Hi guys,
I am working on a buoy project and need to measure vertical displacement. I need a faitly accurate IMU for this because even small inaccuracies can lead to huge drift when integrating acceleration. Budget for IMU is less than 200 USD. Are there any good IMUs you would reccomend?
Consumer grade IMUs are not accurate enough to allow use of double integration to measure displacement.
The main problem is subtraction of the acceleration due to gravity (vector), as explained in this note: https://web.archive.org/web/20180112063505/http://www.chrobotics.com/library/accel-position-velocity
Not even if I increase the budget to 3-400$? How about adding filtering techniques to a consumer grade IMU, would that work?
I'm not the right person to answer this, but maybe it would help to hear a bit more about your project. What is the displacement measurement used for?
If so, it would have been done years ago, and you could buy such devices cheaply on Amazon.
Remember analysis, there is always a "+ c" term when integrating. So no matter how much money you throw into the project, it will drift.
Its for a buoy project, so the vertical displacement will be used to measure ocean wave heights.
It will drift, but there are ways to correct the drift such as for instance sensor fusion with a GNSS system.
I don't have the papers anymore, I need to dig for them, but I have read several papers where people have demonstrated that it can be done.
I don't have any experience in the area, but if you don't need any fixed reference, would it be enough to measure vertical displacement relative to a running mean?
The buoy is not going to be moored, it's free-floating, so there isn't a consistent, fixed reference point to measure against. I would measure the displacement relative to a running mean. But the only way I know of is using the IMU's vertical acceleration and performing double integration over time to obtain the displacement, and comparing the buoy's position, derived from the double integration, to a running mean. That is why I need a fairly accurate IMU. The IMU will eventually drift as suggested by the other users. But there are ways of correcting the drift. I just haven't gotten that far yet.
I'm probably misunderstanding the issues involved, but if your readings drift, and your mean drifts along with them, is there an issue?
Do that and post the links. Why have you not simply built or used the apparatus they described?
Please explain your project in more detail, with realistic goals, distances and times to be measured, and perhaps someone can help.
There is an issue because you need the absolute vertical displacement to determine the wave heights. If both the IMU readings and the running mean drift significantly from the true vertical displacement of the buoy, then the wave height measurements will be inaccurate.
It's been a while since I read the papers. I had to take some time off for personal reasons.
The project is to create a small low-cost wave buoy that can be used to determine wave heights. Traditionally, wave buoys are very large, heavy and expensive. They are difficult to deploy and take up too much space. I want to create something similar to this, without costing a fortune. Spotter Buoy by Sofar
The goal is to eventually have an open-source low-cost buoy that is cheap enough to be deployed whenever and wherever. The buoys will estimate wave heights, take measurements and upload everything to a website. More sensors will be added eventually, like c02, pH, temperature etc for water quality. But that part is easy. The difficult part is to get accurate wave estimations without spending 2k on an IMU. Its difficult but it definitely can be done.
I will do some more research and post the links I find.
You talked about IMU, not IMU + GPS.
Anyway, where's the problem?
You could use Pressure Sensor or a TOF distance sensor to measure vertical distance.
The offset would be only few millimetres
Yes, because I am asking advice on which IMU to get. Theres no problem yet, there will be in the future though, but I need to start somewhere. I currently only have the GY-86 which is not good enough.
I know members here have tons of expertise and invaluable experience and probably played around with different IMUs. I want to know what's the best IMU I can get for around 200 USD.
you could just get an ardupilot flightcontroller, it comes with everything you might want .. plus logging etc.
What can you get for $200?
The ones suitable for dead reckoning (double integration) start in the range $5000-$10000, last time I checked.