✳flâneur — a map of the web's best reading

An application of Numerical Solution to Maximum Likelihood Estimation in GraphSLAM | by Shiva Chandrachary | Medium

shivachandrachary.medium.com · 1,018 words · saved by 1 readers

In the previous post, we looked at a robot taking repeated measurements of the same feature in the environment. This example demonstrated the fundamentals of maximum likelihood estimation(MLE) but was very limited since it was only estimating one parameter — z1​. Below we will walk through a more complicated 1-dimensional estimation problem. The robot starts at an arbitrary location that will be labeled 0, and then proceeds to measure a feature in front of it — the sensor reads that the feature is 7 meters away. The resultant graph is shown in the image below. After taking its first measurement, the following Gaussian distribution describes the robot’s most likely location. Recall that since we constrained the robot’s initial location to 0, x_0x0​ can actually be removed from the equation. Next, the robot moves forward by what it records to be 10 meters and takes another measurement of the same feature. This time, the feature is read to be 4 meters behind the robot. The resultant graph

An application of Numerical Solution to Maximum Likelihood Estimation in GraphSLAM Shiva Chandrachary 5 min read · Jan 11, 2021 -- Listen Share In the previous post , we looked at a robot taking repeated measurements of the same feature in the environment. This example demonstrated the fundamentals of maximum likelihood estimation(MLE) but was very limited since it was only estimating one parameter — z 1​. Below we will walk through a more complicated 1-dimensional estimation problem. Motion and Measurement Example The robot starts at an arbitrary location that will be labeled 0, and then proc

Explore this link on the map →

related reading