Sensor error calibration and optimal geometry analysis of calibrators.

The target localization performance is not only affected by the measurement noise, but also related with the sensor error such as sensor position error and measurement biases. If not properly estimated and calibrated, sensor error may cause large localization error or ghost tracks. This paper investigates the sensor error calibration problem with imperfect calibrators whose positions are not accurately known when both the sensor position error and measurement biases exist. The performance loss without considering the sensor error is analyzed by evaluating the mean square error (MSE). A closed-form solution based on weighted least squares (WLS) is proposed to estimate the sensor measurement bias and update the sensor position by using the biased range and angle measurement about calibrators. Moreover, the optimal geometry for calibration objects is analyzed and the optimum criterions for updating sensor position are proposed by minimizing the trace of the CRLB. The optimal calibrator-sensor geometries in the presence of measurement bias are generally different from the optimal calibrator-sensor geometries in the absence of measurement bias. Finally the performance of the proposed solution is verified and the optimal calibrator-sensor geometry results for 2 or 3 sensors are solved efficiently by using Nelder–Mead simplex method in simulation. • Formulation of sensor error calibration problem using imperfect calibrators. • Theoretical analysis of the performance loss without considering the sensor errors. • Constructing closed-form solution based on WLS to calibrate sensor errors. • Formulation and analysis of the optimal geometry for calibration objects. • Demonstrating performance improvement of the proposed solution and propositions. [ABSTRACT FROM AUTHOR]