Jump to content
I2Cdevlib Forums

Ferrocen

Members
  • Posts

    13
  • Joined

  • Last visited

Everything posted by Ferrocen

  1. See second post at Using the MPU6050 to analyse human gait. You can find some interesting undocumented function in source code
  2. New small update: added a "Directional Gyro" device. NOTE: This source code does not contain any filters prevent drift from y-axis (YAW).
  3. Small update: added one of several flight devices such as Attitude Indicator for data visualization.
  4. Hi all! I uploaded my simple example, based on "Embedded MotionApps 5.1". It's available at GitHub. This is first release without any optimization. It tested on "Arduino Mega 2560" and MPU-6050(MPU-9150). This code provided "as is". Many functions like as "pedometer" and "taps" not tested. Maybe you will find something useful for your project. WARNING: INT pin from MPU-6050 must be connected to 19 pin ((RXD1/INT2)PD2) on Arduino Board.
  5. I also uploaded my simple example, based on "Embedded MotionApps 5.1". It's available at GitHub. This is first release without any optimization. I use "ATmega2560" and MPU-6050(MPU-9150). This code provided "as is". Many functions not tested. Good luck!
  6. I don't have any info about "Embedded MotionApps Platform Release 2.0.0". It's currently unavailable for downloading. I have nothing to compare the latest version available. I think it is more correct to use the latest version. Please pay attention to the achievements of Pansenti GitHub MPU9150Lib.
  7. Hi Jonathan! About ready solutions i can say that "Embedded MotionDriver 5.x.x" has a "pedometer" function, with which you can get the number of detected steps and duration of walking time. The driver also can detect of the taps along the X, Y, and Z axes and generate interrupt. You can set some parameters such as: tap threshold for a specific axis axes which will register a tap length between valid taps minimum number of taps needed for an interrupt etc... But you have to adapt this driver for your project and conduct some experiments with it. I hope it will be useful to you.
  8. No no no ... When I gave links to the Android repository, I was talking only about the possible usefulness of the source code - fully they are not suitable. As for the "Embedded MotionDriver 5.1", it quite the contrary is suitable for adaptation for the "AVR platform" - in this case you need only include the header files which specific for your compiler or IDE and assigning "read\write" functions of I2ะก bus. Method of "read\write" data for MPU in this source code is fully compatible with the functions (for ex. "I2Cdev::writeBytes" and "I2Cdev::readBytes") from the library (I2Cdevlib) of Jeff Rowberg. In addition, in the description of the driver said: this driver currently works for the following devices: MPU6050 MPU6500 MPU9150 (or MPU6050 w/ AK8975 on the auxiliary bus) MPU9250 (or MPU6500 w/ AK8963 on the auxiliary bus) About Pansenti's library for MPU-9150. This code fully based on "Embedded MotionDriver 5.x.x" and more adapted to work specifically with MPU-9150, although compatible with the MPU-6050. And also focused on the Arduino platform. So the best solution would be to base the "Embedded MotionDriver 5.xx" and by simple manipulations to adapt it for your project.
  9. As was mentioned earlier by Jeff Rowberg turnkey solution for platform Arduino, based on "Embedded Motion Driver v5.1.1 SDK", located on the Pansenti GitHub MPU9150Lib. It has some simple sketches showing how the software can be used and also to calibrate the accelerometers and magnetometer (if it is connected).
  10. I rewrote on whip up (crude method) standard functions from the Arduino libraries for working with I2C interface (twi.cpp, wire.cpp) and a library "I2Cdev" (many thanks to Jeff Rowberg for his invaluable work!) on "pure C" (I do not use Arduino-oriented libraries in my projects). Then, in the same way i adapted the driver source code of "Embedded MotionDriver 5.1" for compile with WinAVR. And it works (look at the picture)! Now i need to clean up the source code (at the moment the code size is greatly inflated), and to understand in detail the interaction of functions with the DMP-kernel. This is all to the fact that you can quite successfully use the available driver from "Android Kernel". Testing "MotionDriver v5.1" on Atmega2560
  11. Apologies for perhaps a strange question, but somebody found something useful from source files of InvenSense (drivers for InvenSense MPL-sensors, MotionApps 5.x.x code, example initialization DMP, example of "auto calibration", etc) for Android Kernel? Look for example on these android Git repositories: Link 1 and Link 2 And some files: Embedded MotionDriver 5.1 UMPL Source Code Embedded MotionApps Platform Functional Specification 2.0 MotionDriver Tutorial I look that in different revisions of the source code "InvenSense MotionDriver" observed different size of "DMP Firmware." In some "DMP_CODE_SIZE = 2634", and in other "DMP_CODE_SIZE = 3062" (in MPU memory banks). If I'm right then in the firmware laid some functions and data processing algorithms, which are updated\corrected\complemented with new release of the "InvenSense MotionDriver", and to which access is provided through the "user library". I think that can be quite successfully explore and adapt the available source code in accordance with our needs. Correct me if this is not so. UPDATE: I also uploaded my simple example, based on "Embedded MotionApps 5.1". It's available at GitHub. This is first release without any optimization. I use "ATmega2560" and MPU-6050(MPU-9150). This code provided "as is". Many functions not tested. UPDATE2: Added small optimization options for GCC compiler and linker. Also fixed some code for memory optimization.
×
×
  • Create New...