Simultaneous Localisation and Map Building Using the Probabilistic Multi-Hypothesis Tracker
Abstract
This report presents an algorithm for efficiently solving the Simultaneous Localisation and Map Building (SLAM) problem. The SLAM problem requires both the dynamic estimation of the sensor location and the tracking of features of interest in the environment using the sensor measurements. The problem is difficult because the unknown sensor and feature locations are coupled through the sensor measurement. It has been shown that under linear Gaussian conditions, a Kalman Filter solution converges to a solution relative to the unknown starting location. However, this approach does not scale well with the number of features in the scene, and is unfeasible for large maps. The algorithm introduced here is based on the Probabilistic Multi-Hypothesis Tracker (PMHT) and exploits a factorisation of the problem to reduce the computational requirements of the Kalman Filter approach. The new algorithm is demonstrated on a benchmark data set recorded in Victoria Park.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 2005
- Accession Number
- ADA432536
Entities
People
- Samuel Davey
Organizations
- Defence Science and Technology Group