1. DTW Overview
DTW measures similarity between signals with different time scales by allowing nonlinear alignments.
Unlike transform-based methods that analyze signal components,
DTW directly compares structural patterns in the time domain by locally stretching or compressing the signals.
2. Signal Processing
DTW uses dynamic programming to find the optimal path.
This path minimizes the cumulative Euclidean distance between the two signals.
It first computes a local cost matrix d(i, j), then constructs an accumulated cost matrix D(i, j).
3. Representation
The primary output of DTW is the DTW Distance.
A smaller DTW distance indicates higher similarity between signals, even when their timing or pacing differs.
In this example, the DTW distance is the final accumulated cost “1.7” located at the top-right corner, (i, j)= (8,7).
4. Interpretation
Beyond the distance score,
the resulting alignment path reveals exactly where and how the signals are temporally distorted to achieve the optimal match.
5. Radar Application
In radar systems, DTW is used to compare motion patterns that vary in time.
It enables the recognition of similar behaviors, such as human movement or repetitive actions, even when they occur at different speeds.
Most importantly, it allows for fast and simple implementation while minimizing MCU processing power and resource consumption.
Radar Sensor SoC & Module
WiFi IoT Module
5G/LTE/CAT-M1/NB-IoT
WiSUN/HaLow/Thread Module
