stack!) on an Arduino. When you have many different functions you may find it better to use separate "Modules" that are not in "Shield" shape. That gives you flexibility in mounting them, and you can use whatever Arduino pins you want, with no dictates from a "shield"..

  • Think power design and "Where is Ground" from the beginning.

Some suggestions here: http://arduino-info.wikispaces.com/Arduino-Project-Planning-Electrical

Later you will be combining the separate parts that worked on their own. Some suggestions on combining them are HERE: http://arduino-info.wikispaces.com/CombiningArduinoSketches

Good Luck. (But what you REALLY need is Patience and Persistence)...

Any Arduino can handle a GPS module, an absolute orientation sensor like the BNO055, and drive a simple robot around a series of waypoints. There would also be memory and pins left over for collision sensors. I've used a tiny Pro Mini board to drive an autonomous boat around a lake.

That said, a fundamental limitation for your project will be GPS accuracy. At the very best expect your mower to miss waypoints by 1 to 2 meters, and at times by 10 to 20 meters, depending on trees overhead, nearby buildings, how many satellites are in the sky and where they are. You will want to investigate that very thoroughly before investing too much in the project.

Agreed with the comments about the GPS, I test various bits of trackers in my Garden and I know the position can vary a lot.

So that would be the first thing to check, build a test unit and log the GPS change in position to see what sort of variation you get. Nearby houses, trees etc can affect accuracy.

jremington:
[...] a fundamental limitation for your project will be GPS accuracy. At the very best expect your mower to miss waypoints by 1 to 2 meters, and at times by 10 to 20 meters, depending on trees overhead, nearby buildings, how many satellites are in the sky and where they are. You will want to investigate that very thoroughly before investing too much in the project.

I do agree. That's why I was considering proximity sensors. Thank you for the hint. I will definitely require more info about this.

Hi Terry and thank you for the hints.

terryking228:
Hi,
When you have many different functions you may find it better to use separate "Modules" that are not in "Shield" shape. That gives you flexibility in mounting them, and you can use whatever Arduino pins you want, with no dictates from a "shield"..

That's good advice. I will keep it in mind. Once all systems have been checked, I can see how to arrange my circuitry in a box.

terryking228:
Good Luck. (But what you REALLY need is Patience and Persistence)...

The force is strong in me... :smiley:
Thank you again and take care.