Dictionary-based map compression using modified RANSAC map-matching

Obtaining a compact representation of a large-size map built by mapper robots is a critical issue for information sharing in robotic sensor networks. In the paper, this map compression problem is explored from a novel perspective of lossless data compression. The primary contribution of the paper is the proposal of an efficient dictionary-based map compression approach employing RANSAC map-matching techniques. We also present several efficient schemes by using preemptive RANSAC as well as adaptive sample size. Experiments show promising results in terms of map compression ratio as well as computational efficiency.