can mpu6050 be used in mopping robot

By stevewee , 10 December 2012
Forums
SmartMotion® MEMS Motion Sensors

a mopping robot should run straight foward,but some times with the inference of other problem，after 10meters running,the angle is changed.so straight foward is very hard to achieve.
can mpu6050 be used to cope with the very slowly angle shifting?
from the datasheet ,it seems the chip is fit for fast angle change. :)

stueckrath.c

13 years 3 months ago

The magnetometer on board should give you a true accurate heading. However, this is dependent on there not being large metal (magnetic) objects in the nearby environment (Refrigerator, Oven, Steel Filing Cabinet, etc), or lots of current through electrical wires that would generate magnetic fields. These all impede the magnetometers ability to read North accurately. You can just hold a regular compass near these objects to see where you may have problems.

phpbb Post ID
22086

stevewee

13 years 3 months ago

thanks a lot yorisimo .
my mopping robot has installed a strong magaetic device,so electronic compass is not fit for adopted ,is there any other method to navagination(indoor navagation)?as this is consumer device, the navigation price shoud be cheap,not extend 10$.the ordaniary indoor navagation method:infared,ultrasound etc have problem:unstable and expensive.

now i installed the mpu6050,but really unsure if it can used for navagation.

this robot is our latest housekeeper robot,but is sitill on the development.

phpbb Post ID
22089

stueckrath.c

13 years 3 months ago

Oops, I read 9150 not 6050 because that's what I'm using. The 6050 does not have a magnetometer, but without that you will not get accurate heading information over time. In about 2 minutes you will be off by a few degrees. This may be sufficient if you can reset the heading based on the walls in the room or something

phpbb Post ID
22092

guy_mcilroy

13 years 3 months ago

MPU 6050 would be suitable for the application, though 9 axis sensing is typically desired for yaw stabilization. The gyro will be able to respond reliably to angle change down to its minimum resolution of 27.5 degree per hour. However, our eMPL library has some features which are tuned to hand-held operation which would possibly cause isssues on a robot vacuum, including no-motion calibration and magnetometer bias estimation. I would recommend to take a look at the raw data using our MotionFit SDK board.

phpbb Post ID
22095

stevewee

13 years 1 month ago

i have tested that mpu6050 will shift 0.26degree every minute when the moto is still. but when running on mopping robot,the angle will changed a lot. i guess maybe the inerfere or other EMC problem affect the 6050?
this is my product that adopt mu6050 to stablilize the direction.
as running straight is the basic for the algorithm of full cleaning coverage of the room,so this is very important.

the video of my product that adopt 6050 can view here
http://player.youku.com/player.php/sid/XNTEwNDQyODY0/v.swf

phpbb Post ID
22098

guy_mcilroy

13 years 1 month ago

Integration error during the periods of increased noise is likely the culprit for your testing error. I would recommend to use the gyro raw data at a fast rate (up to 8kHz), and forgo the 6 axis sensor fusion for the sake of increased integration speed, to see if that results in a more accurate angle estimate for your environment.

phpbb Post ID
22101
phpbb Topic ID
14904