With the increasing intelligence of human society, people's demand for human-computer interaction is also increasing. As an important communication medium for human to express information, gesture has always been an important topic in human-computer interaction. Using gesture recognition technology in the vehicle environment can reduce the operation difficulty during driving, reduce the possibility of driver distraction, and greatly improve driving safety and driving experience. Millimeter wave radar can effectively protect the privacy in the car from being leaked, and can still work normally in the dark interior environment. Moreover, with the development of millimeter wave technology from 24g to 60g and 77g, the improvement of its resolution further improves its ability to detect small displacement. Therefore, the gesture recognition technology using millimeter wave radar has been developed. In this paper, 3DCNN and series LSTM network structures are designed, and 3DCNN network, 3DCNN-LSTM and 3DCNN-ECA-LSTM networks are used for training and verification respectively. The experiment shows that the average recognition accuracy of 3DCNN-ECA-LSTM is the highest, up to 98.2%, which is about 5% higher than that of traditional 3DCNN, and the ECA3D module improves the accuracy of the whole network by 2%. Then 3DCNN-ECA-LSTM is used to take range - Doppler and horizontal angle - range image as the first group of inputs, range- Doppler and pitch angle - range image as the second group of inputs, and range Doppler and horizontal angle - range map and pitch angle -range map as the third group of inputs. It is analyzed that the average recognition rate of the network integrating the three features is more than 4% higher than that of the other two networks.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Research on Gesture Recognition Algorithm Based on Millimeter-Wave Radar in Vehicle Scene


    Additional title:

    Sae Technical Papers


    Contributors:
    Bai, Jie (author) / Huang, LiBo (author) / Xiong, Min (author) / Zhu, XiChan (author) / Wang, Xiuqing (author)

    Conference:

    SAE 2021 Intelligent and Connected Vehicles Symposium Part II ; 2021



    Publication date :

    2022-03-31




    Type of media :

    Conference paper


    Type of material :

    Print


    Language :

    English




    Research on Gesture Recognition Algorithm Based on Millimeter-Wave Radar in Vehicle Scene

    Wang, Xiuqing / Bai, Jie / Zhu, XiChan et al. | British Library Conference Proceedings | 2022


    Micro Gesture Recognition of the Millimeter-Wave Radar Based on Multi-branch Residual Neural Network

    Weng, Caien / Li, Dehai / Zhao, Guiquan et al. | SAE Technical Papers | 2022


    Road-Shape Recognition Using On-Vehicle Millimeter-Wave Radar

    Miyake, Y. / Natsume, K. / Hoshino, K. et al. | British Library Conference Proceedings | 2007


    Road-Shape Recognition Using On-Vehicle Millimeter-wave Radar

    Miyake, Yasuyuki / Natsume, Kazuma / Hoshino, Koichi | IEEE | 2007


    Research on Tracking Algorithm for Forward Target-Vehicle Using Millimeter-Wave Radar

    Song, Shiping / Wu, Jian / Yang, Yu et al. | British Library Conference Proceedings | 2020