# How to do localization when landmarks can't be identified?

 1 I'm working on a more realistic robot simulation in preparation for possibly building a robot that does localization. I'm having trouble adapting the measurement probability algorithm from the simple lecture/homework examples to more realistic ones. In the unit3 lecture we measured the distance to each landmark. In the homework 3.5 we measured bearing to each landmark. What I imagine is more realistic for a robot with a rotating distance sensor is to get pairs of bearing, distance measures. Writing code to take a number of bearing and distance measurements to each landmark that is in range is straight forward. I'm having trouble comparing the actual measurement from the robot with each of the particle's sense data. There are two problems 1) coming up with a single goodness measure or probability because angle and distance are very different quantities. My only guess here is that a geometric mean should be used rather than an arithmetic mean. 2) the bigger problem is deciding which angle distance pairs in the actual measurement to compare with which angle distance pairs in the particle measurement. The problem is that the landmarks have no known identity. In the lecture and homework in all cases we knew the identity of the landmark. I don't think that is very realistic at least not for the simple robot I'm thinking of building. For example Robot sense data particle j sense data i angle distance i angle distance 0 0.78 11.0 0 2.79 9 1 2.81 14.5 1 3.15 17 2 3.44 6.2  The robot sensed 3 landmarks and the particle sensed 2. Should I compare Robot and particle measurements as 0 - 0, 1 - 1, 2 - or 0 - , 1 - 0, 2 - 1 or ... Even if a particle senses the same number of landmarks it doesn't make it any easier to figure out how to match each one up. Even in later material like SLAM we assume we can identify the landmarks. Identifying the landmarks seems to be a major problem that isn't covered. In general what is done? landmark identification possibly through some visual pattern matching or statistical correlation between measurements to unidentified landmarks I was thinking of building a robot without a camera to keep it simple but in either case I need more knowledge to get localization working. Any help or pointers would be appreciated. Thanks, -John asked 30 Mar '12, 18:15 JohnSnyders 106●3●10 accept rate: 0%

 0 That is definitely a huge problem that I guess was deemed out of scope for this class. What I have seen in the past is the Expectation-Maximization (EM) algorithm to estimate correspondences for landmarks which are modelled as Gaussians, but I'm sure there are several others. That method is described in prof. Thrun's book! answered 30 Mar '12, 18:39 George Brind... 8.1k●16●52●98 I just looked at the book and it doesn't explicitly refer to the EM algorithm, just "Maximum Likelihood Correspondence". If you are able to have access to it, the information is at page 215 under section "7.5 Estimating Correspondences". (30 Mar '12, 18:49) George Brind... From what I understood, you basically take the correspondences that maximize the probability of all measurements. So you need that "measure of goodness", which can just be the landmark's Gaussian belief representation, and you need to evaluate all measurements against all landmarks in order to be able to maximize that "goodness". Concretely, in the example you provided, I expect that the associations would be between 1-0, 2-1 and robot measurement 0 would indicate either a new landmark or that the particle's pose is unlikely. I am also curious about this problem myself, but I haven't tried to tackle it with any of the proposed solutions yet. (30 Mar '12, 18:57) George Brind...
 0 I have been working on a similar question, but I am planning on using a camera. I described my general approach here. answered 30 Mar '12, 21:32 Michael Bangert 788●5●17●21
Question text:

Markdown Basics

• *italic* or _italic_
• **bold** or __bold__
• image?![alt text](/path/img.jpg "Title")
• numbered list: 1. Foo 2. Bar
• to add a line break simply add two spaces to where you would like the new line to be.
• basic HTML tags are also supported

×5,188
×76