Track-Before-Detect approach on LIDAR signal processing for low SNR target detection

In recent years, LIDAR sensor has been getting higher interests as one of the prospective sensors for future intelligent vehicles. In order to enable advanced applications on variety of road environment, it becomes more important to detect various types of objects from further distance. In this article, we focus on LIDAR signal processing and propose an algorithm to detect low SNR target. It is based on Track-Before-Detect technique in Bayesian probabilistic framework and derive point measurements from the raw data. Fundamental evaluation in comparison to constant thresholding with single observation has shown significant improvement by the proposed detector which enables to produce further point measurements.