Related Experiment Video
Updated: Dec 10, 2025
11:41
Evaluation of an Exclusive Spur Dike U-Turn Design with Radar-Collected Data and Simulation
Published on: February 1, 2020
20.7K
An Adaptive Ellipse Distance Density Peak Fuzzy Clustering Algorithm Based on the Multi-target Traffic Radar
Lin Cao1,2, Xinyi Zhang1,2, Tao Wang1,2
1Key Laboratory of the Ministry of Education for Optoelectronic Measurement Technology and Instrument, Beijing Information Science and Technology University, Beijing 100192, China.
Sensors (Basel, Switzerland)
|September 4, 2020
Summary
A new adaptive ellipse distance density peak fuzzy (AEDDPF) clustering algorithm improves vehicle clustering accuracy in dense traffic radar scenes. This method enhances data structure description and density peak selection for better results.
Area of Science:
- Computer Science
- Artificial Intelligence
- Data Science
Background:
- Clustering accuracy for closely spaced vehicles in multi-target traffic radar scenes is a significant challenge.
- Existing algorithms struggle with the complex data structures generated by radar measurements.
Purpose of the Study:
- To propose a novel clustering algorithm, the adaptive ellipse distance density peak fuzzy (AEDDPF) algorithm, to address low clustering accuracy in dense traffic radar scenarios.
- To enhance the description of radar measurement data structure and improve the selection of density peak points.
Main Methods:
- Replaced Euclidean distance with adaptive ellipse distance for more accurate data structure representation.
- Introduced an adaptive exponential function curve in the decision graph for precise density peak selection.
- Utilized iterative calculations of membership matrix and clustering centers for final clustering results.
Main Results:
- The AEDDPF algorithm demonstrates higher clustering accuracy compared to DBSCAN, k-means, FCM, GK, and Euclid-ADDPF on real measurement datasets in specific scenarios.
- Experimental results validate the algorithm's superior clustering performance in close-range vehicle scenes.
- Analysis indicates the AEDDPF algorithm possesses generalization capabilities for other data types.
Conclusions:
- The proposed AEDDPF algorithm offers a significant improvement in clustering accuracy for multi-target traffic radar scenes, particularly for vehicles with close driving distances.
- The adaptive ellipse distance and enhanced density peak selection effectively handle complex radar data structures.
- AEDDPF shows promise for real-world applications in intelligent transportation systems and beyond.
