A comparison of vehicular trajectory encoding techniques

Abstract

The transmission of vehicular trajectory information is one basic building block of car-to-car communication. Frequently, this information is transmitted as raw data, i. e., as a sequence of location measurements. In this paper we argue that, due to the laws of physics and the requirement to follow a road, vehicular mobility has very specific characteristics. Hence, vehicular trajectory information can be compressed very efficiently using domain-specific lossy compression schemes. We discuss and compare three promising approaches that can be used to this end: linear approximation, cubic splines, and clothoids.