Wikipedia SLAM

Source:  http://en.wikipedia.org/wiki/Simultaneous_localization_and_mapping Parents: SLAM Index , slam-resources

Different types of sensors give rise to different SLAM algorithms whose assumptions are most appropriate to the sensors.

• At one extreme, visual features provide details of many points within an area –> rendering SLAM unnecessary
  • shapes in these point clouds can be easily and unambiguously aligned at each step via image registration .
• At the opposite extreme, tactile sensors are extremely sparse
  • they contain only information about points very close to the agent
  • require strong prior models to compensate in purely tactile SLAM.
• Most practical SLAM tasks fall somewhere between these visual and tactile extremes.

Algorithms

• Statistical techniques
  • Kalman
  • Particle/Monte Carlo
• Set-membership techniques
  • Bundle adjustment
  • Maximum a posteriori estimation (MAP) – SLAM for image data

Given:

• Controls u
• Sensor measurements o
• Time steps t

To estimate:

• Agent’s state x
• Map of environment m

All quantities are probabilistic.

Objective is to compute:

Use Bayes' rule, EM algorithm

Sensor models

• landmark based
• raw data based
  • make no assumption that landmarks can be identified