12 May 2009

Goals

• Testing electronic chaining and leashing with a single UA
• The SNR measurements from this flight were used offline to test a UKF that attempts to localize the MNR.
• The radio parameters, power (k0) and decay constant (alpha), were found from a different flight day, 05-20-2009 by using best fit.

Flight Path (blue) with MNR Estimate (red)

The red points near the runway represent the initial estimate of the location of the MNR. The true position is the black square, while the final estimate is the black star. Note that the initial position was chosen randomly on the field with no apriori knowledge. It's position estimate was over away from the true position, yet the filter still did a reasonable job of converging.

Parameter Esimation with Truth

The next 2 plots show the convergence of the estimates of the position and the radio parameters. The "true" radio parameters were found by using the best fit from this day.

Lat-lon residuals with error bars.

The final plot shows the Lat and Lon residuals with error bars. Note that the filter satys consistent for most of the run, except the end. This may have been caused by the trajectory flown. The last portion of the flight was ~500m North of the target, and may have biased the estimate in that direction. Note that before this biasing occured the error in position was below 10m, which is quite good concerning the noise in the data.