Better EKF performance with new state covariance Q approach. Tuned some parameters for EKF only in GeTMeasurementEvents.m

	- Bugfix on estimate calculation, based on weights now and not mean.
	- Same for estimates covariance.
	- Attempted to implement better state covariance Q in Q_dt(dt) lambda
	- Constants changed in GetMeasurementsEvents
	- Moved back to including IMU in measurement equations

	- GetMeasurementEvents now separates tuning for EKF and PF, use arg = 2 for PF
	- Naive approach to bootstrapped PF in task7_PF

- the cam estimate covariance matrix wasn't computed incorrectly. Fixed now.
- the estimates and the covariance matrices R_cam are updated.

- the weights for the heading angle are now 100 (compared to 1000 before).

- State-from-camera algorithm now also returns covariance matrix for every state estimate
- Update the state measurements MAT file with the matrices R

	- Removed dead code etc.
	- Moved EKF to task7_EKF.m where it belongs

- scripts now prints out the tables with the means and the variances of the data from camera provided for the task 5

- used algorithm parameters: l = 1; GN1wGridLS(J, d_term1, d_term2, x0, 20, 0.01);

- update to the state estimation algorithm

- also, made some small cosmetic improvements (for better readability)

- algorithm settings: GN1wGridLS(J, d_term1, d_term2, x0, 10, 0.01); l = 0.

- algorithm call settings: GN1wGridLS(J, d_term1, d_term2, x0, 20, 0.01);