flâneur

Application of Maximum Likelihood Estimation in GraphSLAM | by Shiva Chandrachary | Medium

shivachandrachary.medium.com · 1,083 words · saved by 1 readers

At the core of GraphSLAM is graph optimization — the process of minimizing the error present in all of the constraints in the graph. Let’s take a look at what these constraints look like, and learn to apply a principle called maximum likelihood estimation (MLE) to structure and solve the optimization problem for the graph. The likelihood is a complementary principle to probability. While probability tries to estimate the outcome given the parameters, the likelihood tries to estimate the parameters that best explain the outcome. For example, Probability: What is the probability of rolling a 2 on a 6-sided die? Answer: 1/6 Likelihood: I’ve rolled a die 100 times, and a 2 was rolled 10% of the time, how many sides does my die have? Answer: 10 sides When applied to SLAM, likelihood tries to estimate the most likely configuration of state and feature locations given the motion and measurement observations. Let’s look at a very simple example — one where our robot is taking repeated measurem

Application of Maximum Likelihood Estimation in GraphSLAM Shiva Chandrachary 6 min read · Jan 11, 2021 -- 1 Listen Share At the core of GraphSLAM is graph optimization — the process of minimizing the error present in all of the constraints in the graph. Let’s take a look at what these constraints look like, and learn to apply a principle called maximum likelihood estimation (MLE) to structure and solve the optimization problem for the graph. Likelihood The likelihood is a complementary principle to probability. While probability tries to estimate the outcome given the parameters, the likelihoo

related reading