0
$\begingroup$

blue and red box arise doubtsI am novice in Slam. I want to implement EKF based slam. For that I consider the book PROBABILISTIC ROBOTICS by Sebastian THRUN. The algorithm they derive in this book for EKF SLAM FOR KNOWN CORRESPONDENCE. I have some doubt with this algorithm.

Hope it will be clear after this post.

I post a picture to show my doubts in this algorithm. Blue and red box arise doubts. Unable to understand what did it mean.

In this algorithm line no: 6 they calculate $Q_t$. As per my knowledge $r,p_i$ they are range and bearing measurement. I have a data file which consist of 88000 value for range and bearing for different time step. So can I consider all this value to calculate $Q_t$ matrix?

In this algorithm can anyone explain me line no 9? $\mu(j,x), \mu(j,y), \mu(j,s)$ are they Cartesian co-ordinate of landmarks? $\mu(t,x), \mu(t,y)$ are they Cartesian co-ordinate of Robot Pose? If so, there is no sign of timestamp in $\mu(j,x), \mu(j,y)$ but there is a $t$ associate with $\mu(t,x), \mu(t,y)$. which $t$ it is?

There is written that if landmark $j$ never seen before then only calculate line no: 9.

My range and bearing data file are look like

# Time [s]    Subject #    range [m]    bearing [rad] 
1248272276.038    90      2.148       0.025 
1248272276.727    14      3.387       0.006 
1248272277.773    14      3.309      -0.015 
1248272277.773    90      2.033       0.024 
1248272278.237    14      3.272      -0.023 
1248272278.969    90      1.946       0.040 
1248272279.453    14      3.165      -0.051 
1248272280.199    14      3.118      -0.070 

So on.....

So as per my data file same landmark repeated several time dependent upon Robot movement. How could I manipulate them as per line No. 10

My odometry data file look like that

# Time [s]    forward velocity [m/s]    angular velocity[rad/s] 
 1248272272.841       0.074       0.229 
 1248272272.852       0.074       0.229 
 1248272272.903       0.074       0.229 
 1248272272.964       0.074       0.229 
 1248272272.995       0.074       0.229 
 1248272273.005       0.074       0.230 

So on..

So the odometry data's timestamp are not matched with range bearing timestamp. They are different. So How could I calculate? Because in line no 3 of this algorithm I am using odometer timestamp but in line no. 9 I am using range bearing time stamp but they are different. Then How could I calculate mu(j,x),mu(j,y) from that?

$\endgroup$

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.