O'Reilly logo

Bayesian Estimation and Tracking: A Practical Guide by Anton J. Haug

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

20.6 Performance Comparison Analysis

The performance analysis to be presented below was conducted using the synthetic test data described in Section 20.4. Note that this is the same synthetic trajectory (with identical noise in the image data) as was used in Section 19.6. The performance of the sensor fusion estimation filter (UKF(2)) is compared against two estimators that operate only on photogrammetric data (NLLSQ and UKF(2)) and also against two estimators that use only IMU data (Predictor–Corrector and UKF(2)). The performance metric used in all cases is the root mean squared (RMS) error. As we will show below, the sensor fusion estimator had the lowest RMS errors and was therefore the best estimator.

Figures 20.2 through 20.4 show the estimated track components of a single run of the sensor fusion filter against the truth track components. Figure 20.5 shows a close-up of a small section of the lateral position truth trajectory with the estimated trajectories form multiple estimators superimposed over the truth. Identical noisy observations and initialization were used as inputs to all estimation methods. Analysis of these figures leads to a number of observations.

Figure 20.2 Translational and rotational position using the sensor fusion estimator.

img

Figure 20.3 Translational and rotational velocity using the sensor fusion estimator.

Figure 20.4 Translational and rotational ...

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, interactive tutorials, and more.

Start Free Trial

No credit card required