Classification and tracking of moving objects for automotive radar
With the development of science and technology, automotive driving has become a study of concern. Various types of sensors have been installed on self-driving vehicles to monitor the surrounding environment. Among them, the high adaptability to different weather conditions and low cost of radar sens...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
2017
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/71157 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | With the development of science and technology, automotive driving has become a study of concern. Various types of sensors have been installed on self-driving vehicles to monitor the surrounding environment. Among them, the high adaptability to different weather conditions and low cost of radar sensor make it one of the most popular types of sensors. Specifically, the 77 GHz automotive radar can be applied as both Long Range Radar (LRR) and Short Range Radar (LRR). The radar sensor is expected to achieve better situation awareness for autonomous vehicles or drivers by detecting the obstacles in the environment.
In this project, 77 GHz radar sensors are used to collect data from real scenarios where vehicles and pedestrians can be detected. The challenge is to extract meaningful information from the raw radar data. The raw radar data only contains discrete data points representing detected targets. Therefore, the display of radar data, clustering of data, tracking of moving objects, and classification of moving objects are implemented step by step to achieve awareness of road situations. The radar data is displayed dynamically in frames. The clustering algorithm is based on DBSCAN clustering method, whereas the tracking algorithm is based on Kalman Filter. In the end, the classification of objects is realized by studying the features of clusters of different categories of road users. |
|---|