Behavior Forests: Real-Time Discovery of Dynamic Behavior for Data Selection

TL;DR

This paper introduces Behavior Forests, a real-time data selection framework for automated driving systems that identifies relevant dynamic behaviors in vehicle data, reducing data volume while preserving important information.

Contribution

It presents a novel Behavior Forest framework that constructs a Behavior Graph during vehicle operation, enabling adaptive, analytical pattern discovery without pre-training.

Findings

Discards 96.01% of data while retaining relevant behaviors

Effective on synthetic, ECG, and automotive datasets

Provides analytical descriptions of discovered patterns

Abstract

Automated Driving Systems (ADS) development relies on utilizing real-world vehicle data. The volume of data generated by modern vehicles presents transmission, storage, and computational challenges. Focusing on Dynamic Behavior (DB) offers a promising approach to distinguish relevant from irrelevant information for ADS functionalities, thereby reducing data. Time series pattern recognition is beneficial for this task as it can analyze the temporal context of vehicle driving behavior. However, existing state-of-the-art methods often lack the adaptability to identify variable-length patterns or provide analytical descriptions of discovered patterns. This contribution proposes a Behavior Forest framework for real-time data selection by constructing a Behavior Graph during vehicle operation, facilitating analytical descriptions without pre-training. The method demonstrates its performance using a synthetically generated and electrocardiogram data set. An automotive time series data set is used to evaluate the data reduction capabilities, in which this method discarded 96.01% of the incoming data stream, while relevant DB remain included.