Feature Extraction for Human-Vehicle Classification in FMCW Radar

Rizik, Ali; Randazzo, Andrea; Vio, R.; Delucchi, Alessandro; Chiblé, Hussein; Caviglia, Daniele D.

doi:10.1109/icecs46596.2019.8965072

Cited by 10 publications

(11 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Radar systems can be used for detecting and classifying different targets, such as pedestrians and vehicles [3][4][5]. Indeed, such systems produce, through a proper antenna, an electromagnetic wave that propagates to the objects eventually located in the inspected scenario.…”

Section: Pedestrian-vehicle Classification Reviewmentioning

confidence: 99%

“…The setup considered in the present paper includes a "Distance2Go" radar demo board developed by Infineon technologies, which is able to produce range-Doppler maps by performing a double fast Fourier transform (FFT) on the raw data measured using a frequency-modulated continuous wave (FMCW) scheme [5,10]. Such maps are characterized by a peak in correspondence to the frequency shift due to the propagation delay and to the Doppler effect (which is always present when dealing with moving targets).…”

Section: Pedestrian-vehicle Classification Reviewmentioning

confidence: 99%

“…From these measurements, it is possible to derive features to be used for machine learning classification. Specifically, the machine learning features used for training are the extension of the range and velocity profiles, as well as the standard deviation, mean, and variance for the same variables, in addition to the radar cross-section and the estimated target velocity [5]. However, in vehicle classification, there is the problem of vehicles moving crosswise (i.e., along a direction perpendicular to the radar axis), which can be mistaken for pedestrians [11].…”

Section: Pedestrian-vehicle Classification Reviewmentioning

confidence: 99%

“…This fact makes them ideal, in addition to classical closed-circuit television (CCTV) systems [2], for operating round-the-clock automatic surveillance in an urban environment. Radars [3][4][5] can be used for vehicle and pedestrian classification by relying on feature extraction from the range and Doppler profiles of each target. The data collected by radar measurements can be used as input to machine learning (ML) algorithms for classification.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Pedestrian and Multi-Class Vehicle Classification in Radar Systems Using Rulex Software on the Raspberry Pi

et al. 2020

Self Cite

View full text Add to dashboard Cite

Nowadays, cities can be perceived as increasingly dangerous places. Usually, CCTV is one of the main technologies used in a modern security system. However, poor light situations or bad weather conditions (rain, fog, etc.) limit the detection capabilities of image-based systems. Microwave radar detection systems can be an answer to this limitation and take advantage of the results obtained by low-cost technologies for the automotive market. Transportation by car may be dangerous, and every year car accidents lead to the fatalities of many individuals. Humans require automated assistance when driving through detecting and correctly classifying approaching vehicles and, more importantly, pedestrians. In this paper, we present the application of machine learning to data collected by a 24 GHz short-range radar for urban classification. The training and testing take place on a Raspberry Pi as an edge computing node operating in a client/server arrangement. The software of choice is Rulex, a high-performance machine learning package controlled through a remote interface. Forecasts with a varying number of classes were performed with one, two, or three classes for vehicles and one for humans. Furthermore, we applied a single forecast for all four classes, as well as cascading forecasts in a tree-like structure while varying algorithms, cascading the block order, setting class weights, and varying the data splitting ratio for each forecast to improve prediction accuracy. In the experiments carried out for the validation of the presented approach, an accuracy of up to 100% for human classification and 96.67% for vehicles, in general, was obtained. Vehicle sub-classes were predicted with 90.63% accuracy for motorcycles and 77.34% accuracy for both cars and trucks.

show abstract

Section: Pedestrian-vehicle Classification Reviewmentioning

confidence: 99%

Section: Pedestrian-vehicle Classification Reviewmentioning

confidence: 99%

Section: Pedestrian-vehicle Classification Reviewmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Pedestrian and Multi-Class Vehicle Classification in Radar Systems Using Rulex Software on the Raspberry Pi

et al. 2020

Self Cite

View full text Add to dashboard Cite

show abstract

“…Xu et al improved the range and Doppler velocity resolution assuming the micro-Doppler analysis in the line-of-sight (LOS) scenario [10]. The study [11] explored differences between Doppler signatures resulting from the motion of two different observation objects, the human body, and the vehicle, with a 24 GHz radar. In [12], pedestrian body movements and bicycle rotation-related micro-Doppler frequency shifts with significant variations were used to distinguish pedestrians and bicyclists from other objects on the road.…”

Section: Introductionmentioning

confidence: 99%

Diffraction Signal-Based Human Recognition in Non-Line-of-Sight (NLOS) Situation for Millimeter Wave Radar

Terashima

Yamada

et al. 2021

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

In driver assistance or self-driving systems, millimeter-wave radar is an indispensable sensing tool because of its applicability to all weather conditions or non-line-of-sight (NLOS) sensing. This study focuses on a human recognition issue in the NLOS scenario by applying the support vector machine (SVM)-based machine learning approach to a diffraction signal. We show that there is a significant difference in diffraction signals between man-made objects (e.g., metallic cylinder and human body) even without motion. Hence, by exploiting such difference, SVM achieves a high recognition rate, even in deeply NLOS situations. The experimental investigation, using a 24 GHz millimeter-wave radar in an anechoic chamber demonstrates that a diffraction signal based recognition accurately classifies the real human and human mimicking manmade object, even in the NLOS scenario shielded by the parking vehicle.

show abstract