Presentation is loading. Please wait.

Presentation is loading. Please wait.

Zach Zira & Kevin Roy ISE 533

Similar presentations


Presentation on theme: "Zach Zira & Kevin Roy ISE 533"— Presentation transcript:

1 Zach Zira & Kevin Roy ISE 533
GPS Comparison Study Zach Zira & Kevin Roy ISE 533

2 Overview About GPS The Experiment Our Hypothesis The Data ANOVA
Normal Plots Residual Plots Main Effects Conclusion 11/12/2018

3 About GPS More than 24 GPS satellites are currently orbiting the earth
Traveling 8,700 mph (3.9 km/sec) at an average height of 12,500 miles (20,200 km) Guaranteed to have signals of at least four satellites (at any location at any time) Circulation time of 12 h sidereal time = 11 h 58 min earth time There is a lot more other different types of satellites (communication, weather, etc.) 11/12/2018 **

4 GPS Locations Day 2 Day 1 11:56 12:00 Day 3 Day 4 11:52 11:48
11/12/2018

5 GPS Error Sources: Atmospheric effects Shifts in the satellite orbits
Errors of the satellite clocks Multipath effect Rounding Describe how the coordinates move Describe why this contributed to picking this topic 11/12/2018 **most to least significant

6 The Experiment Set three GPS units next to each other
A+, GSAT, and TomTom Mkii Log the output via Bluetooth Repeat (1 & 2) 11 hrs 58 minutes later Conduct the process for a total of 3 times GPS devices don’t interfere with each other 11/12/2018

7

8 Our Hypothesis Using Latin Square design Statistical Model:
Row effect (α): GPS Device Column effect (β): Time Treatment effect (τ): Day Statistical Model: yijk = µ + α i + τj + βk + εijk Hypothesis: 9 hypothesis tests (3 for lat, 3 for long, 3 for number of satellites) 11/12/2018

9 Data 11/12/2018 Column Response 1 2 Row A+ day 1 day 2 day 3 Gsat
Column Response 1 2 Row A+ day 1 day 2 day 3 Gsat TomTom Time Latitude 1 2 Device A+ Gsat TomTom Time NOS 1 2 Device A+ 11 9 Gsat 8 TomTom Time Longitude 1 2 Device A+ Gsat TomTom 11/12/2018

10 ANOVA ← The device is significant when observing longitude
General Linear Model: Long versus Device, Time, Day Factor Type Levels Values Device fixed , 2, 3 Time fixed , 1, 2 Day fixed , 2, 3 Analysis of Variance for Long, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P Device Time Day Error Total S = R-Sq = 98.10% R-Sq(adj) = 92.41% ← The device is significant when observing longitude General Linear Model: Sat versus Device, Time, Day Factor Type Levels Values Device fixed , 2, 3 Time fixed , 1, 2 Day fixed , 2, 3 Analysis of Variance for Sat, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P Device Time Day Error Total S = R-Sq = 86.54% R-Sq(adj) = 46.15% Nothing is significant when → observing number of satellites General Linear Model: Lat versus Device, Time, Day Factor Type Levels Values Device fixed , 2, 3 Time fixed , 1, 2 Day fixed , 2, 3 Analysis of Variance for Lat, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P Device Time Day Error Total S = R-Sq = 86.89% R-Sq(adj) = 47.57% ← Nothing is significant when observing latitude 11/12/2018

11 Normal Plots Questionable normal probability plots
All normal probability plots passed equal variance test 11/12/2018

12 Residual Plots

13 Main Effects With additional number of satellite data, we could have done a regression analysis Determine if there is correlation between the number of satellites and the variance of the locations. 11/12/2018

14 Conclusion Significant Not significant
Longitude vs Device (p= 0.024) Moderately significant: Longitude vs Day (p= 0.115) Not significant All other factors were not significant From our analysis, the GPS did not behave as predicted 11/12/2018

15 Questions? 11/12/2018

16 Wavier Please do not use any of the information in this presentation or in the term paper for publication without written consent from both Zach Zira and Kevin Roy


Download ppt "Zach Zira & Kevin Roy ISE 533"

Similar presentations


Ads by Google