Hi! I am an engineering student, I'm in my second semester. for this one, I have been left a project to make a robot that receives some coordinates assigned (by me) and must move after entering these values (x,y) would be approximately 5 positions in a space of 5*5 meters. I have thought that for the positioning and coordinates system the use of gyroscope-accelerometer, incremental enconders and just saw that with an optical mouse sensor I could get my x,y coordinates, I have thought this because I have seen that an error accumulates between the gyroscope and wheels, so I would use these sensors to give my initial position and the rest of positions correctly. But I have many doubts before starting my project, do you see this viable? Or should I just keep only the gyroscope and accelerometer with the enconders? In addition to that, are there other economical options? (I have seen the odometry but also over time I will accumulate error) someone suggested me to do this with servos, it could make things easier for me but I do not see that option. I appreciate any ideas or corrections.
Hi @elementart,
welcome to the arduino-forum.
The most important question to answer is:
what is the demanded precision of reaching a certain position?
Very interesting idea which in my opinion is worth testing it.
How good the readings will be will depend on the surface the robot has to drive on.
- Is it a superflat homogen surface like in a class-room?
- Is it the parking lot in front of the university?
- gras?
Some more questions:
Are you allowed to do referencing before each new move to a xy-coordinate?
Referencing means drive the robot to a certain position
example: lower left edge
And align the body of the robot to x-axle and then say new
"now I am at position x=0 y=0 looking straight forward along x-axle.
I guess it will be a very interesting experiment what precision you can achieve with using
to calculate your new position.
Without having any experience I am sceptical about the precision.
Does the field have vertical walls at the borders of 5 x 5 m?
Then using Time of flight sensors (ToF) in combination with a compass might be an option.
These sensors are available for very different distances.
You can even measure a distance of 100 m at a precision of 2 mm if you pay the price.
Four sensors each one looking horizontal with angles between them 90 degree measuring the distance to the wall the sensor is looking at. Combined with the measuring of the compass that delivers the rotational angle in relation to the coordinantesyou could calculate the position.
You wrote:
becoming an engineer means to be able to work in a certain way to develop new things.
You should analyse or ask what are the criteria to achieve the best possible grade for this project.
Is it for working and documenting in a certain way even if the result would be
with using this kind of equipment the achievable precision is as low as ....
I used the following methods you measure the perfomance ....
or it does just not work:
the reasons are:
1.
2.
3.
The problems are ....
the following things ..... must be improved
With which would show you can work with methods that engineers use.
or
do you get the best possible grade only in case the project is successful
and then again what is successful in regarding the precision?
Startposition x=0 y=0
Would be successful if you enter coordinates 1m 1m
and the robot moves something in between 40 to 50 degrees and stops at a position with allowed deviation +- 50 cm?
= very low precision.
I am very sure the precision must not be within less than a micrometer.
But what is the demanded precision?
best regards Stefan