9-Axis IMU LESSON 8: Using Gyros for Measuring Rotational Velocity and Angle

[@media1critter]

Paul, I just wanna say thank you so much for doing this :) This series would be worth $5,000 at a private university, maybe even more, and in my opinion is far more engaging. Your videos are extremely relevant to my current interests, full of humor, very laid back, accessible from anywhere (I was literally just watching this on the subway) and can be replayed over and over so that I’m truly learning at my own pace. Because of folks like you, the whole world gets to enjoy electrical engineering lectures that are so good that they equip us with the skills needed to make an impressive portfolio of projects even by professional standards. Much love! ❤

[@quaternion-pi]

Great motivation for the need for a complementary filter...very clearly explained. Wish I had teachers like you when in school. Thanks!

[@paulmcwhorter]

I do think we are hitting some really important engineering concepts in this series, which are themselves much more important than IMU. Low pass filter and fusion of multiple sensors providing related data. Love this stuff

[@blinkmaker]

@@paulmcwhorter Could you show us how to program other types of filters? Including a Kalman filter and explain how it works?

[@gweliver]

I give you a thumb ups for all the information you provide with this TUTORIAL and now I will watch Lesson 8 with comments (if needed).

[@Hino_55]

At first I was confused on as to how the millis() command really work. Soon, I realized that it was a "time-grabber"...more specifically "elapsed time-grabber". As soon as we turn ON the arduino, its internal clock starts. The millis() command helps us to grab the instantaneous time. (instantaneous in a sense that it grabs the time as soon as the program reads that line of code containing the millis() command). Thank you for this wonderful lesson Sir Paul! Trade-offs upon trade-offs these past few lessons. I'm excited how we can get the best of both the accelerometer and gyro to get more accurate and stable pitch and roll angles.

[@Fogaca135]

I lol every time Mr.Pault upload the code!!! Great lessons!

[@juanramirezjardua2082]

Thanks Paul for all these amazing videos. During the corona virus , staying at home, I’ve been playing with gyros that I had from different projects, including the BNO055, and your videos have made to understand lot of things that were black boxes when I only use libraries and prebuild code.

[@jstro-hobbytech]

Paul. Ive spent the last two days studying control systems, complimentary feedback loops, trigonometry, linear algebra, quaternions. i didnt want to write the code as i didnt want to jump ahead the lessons and im not entirely sure what syntax you are going to use but i know most of it. ive put like 9 to 10 hours into learning the maths and such all while renovating my kitchen, learning pyqt5 to help a friend debug his application plus studying bode plot analysis for different components for a swictcher psu, teaching my wife basic arduino and trying to decide on a more advanced visualization tool im going to learn for a long term robotics project im working on. i dont sleep much haha. it's all fun though and you only live once.

[@Happynewmama]

Thank you very much sir! I had thought that you could get an accurate angle measurement just from gyro readings. Now I realise it is much more complicated than that.

[@federext]

Paul, amazing videos!! Thank you very much for sharing all these knowledge in such a didactic way.

[@paulmcwhorter]

My pleasure!

[@federext]

@@paulmcwhorter do you recommend any special library for filters? Thanks Fede

[@jesser42]

Yes I got the HW solution right. Glad to know that my years of studying Mechanical Engineering wasn't a waste of time lol.

[@marcelolopes6228]

Amazing explanations! It gives me a lot of joy seeing your passion explaining all this stuff. Congratulations and thank you Sir!

[@neomatrix2526]

Great video. This is so helpful. You make excellent content. Clear, concise, interesting. Well done.

[@gregh9237]

thank you so mouch Paul!

[@toddduffett1378]

Thanks Paul, another great presentation.

[@jaynone9262]

I am still trying to get the temperature humity sensor to work. Been at it for 3 months. Lol. 1st project is easy. You do the led blink thing, no big deal. Just uploaded led blink and play with time on/time off. Figuring out DHT11 is driving me crazy. Almost ready to try something new. Been trying different script on the DHT11. No go... Rotational velocity way way over my head. Maybe need to read instructions again. Lol... Love the vids.keep them coming.

[@dcpowered]

YAY! Another tasty IMU lesson!!

[@陈霓-e8d]

cute papa, true reaction from a programmer when we download the program and wish to be workout!

[@Bob-zg2zf]

Crystal clear. 😍

[@anamaybelekar]

Great explanation...!!! But one thing I want to know, as you explained internal structure of accelerometer in previous lectures, can you please explain internals of the gyroscope, means how it works?

[@GCRickerNH]

Great job Paul. Can you please explain how the gyroscope works within the chip like you did with the accelerometer. Is there something in there spinning?

[@paulmcwhorter]

Not spinning, but something vibrating, which creates the inertia needed to measure rotational velocity

[@troyfrei2962]

Great video. It seem it needs a reset to 0 deg, to clear drift. But the problem is when to 0 the angle.

[@danielsaenz5570]

You went very in-depth on how accelerometers work, but not the gyros. Is that because you don't know how they work? Or are they very similar? Thanks Paul!

[@paulmcwhorter]

I understand gyros, but they are harder to explain. Has to do with rotational inertia, and how vibrating bodies dont want to change direction;

[@danielsaenz5570]

@@paulmcwhorter okay.

[@wayneswan3092]

HA HA! see my comment on lesson 7... I knew that we could use the gyro, but I'm still learning how to apply math to programming. I can do algebra and geometry all day long on paper. but when it comes to putting into a program, I get lost in that translation. So I folded like a cheap lawn chair! BUT I at least knew what needed to be done!

[@ManchesterBlackSheep]

Thanks for this. When you say fee you mean phi? What we'd pronounce as rhyming with high in British English?

[@arghachatterjee1353]

Please can you refer some materials or books which I can use to make a thrust vector controlled rocket both for sensor and PID part as you say that we should not rely on these for advance projects

[@nitzanbreitman]

Hey Paul, great videos! This whole playlist is truly amazing and helpful. I wanted to know, how come you are not calibrating the sensor after each reset? Earlier you showed that it is needed after a reset..

[@TheAneesha009]

Thanks Paul..Wonderful videos and I am following your all lessons. My question, Is it possible to get the trajectory of motion when we move BNO055 sensor in free space? Say, I moved IMU sensor diagonally in space to make semi-circular shape. My initial and final point of BNO055 sensor is fixed.

[@NathanPK]

I don't understand why we can add the gyro rates in rad/s to angles in degrees and have it work? Why don't we have to convert to deg/s with a 180/pi conversion factor?

[@deepakbhola1585]

Hey paul, cant we measure linear velocity from BNO005 and fuse this velocity along with the wheel encoders to get better estimate of velocity using kalman filters.. kindly reply. cant we convert this rotational velocity into linear velocity.

[@phildunkerley5855]

Thanks Paul, very good stuff. Can you clarify please how the Y axis is in degree units when the output from the gyro is in rads/sec as used in the Theta/Phi calcs?

[@johnganci933]

The BNO055 datasheet ( cdn-learn.adafruit.com/assets/assets/000/036/832/original/BST_BNO055_DS000_14.pdf) states that the default units for the gyroscope is degrees per second. If you look at Paul's output, you will see that it confirms this.

[@johnganci933]

Additional comments about BNO055 gyroscope output being deg/s instead of rad/s. My previous reply glossed over the real reason why this occurs. What follows explains in more detail why. If you read the Adafruit documentation for the Adafruit 9-DOF Absolute Orientation IMU Fusion Breakout - BNO055 and the Adafruit BNO055 Library (located on Github), you will see that the units for gyroscope values are supposed to be rad/s. This does not agree with the output shown in Paul's code, which is definitely in deg/s. 1. The original release, 1.0.0, of the unified sensor driver for the Adafruit BNO055 orientation sensor breakout, came out on May 1, 2015. The gyroscope vector values were returned as rad/s. However, the code in the Adafruit_BNO055.cpp source to set the gyroscope units to rad/s is commented, so the sensor is set to the default gyroscope units, which are degrees/s. The Adafruit code did a conversion assuming that that the values were in rad/s/ Therefore, the gyroscope values being returned are wrong! 2. On October 21, 2016, someone asked about "BNO055 wrong scaling for gyroscope" on the Adafruit forums. This issue was addressed by a fix to the Adafruit source that changed the code to return the gyroscope values in deg/s. The issue was closed April 20, 2017. On Github, this fix appears in the 1.1.3 release, dated July 5, 2018. The code is the same in the current release, 1.1.10. 3. A newer issue from May 21, 2018, which is still open, commented that "Not a problem per say, but looks like the library outputs gyro data in degrees/s, whereas the tutorial claims it's in radians per second?" and "Mainly a documentation issue (though I know the Unified Sensor lib prefers r/s, but that would be a breaking behaviour for this lib...)" This is precisely the behavior we are now seeing. It is not clear why Adafruit has never uncommented the code that sets the gyroscope units to rad/s. If you search on the Adafruit forums, you will find that some people have tried uncommenting the code, but state that it did not work. Perhaps Adafruit ran into the same problem.

[@jackd6134]

Possible to write code to find best coefficients to optimize the signal per given specs rather than guessing? Would require reading RT data to test.

[@sanasabah8460]

Dear paul, thank you for this tutorial, I followed each lesson in this series and things was great until Now, when I try to write the code ,The thetaG and phiG have a strange behavior, it it keep increasing and suffering from noise, I even tru to copy past the code from your website and I get the same problem. What should I do?

[@id-face2518]

me too I have the same problem, and I really need the solution !!

[@shardulsaptarshi5575]

@@id-face2518 @Paul McWhorter can you help?

[@T--gl3xp]

amazing video sir am using the same hardware that you are using but I can't get the same measurements of (thetag)and the (phig) from the gyroscope

[@matthewcollins1858]

Present, I was able to pull the data but didn't know what to do with it.

[@deepakbhola1585]

Will all these codes work for MPU 9250 ?

[@mechabotix]

Hello.paul Can we use 9250 with this code?

[@wayneswan3092]

Well, my first thought would be to average the two. Multiply them together and then divide by 2. But somehow, I'm guessing I'm way off the mark. But, I'm going to try and and see what happens. (Edit: Nope, didn't work, but noe I'm bettinging we'll implement the magnetometer now, aren't we? Wish I could figure out how to do it myself. But that's why I'm here, to learn!)

[@prashantajabe6636]

Please also provide the resourse you use for this

[@BizuMgr]

How Gyro's Theta and Phi are in Degrees without conversion ?

[@jonathangonzaga95]

Can I have a constant dt? I read my gyro data every 10ms.

[@陈霓-e8d]

Hello Paul, Thanks for your video! I am testing l3gh20 on a stm32f3discovery board. When the gyro stays still, the results (without calibration) don't approach to 0 at all. Do you have any suggestions?

[@陈霓-e8d]

Ok, i figure out myself. The unit of sensitivity is mdps/LSB, so the result I've got is in mdps!

[@wouter.k1570]

should this also work with the MPU6050? when I use these formulas with a delay(500) the output in the serial monitor will not go back to zero when my gyro is back in the stable position, is this because the code is not compatible or should i use a lower delay?

[@kevinshen9391]

use a lower delay like 100 or even 50, it should be better

[@KB1QVO]

I'm afraid that if you keep reacting to your code downloads that way, you're going to develop hyper-tension :) and … getting some sensor fusion in the next lesson would rock!!!

[@paulmcwhorter]

. . . wait for it! Should be next thursday the magic starts happening.

[@waltmetcalf5354]

Paul, every time I stop the serial monitor and start it again I loose calibration. Is this normal?

[@paulmcwhorter]

Yes, then the sensor constantly tries to calibrate itself, and then sometimes loses its calibration in operation. Wish we could just turn calibrate off and use known good values.

[@codecage9333]

@@paulmcwhorter But we're not 're-calibrating" when using the SerialPlot app? Actually the code has started the re-calibration steps but we seem to just ignore that when we look a data via SerialPlot.

[@pulesjet]

Would these work outside of the Earths Magnetosphere ?

[@kevinshen9391]

yes, it uses an internal measurement (I think, at least its true with most of these)

[@林韋辰-o8m]

"22:14" dt

[@superepicproportions]

All this talk about gyros is making me hungry...

[@jstro-hobbytech]

this serial plotter app is horrible paul. i'm going to try to find a better one and i'll let you know what it is so it can be used along with the course.

[@gionibegood6950]

why you are not using bn085? it is making all work for you, low noise, is computing everything internally, you made 10 movies telling what bno085 is doing for nothing