The effect of inertial measurement unit harmonic oscillations on reckoning algorithm error of pedestrian navigation system

Аuthors

Kuporosova E. S.

Kazan National Research Technical University named after A.N. Tupolev, 10, Karl Marks str., Kazan, 420111, Russia

e-mail: kuporos_lena@mail.ru

Abstract

To study the accuracy of inertial reckoning algorithm operation with pedestrian navigation system, a robot’s movement, imitating human gait, was modeled. The experiment with a human movement over closed rectangular rout with inertial measurement unit (IMU) mounted on the users’ waist was carried out to evaluate quantitative and qualitative characteristics of a robot’s movement. The study focused on the accuracy of the input data processing algorithm (without an external correction).

The human motion model with IMU was developed as a kinematic model of a robot, imitating a human gait. Sensors transfer functions are assumed equal to 1. The real experiment and data recordings were carried out with IMU ADIS16405. Sampling time  ≈ 0.025 s. Sensor readings arrays were processed by means of moving average with 15 samples window extent using MATLAB Curve Fitting Tool libraries. Fourier transform and Fourier spectrum modeling were applied to smoothed signals. Reference oscillations frequency of the pitching motion parameter is f₁ ≈ 1.55 Hz, and Reference oscillations frequency of lateral motion parameter is f₂ ≈ f₁/2 ≈ 0.77 Hz. Evaluation of oscillation amplitudes for the roll (γ_m = 1.13°) and pitch (ϑ_m = 0.69°) were made. The obtained experimental results were used in the robot motion model.

Conclusions:

1. Harmonic oscillations of IMU cause error in pedestrian navigation system without correction even during sensor ideal performance.

2. The greatest impact on human localization error is caused by IMU oscillations with respect to the horizon plane. The error value is linearly dependent on pitch and roll IMU oscillations amplitude.

3. With equal pitch and roll, IMU angular and pitch oscillations amplitudes can cause human localization error seven times greater than roll oscillations.

Keywords:

accelerometer, angular velocity sensor, angular oscillations, trajectory, а frame of reference, kinematic movement model, apparent acceleration, Fourier transform

References

1. Devidson P., Takala Ya. Giroskopiya i navigatsiya, 2013, vol. 80, no. 1, pp. 86-94.

2. Nguen N’y Man. Trudy MAI, 2012, no. 58, pp. 1-22.

3. Matveev V.V., Raspopov V.Ya. Osnovy postroeniya besplatformennykh inertsial’nykh navigatsionnykh system (Bases of construction of strapdown inertial systems), Saint-Petersburg, Elektropribor, 2009, 280 p.

4. «Analog devices». Datasheet ADIS16405, URL: http://www.analog.com

5. MATLAB & Simulink Release Notes for R2008a, URL: http:www.mathworks.com

6. Bendat D., Pirsol A. Prikladnoi analiz sluchainykh dannykh (Applied analysis of random data), Moscow, Mir, 1989, 540 p.

7. Marinushkin P.S., Bakhtina V.A., Podshivalov I.A., Stukach O.V. Nauka i Obrazovanie, 2015, no. 6, pp. 157-173.

Download