I have read in the Internet that in the most cases people prefer using INS (accelerometer + gyroscope) + GPS sensors. I have read they can compensate each other errors, but is it possible to obtain the same result using accelerometer + magnetometer + GPS sensors? Or accelerometer + gyroscope + magnetometer? I have also read that gyroscope requires lot of battery, so my question is: what is the system which can give me the same result of IMU + GPS, without using gyroscope? If it is not possible, can I just use 9 axis system without GPS?

I know that it depends on the application I want to use, but just let's consider I would like to realize a medium/high level application, so I need it to be precise, but I can afford to have some (very) small errors. If there is a paper which compares the performance of the system I mentioned, please link it :)

Thank you in advance for your answers.


closed as too broad by PeterJ, Neil_UK, Daniel Grillo, Bimpelrekkie, Chetan Bhargava Apr 1 '16 at 19:27

Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 1
    \$\begingroup\$ "Precise" means very little until you tell us what it means. Does it mean millimeters, meters, kilometers? Minutes of arc? Tens of degrees? \$\endgroup\$ – Scott Seidman Mar 25 '16 at 12:40

Each sensor has specific characteristics:

  • Accelerometer: Provides detailed information about linear acceleration. Can also measure direction of gravity (partial orientation) under some circumstances. High sample rate, but subject to scale and offset errors. Integration to get velocity and position leads to drift.

  • Gyroscope: Provides detailed information about rotation rates. High sample rate, but subject to scale and offset errors. Integration to get orientation leads to drift.

  • Magnetometer: Provides drift-free information about orientation, but absolute accuracy is low, depending on knowledge of local magnetic field and the distortion effects of nearby iron. Moderate sample rate.

  • GPS: Provides drift-free information about position and velocity. Interferometry can be used to get information about orientation as well. Low sample rate.

The INS provides high sample rate data, but is subject to drift. Therefore, high-accuracy systems combine that data with magnetometer and GPS data to eliminate the drifts and get the best possible answer at a high sample rate.

  • \$\begingroup\$ I would disagree with last sentence, the GPS can't be used to compensate drift, it's just an additional device that provides position and velocity. If the system is at the same position, and moves up/down, rotates,...the GPS isn't of any help. \$\endgroup\$ – Marko Buršič Mar 25 '16 at 13:07
  • 1
    \$\begingroup\$ @MarkoBuršič: You'd be wrong. GPS supplies absolute position in 3 dimensions, albeit with some error distribution. It does NOT drift. \$\endgroup\$ – Dave Tweed Mar 25 '16 at 20:51
  • \$\begingroup\$ Ok, would you please explain how would you measure gravity angle with GPS, or rotating speed, or heading (all at same position of course)? You can't compensate an IMU with GPS. \$\endgroup\$ – Marko Buršič Mar 25 '16 at 20:58
  • 1
    \$\begingroup\$ @MarkoBuršič: I can and I do. I've been working in this field for a while now, supplying high-accuracy, high-rate GPS+INS systems for both military and commercial applications. A thorough answer to your question would be far too long for this format, but as I said in my answer, you CAN get orientation information from GPS by using multiple receivers, and either DGPS or interferometry techniques. My systems use the latter. \$\endgroup\$ – Dave Tweed Mar 25 '16 at 21:06
  • \$\begingroup\$ Please, I am intersted, show some links where the GPS can replace the IMU. From your point of view you don't even need the IMU. \$\endgroup\$ – Marko Buršič Mar 26 '16 at 8:19

It depends what are you needs. With acc+gyro+mag you get attitude sensor, but you don't have the speed nor exact heading. Never heard of gyro using lots of batteries, probably you confuse it with true mechanical gyro with motor and inertia. There is no comparison between GPS and IMU, GPS gives position, path speed, heading; IMU gives acc, speed of rotation, attitude.

  • \$\begingroup\$ Thank you for you answer! I understood what you mean. About the gyro I have found it here (slide 38) google.it/… They wrote that a system without gyro uses 90% power less than the one with a gyro. So I suppose gyro uses lot of battery. \$\endgroup\$ – Enrico123 Mar 25 '16 at 10:21
  • \$\begingroup\$ @Enrico123 If the device is a TV remote controller then you can use also without gyro, but in the same way why would you need a GPS? If you need high dynamics then gyro is a must. There is also a pressure sensor for altitude, you forgot to mention. New what is your need? \$\endgroup\$ – Marko Buršič Mar 25 '16 at 13:11
  • \$\begingroup\$ Well I am just thinking about system (made by arduino) which can track my movements. I am at the beginning, but I am just trying to understand what kind of sensors I need:) \$\endgroup\$ – Enrico123 Mar 25 '16 at 13:20
  • \$\begingroup\$ Probably 9 sensor IMU for sports tracking, it will track your posture. If you want also the entire path where have you been and velocity (for runners) then also the GPS. There are various fusion algorithms, most known are Madgwick's x-io.co.uk/open-source-imu-and-ahrs-algorithms and propitery Invensense DMP already residing in their MPU devices invensense.com/products/motion-tracking/9-axis. \$\endgroup\$ – Marko Buršič Mar 25 '16 at 17:01

Not the answer you're looking for? Browse other questions tagged or ask your own question.