C12328: Method for Magnetometer and Accelerometer CalibrationNovelty:
A novel technique for estimating sensor bias in magnetometers and accelerometers using angular rate data obtained from a gyroscope.
Value Proposition:
Magnetometers, accelerometers, and gyroscopes are used together in a number of applications, including vehicle navigation systems and personal electronics devices, such as smart phones and tablet devices. The combination of these sensors allows for the estimation of the full attitude (heading, pitch, and roll) of a device; however, all of these sensors are affected by biases, scale factors, and non-orthogonality. Existing methods of device calibration require 360 degree movement of the device in one or several degrees of freedom in order to estimate accelerometer and magnetometer bias from gyroscope data and defined start and end positions; these methods are computation-intensive and require users to manually change device attitude to calculate bias and recalibrate. This method allows for the convenient and accurate estimation of the bias of magnetometers and accelerometers, eliminating the need for user intervention. Additional advantages of this method include:
• Does not require numerical integration of gyroscope data or defined start and end positions
• Can be calculated continuously
• Requires less processing time and can be easily implemented
Technical Details:
Johns Hopkins researchers have developed a novel methodology for correcting magnetometer and accelerometer bias. This method solves for unknown sensor bias within the inertial reference frame using the sensor measurement and the inertial reference measurement within the same inertial reference frame. The proposed solution relates the sensor measurement rate with the angular rate measured by the gyroscope. This solution can be implemented using least squares estimation, adaptive identification or Kalman filter. Notably, this method does not require numerical differentiation of sensor measurements, allowing for accurate, real-time estimation of sensor bias.
Looking for Partners:
To develop and commercialize this technology with vehicle navigation systems and with handheld devices requiring motion capture features.
Stage of Development:
Mathematical proof of concept
Data Availability:
Prototype
Publications/Associated Cases:
Robotics, 2013
