Robotics StackExchange | Archived questions

doubt regarding acceleration values from phidgets IMU

Hi all,

I am using Phidgets Spatial 3/3/3 IMU sensor along with rotary encoders to feed to the robotposeekf. For IMU, I am using phidgets_imu and imufiltermadgwick to get the sensor_msgs/Imu message. Now the problem is although the robot is stationary, the linear acceleration values in X and Y direction are not zero, both from phidgetsimu and imufilter_madgwick. See below:

---
header: 
  seq: 12245
  stamp: 
    secs: 1435257458
    nsecs: 927698280
  frame_id: base_link
orientation: 
  x: 0.0
  y: 0.0
  z: 0.0
  w: 0.0
orientation_covariance: [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
angular_velocity: 
  x: -0.00118647482551
  y: -0.000878773278379
  z: 0.0012882275209
angular_velocity_covariance: [1.2184696791468346e-07, 0.0, 0.0, 0.0, 1.2184696791468346e-07, 0.0, 0.0, 0.0, 1.2184696791468346e-07]
linear_acceleration: 
  x: 0.301461312895
  y: -0.283705212135
  z: 9.65745491309
linear_acceleration_covariance: [8.66124974095918e-06, 0.0, 0.0, 0.0, 8.66124974095918e-06, 0.0, 0.0, 0.0, 8.66124974095918e-06]
---

and I guess because of this there is lot of drift in the IMU (not sure). So, I would like to know why is this happening? Shouldn't the acceleration in X and Y direction be zero if the robot is stationary? Please let me know what is wrong here and how can it be corrected or am I misinterpreting something?

Thanks in advance.
Naman Kumar

Asked by Naman on 2015-06-25 13:42:28 UTC

Comments

Answers

Real word sensor data is always noisy (to varying extend. This is also (very) true for low cost IMUs. Noisy acceleration measurements are perfectly normal and you'll have to live with those if you do not want to spend ten thousands of dollars on a tactical grade IMU.

Asked by Stefan Kohlbrecher on 2015-06-26 07:17:38 UTC

Comments