Practical classification of different moving targets using automotive radar and deep neural networks

• DOI: 10.1049/iet-rsn.2018.0103
• Publication Date: 2018-04-10T02:22:16Z
• Authors: Aleksandar Angelov; Andrew Robertson; Roderick Murray‐S...; Francesco Fioranelli
• Keywords: 0202 electrical engineering, electronic engineering, information engineering; 02 engineering and technology

In this work, the authors present results for classification of different classes targets (car, single and multiple people, bicycle) using automotive radar data neural networks. A fast implementation algorithms detection, tracking, micro‐Doppler extraction is proposed in conjunction with transceiver TEF810X microcontroller unit SR32R274 manufactured by NXP Semiconductors. Three types networks are considered, namely a classic convolutional network, residual combination recurrent problems across four recorded. Considerable accuracy (close to 100% some cases) low latency pre‐processing prior (∼0.55 s produce 0.5 long spectrogram) demonstrated study, possible shortcomings outstanding issues discussed.