Scholarlink

<p>In this work, a new methodology for signal analysis is presented, based on simple mathematical properties of signals. The proposed analysis employs a scanning method with a specified embedding dimension, a common technique in Information Theory. The main hypothesis of this work is that a signal classifier can be constructed using solely geometric concepts. In line with this approach, the model incorporates three simple geometric measures: amplitude, along with two newly defined concepts, Zenith angle and shape factor, which capture variations in terms of the signal’s peaks and valleys. For each segment of a signal, these measures generate a 3-tuple. By eliminating redundant information, a feature space is constructed from these resulting 3-tuples, referred to as feature vectors. Subsequently, the cardinality of this feature space is used as a parameter for signal differentiation. The methodology was tested on noise signals and dynamic chaotic systems, demonstrating its effectiveness in distinguishing between the analyzed signals.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An Innovative Method to Distinguish Chaos from Noise in the Time Domain

  • 期刊论文

  • 发表日期：

    2025年8月4日

查看全文
  • Hernán Manuel García Blesa,
  • Juan Vorobioff,
  • Walter Legnani

摘要

In this work, a new methodology for signal analysis is presented, based on simple mathematical properties of signals. The proposed analysis employs a scanning method with a specified embedding dimension, a common technique in Information Theory. The main hypothesis of this work is that a signal classifier can be constructed using solely geometric concepts. In line with this approach, the model incorporates three simple geometric measures: amplitude, along with two newly defined concepts, Zenith angle and shape factor, which capture variations in terms of the signal’s peaks and valleys. For each segment of a signal, these measures generate a 3-tuple. By eliminating redundant information, a feature space is constructed from these resulting 3-tuples, referred to as feature vectors. Subsequently, the cardinality of this feature space is used as a parameter for signal differentiation. The methodology was tested on noise signals and dynamic chaotic systems, demonstrating its effectiveness in distinguishing between the analyzed signals.