Monte Carlo Localization Algorithm

The Monte Carlo Localization (MCL) algorithm is used to estimate the position and orientation of a robot. During the process, we need to determine the number of beams employed for computation of likelihood function. Usually, 300 ldar beams are used as the measurement. Then, it will easily becomes 0 even with 3 or 4 beams. So, in practice, how many bemas do you use for this process?

Answer 1

[Shresth Sharma]

Hi,

It is my understanding that you are using Monte Carlo Localization algorithm and you are trying to determine the number of beams required for computation of the likelihood function. Now for MATLAB the computation of likelihood uses 60 as default value for ‘NumBeams’. But it is suggested for computational efficiency of the likelihood function the number of beams should be specified less than the number available by the sensor.

For more information on the likelihood function and information on the Monte Carlo Algorithm you can go through the following documentation links:

https://www.mathworks.com/help/nav/ug/monte-carlo-localization-algorithm.html

https://www.mathworks.com/help/nav/ref/montecarlolocalization-system-object.html

https://www.mathworks.com/help/nav/ref/likelihoodfieldsensormodel.html