CE 506 Homework Fall 2004

• homework 1. read chapters 1 & 3. see the notes on regression as presented in class. do a similar least squares regression for the 4 points: (x,y) = (1,3), (3,4), (4,5), (7,5). make a plot of the original points and the fitted line. repeat this with 5 points, by including point 3, (4,5), twice. how did the fitted line change with the inclusion of point 3 twice? be sure to show all of your work, the model elements, the estimated parameters, the residuals, the adjusted observations, any matlab code that you used, and the plots. due tuesday, 7 september. solution
• homework 2. read chapter 4 (skip the nonlinear parts for now) 3 problems. please do them by the method indicated. it is highly recommended that you use the matrix approach and make a matlab script. be sure to include your analysis of the problem and all of the condition equations. important results are the residuals, the parameters (if any), and the adjusted observations. summarize these things at the beginning, then include any matlab code and output as supporting information.
• homework 3. finish reading chapter 4. problem set to wrap up linear models. due tuesday, 28 september.
• homework 4. read chapter 10. 2 problems due tuesday, 19 october.
• homework 5. read either of the designated history books and address the appropriate questions located here. assigned 13 october, due monday, 6 december.
• homework 6. prepare general 2D network adjustment as per details located here. you have ~2 weeks to prepare the program and confirm that it works with the sample data. then on friday, 19 november i will distribute a new network. you must submit results on tuesday 23 november. (corrected!) solution to example 1 here also, see the geomatics drive for "function.zip" with helpful matrix functions. see listing of functions also, in case it helps, here is a matlab (.mat) file that contains my B,W, and f matrices on the first iteration. this is before any column elimination. the parameter order is z1,...,z6,x1,y1,...,x6,y6. data files
• homework 7. get the data files. for the circle fit you have initial approximations for X0,Y0, and R. observations are in X,Y. use general least squares. make global test at 0.05 level of significance. sketch 50% confidence ellipse for the center point. use 0.3 as a priori sigma for the coordinate obs. for the gps pseudorange data, see variable "data", the columns are sat. #, time, Xs (km), Ys (km), Zs (km), pseudorange (m), sat. clock offset (usec). refine the pseudorange by adding c*DT to it. constant is given for c. initial approximations for receiver position (wgs84 geocentric, m) in Xr, Yr, Zr. initial approximation for dt, the receiver clock offset, is given in nanoseconds (E-09). for position unknowns use kilometers, for clock bias unknown use nanoseconds. make global test at alpha=0.05. use 20m as a priori sigma for the pseudoranges. due friday, 10 dec. see zipped matlab & output files for the solutions to these two problems. output is in c.lst and g.lst, main matlab files are cir.m and gpsnav.m