Human–vehicle classification using feature‐based SVM in 77‐GHz automotive FMCW radar

Lee, Seongwook; Yoon, Young-Jun; Lee, Jae-Eun; Kim, Seong-Cheol

doi:10.1049/iet-rsn.2017.0126

Cited by 75 publications

(55 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, a 3D human body with a total height of 1.76 m has been implemented. The human body parts have been modelled as described in Section 2 and their sizes, listed in Table 1, have been chosen to be equivalent to the average human body sizes [21][22][23][24]. The distance between the human 3D model and the antennas is 3.4 m, which identifies a near field region considering the chosen dimensions of the body and the frequency range.…”

Section: Rcs Of a Human Standingmentioning

confidence: 99%

“…To this purpose, the RCS of a perfectly conductive sphere has been analyzed at 24 GHz. A diameter of 20 cm has been chosen close to the average size of a human head [21][22][23][24]. A coarse meshing (λ/8) has been used to model the sphere in FEKO TM , representing a good tradeoff between the simulation accuracy and simulation speed.…”

Section: Performance Assessmentmentioning

confidence: 99%

“…Hamdane et al have shown in [20] that the pedestrians are located by the radars less than 20 m from the front of the vehicle. This distance identifies a near field region, considering the high working frequencies and the dimension of the human targets [21][22][23][24], where the PW condition is not verified. It follows that the implementation of the antenna is needed at the numerical modeling stage for the evaluation of the RCS.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Efficient Simulation Tool to Characterize the Radar Cross Section of a Pedestrian in Near Field

Manfredi¹,

Russo²,

Leo³

et al. 2020

PIER C

View full text Add to dashboard Cite

A simulation tool to characterize the radar cross section of a pedestrian in near field is presented in the paper. The tool has been developed in order to predict and optimize the performance of the short-range radar systems employed in autonomous vehicle operations. It is based on an analytical model which joins the modeling of the human body with the theory of the physical optics. Our studies first focused on the implementation of the electromagnetic code where the human body, the radiation properties of the antenna, and the scenario to be analyzed have been analytically expressed. Then, the proposed model has been validated in terms of accuracy comparing simulated, theoretical, and experimental data regarding the radar cross section of metal and lossy spheres and of an adult, in the frequency range 23-28 GHz. In the end, an evaluation of the performance in terms of required computer memory and execution time has been carried out, comparing the proposed simulation tool with other numerical computational methods.

show abstract

Section: Rcs Of a Human Standingmentioning

confidence: 99%

Section: Performance Assessmentmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Efficient Simulation Tool to Characterize the Radar Cross Section of a Pedestrian in Near Field

Manfredi¹,

Russo²,

Leo³

et al. 2020

PIER C

View full text Add to dashboard Cite

show abstract

“…We firstly design and test a binary classifier and a multiclass classifier using the most widely used support vector machine (SVM). Since SVM has been widely applied as a pattern recognition algorithm for detecting human motion by radar [2,5,19,20], we use SVM as a reference algorithm for comparison to our algorithm proposed in this study.…”

Section: Classifier Designmentioning

confidence: 99%

Radar Application: Stacking Multiple Classifiers for Human Walking Detection Using Micro-Doppler Signals

Kwon

Kwak

2019

Applied Sciences

View full text Add to dashboard Cite

We propose a stacking method for ensemble learning to distinguish micro-Doppler signals generated by human walking from background noises using radar sensors. We collected micro-Doppler signals caused by four types of background noise (line of sight (LoS), fan, snow and rain) and additionally considered micro-Doppler signals caused by human walking combined with these four types of background noise. We firstly verified the effectiveness of a fully connected deep neural network (DNN) to classify 8 types of signals. The average accuracy was 88.79% for the test set. Then, we propose a stacking method to combine two base classifiers of different structures. The average accuracy of the stacking method on the test set was 91.43%. Lastly, we designed a modified stacking method to reuse feature information stored at the previous stage and the average test accuracy increased to 95.62%. This result shows that the proposed stacking methods can be an effective approach to improve classifier’s accuracy in recognizing human walking using micro-Doppler signals with background noise.

show abstract

“…In another approach [15] based on a newly defined parameter using the RCS (radar cross-section), three significant features were extracted from the received radar signal and used as classification criteria to identify humans and vehicles. This method can be operated in real time with a simple classification function.…”

Section: Introductionmentioning

confidence: 99%

Doppler-Spectrum Feature-Based Human–Vehicle Classification Scheme Using Machine Learning for an FMCW Radar Sensor

Hyun

Jin

2020

Sensors

View full text Add to dashboard Cite

In this paper, we propose a Doppler-spectrum feature-based human–vehicle classification scheme for an FMCW (frequency-modulated continuous wave) radar sensor. We introduce three novel features referred to as the scattering point count, scattering point difference, and magnitude difference rate features based on the characteristics of the Doppler spectrum in two successive frames. We also use an SVM (support vector machine) and BDT (binary decision tree) for training and validation of the three aforementioned features. We measured the signals using a 24-GHz FMCW radar front-end module and a real-time data acquisition module and extracted three features from a walking human and a moving vehicle in the field. We then repeatedly measured the classification decision rate of the proposed algorithm using the SVM and BDT, finding that the average performance exceeded 99% and 96% for the walking human and the moving vehicle, respectively.

show abstract

Human–vehicle classification using feature‐based SVM in 77‐GHz automotive FMCW radar

Cited by 75 publications

References 19 publications

Efficient Simulation Tool to Characterize the Radar Cross Section of a Pedestrian in Near Field

Efficient Simulation Tool to Characterize the Radar Cross Section of a Pedestrian in Near Field

Radar Application: Stacking Multiple Classifiers for Human Walking Detection Using Micro-Doppler Signals

Doppler-Spectrum Feature-Based Human–Vehicle Classification Scheme Using Machine Learning for an FMCW Radar Sensor

Contact Info

Product

Resources

About