Abstract
Research and/or Engineering Questions/Objective
As the vehicle safety gets higher interest, the automotive radar system become more significant. For the case when a human unexpectedly gets into the driving road, to prevent a car accident, a variety of targets (e.g., humans, vehicles, animals, and etc.) should be properly classified by the radar. To make it possible, it is essential to know the differences in the radar signals reflected from the humans and the cars. In other words, the received signals show different characteristics depending on the targets. Therefore, in this paper, using the properties of the signals reflected from the human and the vehicles, the target classification is conducted.
Methodology
First, in our proposed method, we measure the received signals that are reflected from the human subjects and the vehicles in the automotive 77 GHz frequency modulated continuous wave (FMCW) radar. Since it is difficult to apply the conventional RCS equation in the high frequency FMCW radar system, with the measured data, we define a new parameter that can appropriately represent the radar cross section (RCS) value. Then, based on the newly proposed RCS parameter, some unique features are extracted, and they are utilized to distinguish the human subjects from the vehicles.
Results
With the parameter calculated from the measured data and some signal features derived from the parameter, we can classify the human subjects and the vehicles into two different groups. The estimated RCS values of the vehicles are larger than those of the human subjects. In addition, the variation of the RCS values is much more fluctuated in the result of the human subjects.
Limitations of this study
An important limitation of this study is that it needs a vast amount of measured data for the human subjects and the vehicles. Since the human subjects have various body shapes and the body of the vehicles also different from each other, the received signals can be varied a lot in some cases. Therefore, to get much higher estimation accuracy, lots of measurement data are required.
What does the paper offer that is new in the field including in comparison to other work by the authors?
Until recently, there exist rare studies about the target classification in the 77 GHz FMCW radar system. Moreover, the measurement of the RCS in the radar system using this frequency band is not well presented in the other studies.
Conclusions
In this paper, we propose the human-vehicle classification algorithm in the 77 GHz automotive FMCW radar system. Based on the features extracted from the newly defined RCS parameter, we can get the desirable classification accuracy.
Key Words : Frequency modulated continuous wave (FMCW) radar; Radar cross section (RCS); Target classification