Efficient Protocol to Use FMCW Radar and CNN to Distinguish Micro-Doppler Signatures of Multiple Drones and Birds

Range-Doppler images are widely used to classify different types of Unmanned Air Vehicles (UAVs) because each UAV has a unique range-Doppler signature. However, a UAV's range-Doppler signature depends on its movement mechanism. This is why a classifier's accuracy would be degraded if the effect of the mechanical control system of UAVs wasn't taken into consideration, which may lead to a non-unique signature of a UAV while in-flight. In this paper, a full-wave electromagnetic CAD tool is used to investigate the effect of the control systems of two quadcopters, a hexacopter, and a helicopter UAVs on their range-Doppler signatures. A Mechanical Control-Based Machine Learning (MCML) algorithm is introduced to classify the four UAVs. Different Machine Learning (ML) algorithms were applied to the generated datasets that considered the mechanical control information of UAVs. The Convolutional Neural Networks (CNN) algorithms provided robust performance reaching an accuracy of higher than 90%.

show abstract

“…12. CNNs are powerful algorithms that are frequently used for radar targets classification [34], [35]. Fig.…”

Section: Classification Resultsmentioning

confidence: 99%

Machine Learning for UAV Classification Employing Mechanical Control Information

Sayed

Ramahi

Shaker

2024

IEEE Trans. Aerosp. Electron. Syst.

View full text Add to dashboard Cite

show abstract

“…Convolutional Neural Network (CNN) is one of the powerful algorithms that are used for radar targets classification [29], [30]. The DopplerNet algorithm [29] is used to classify the quadcopter and hexacopter drones, a Maxpooling layer is added to this algorithm to decrease its complexity to avoid over-fitting.…”

Section: Resultsmentioning

confidence: 99%

Machine Learning for UAV Classification Employing Mechanical Control Information

Sayed¹,

Ramahi²,

Shaker³

2022

Preprint

View full text Add to dashboard Cite

<p>Range-Doppler images are widely used to classify different types of UAVs because each UAV has a unique range-doppler signature. However, a drone's range-doppler signature depends on its movement mechanism. This is why the classifier accuracy would be degraded if the effect of the mechanical control system wasn't taken into consideration, which may lead to a non-unique signature of a drone while in-flight. In this paper, a full-wave electromagnetic CAD tool is used to investigate the effect of the control systems of a quadcopter and a hexacopter UAVs on their range-doppler signatures. A Mechanical Control-Based Machine Learning (MCML) algorithm is introduced to classify the two UAVs and its accuracy is found to exceed 90%.</p>

show abstract

“…A Convolutional Neural Network (CNN) is one of the powerful algorithms used to classify radar targets [9], [24]. The DopplerNet algorithm [9] is used to classify the five different types of drones, a Maxpooling layer is added to this algorithm to decrease its complexity to avoid overfitting.…”

Section: Classificationmentioning

confidence: 99%

UAV Classification using Neural Networks and CAD-generated Radar Datasets

Sayed¹,

Riad²,

Ramahi³

et al. 2022

Preprint

View full text Add to dashboard Cite

<p>Using micro-doppler signatures is an effective way to classify different types of UAVs, as well as other airborne objects such as birds. To generate signatures for drones, radar measurements are needed; however, these measurements are limited to the types of available drones, the radar parameters, the targets’ range, and the environments in which these measurements are conducted. In this paper, a new method for generating signature datasets is introduced. The method uses full-wave electromagnetic simulation software. Using this method, radar drones’ datasets can be generated using different types, sizes, drone materials, radar parameters, detected range, targets speed, and rotor RPM for rotary drones. A 77 GHz modeled FMCW radar is used to create dataset for classification purposes. Finally, a Convolutional Neural Network (CNN) algorithm is used to classify five types of drones. Based on the results, the classification of the drones is found to exceed 97% accuracy. </p>

show abstract

Efficient Protocol to Use FMCW Radar and CNN to Distinguish Micro-Doppler Signatures of Multiple Drones and Birds

Cited by 9 publications

References 24 publications

Machine Learning for UAV Classification Employing Mechanical Control Information

Machine Learning for UAV Classification Employing Mechanical Control Information

Machine Learning for UAV Classification Employing Mechanical Control Information

UAV Classification using Neural Networks and CAD-generated Radar Datasets

Contact Info

Product

Resources

About