[CI4R-MULTI3]

Multi-Frequency RF Sensor Network Human Activity Dataset

• Details about radars utilized:
  
  • ​This dataset is acquired in a synched fashion using the following co-located radars:
    • XeThru (10 GHz UWB impulse radar)
    • Ancortek SDR Kit (24 GHz FMCW radar with 1500 MHz bandwidth)
    • TI IWR1443 BOOST (77 GHz FMCW radar with 4 GHz bandwidth)
      
• # Classes:   11

                        WLKT (Walking towards radar)                    WALKA (Walking away from radar)
                        PICK (Pick up object from ground)             BEND (Bending)
                        SIT (Sitting on chair)                                    KNEEL (Kneeling)
                        CRWL (Crawling towards radar)                  LIMP (Limping with right leg stiff)
                        WTOES (Walking on both toes)                  SHTEPS (Walking with short steps)
                        SCSSR (Scissors gait)

• # Samples/class: ~60
  
• GitHub Link: https://github.com/ci4r/CI4R-Activity-Recognition-datasets
  
• Publications to be cited:
  • ​S. Z. Gurbuz, M. M. Rahman, E. Kurtoglu, T. Macks, and F. Fioranelli “Cross-frequency training with adversarial learning for radar micro-Doppler signature classification (Rising Researcher)”, Proc. SPIE 11408, Radar Sensor Technology XXIV, 114080A (11 May 2020)