Generating Synthetic Short-Range FMCW Range-Doppler Maps Using Generative Adversarial Networks and Deep Convolutional Autoencoders
| Autor: | Marcio L. Lima de Oliveira, Marco J. G. Bekooij |
|---|---|
| Přispěvatelé: | Digital Society Institute, Computer Architecture Design and Test for Embedded Systems |
| Jazyk: | angličtina |
| Rok vydání: | 2020 |
| Předmět: |
Computer science
0211 other engineering and technologies Neural Network 02 engineering and technology Doppler-Range 01 natural sciences law.invention Synthetic Data Deep Learning law Radar imaging Radar 021101 geological & geomatics engineering Generative Adversarial Networks Noise measurement Artificial neural network business.industry Deep learning 010401 analytical chemistry Detector 22/2 OA procedure Pattern recognition Autoencoder 0104 chemical sciences FMCW Artificial intelligence Convolutional business |
| Zdroj: | 2020 IEEE Radar Conference (RadarConf20) |
| Popis: | In this paper, we discuss the usage of Generative Adversarial Networks (GANs) and Deep Convolutional Autoen-coders (CAE) for creating synthetic Range-Doppler (RD) maps of Frequency-Modulated Continuous-Wave (FMCW) radars for a short-range situation with moving objects, based on measured RD maps of pedestrians and cyclists. Instead of using regular mathematical functions or heavy radar simulations, we have used an Artificial Neural Network (ANN) model to generate new data. By using our synthetic data, we can automatically have ground-truth data without the need for manual labor; easily create large synthetic datasets; hardly use much computational power after training. To evaluate our method, we have trained a detector system with just synthetic data, and it was capable of detecting moving objects correctly, on actual Range-Doppler maps, 11.6% better than when using a small dataset. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|