EKF-SLAM: Fix prediction step of motion

[m-dobler]

Fixes the computation of the predicted covariances after a motion command, by changing elementwise multiplication to matrix multiplication.

[coveralls]

Pull Request Test Coverage Report for Build 1096

• 2 of 2 (100.0%) changed or added relevant lines in 1 file are covered.
• 6 unchanged lines in 2 files lost coverage.
• Overall coverage increased (+0.03%) to 90.305%

---

Files with Coverage Reduction	New Missed Lines	%
SLAM/GraphBasedSLAM/graph_based_slam.py	1	98.32%
PathPlanning/BatchInformedRRTStar/batch_informed_rrtstar.py	5	81.41%

Totals
Change from base Build 1094:	0.03%
Covered Lines:	6725
Relevant Lines:	7447

---

💛 - Coveralls

[AtsushiSakai]

Fix the bug and comment is welcome. But, could you keep the code before simplifying? . Because, this repo's aim is for learning algorithm. So the code should be same as algorithm. In this simulation case, it is possible to simplify the code, but in other simulation, it might be impossible.

[AtsushiSakai]

@m-dobler Thank you!! Good catch : ).