Yi Ding

Assistant Professor at UT Dallas

Heterogeneity is problem that widely met in research and applications. It is said in [SenSys12, IODetector] that “since the antenna gain may vary across different mobile phone models, it is hard to accurately map different RSS values to different environments.” It is also said in [TMC14] that “environmental changes impact the transmission channel between devices”. Similar conclusions can also been found in [SenSys12, CrowdMon], [iiWAS16].

Taxonomy

  • Heterogeneity types: device, operating system, data density, user, environment,
  • Solution types: deep learning, transfer learning, fingerprinting, clustering, interpolation, feature selection, adversarial network
  • Application: localization, device-free HAR

Comments

  • A common cons is limited heterogeneity
  • No real world application is conducted no feedback from real world applications.

Special Topics

  • Heterogeneity in Mobile Sensing Applications, UbiComp13
    • Taxonomy: Device
  • Modellet: Experiencing and Handling the Diversity, MobiCom 14
    • Taxonomy: Data density, environment, fingerprinting, localization
    • Cons:
      • Limited heterogeneity (13 deployment venues)
      • Device diversity is not considered.
  • Smart Devices are Different, SenSys 15
    • Taxonomy: device, clustering+interpolation
    • Pros: Very solid experiments.
    • Cons:
      • Limited heterogeneity (36 devices, 9 users);
      • Environment diversity is not considered (actually only accelerator related heterogeneities are considered).
  • DeepSense, WWW17
    • Taxonomy: Device, User, Deep learning
    • Solution: CNN + RNN (GRU)
    • Pros: Solid experiments,
    • Cons:
      • Limited heterogeneity (20 mobile phones, 9 users, 6 activities)
      • Mainly consider the heterogeneity of noisy measurement.
  • CrossSense: Cross-Site and Large-Scale WiFi Sensing, MobiCom18-CrossSense
    • Taxonomy: user, environment, transfer learning,
    • Pros: Solid experiments and rich results
    • Cons:
      • Limited heterogeneity (100 users, 40 gestures, 1.2 million wireless activity samples)
      • Training user needed;
      • Training user needs to do the same thing in both training and deployment environment;
  • Towards Environment Independent Recognition, MobiCom18-EI
    • Taxonomy: user, environment, deep learning, adversarial network, device-free HAR
  • Widar3.0: Zero-Effort Cross-Domain Gesture Recognition with Wi-Fi, MobiSys19
    • Taxonomy: user, environment, Feature selection
    • Pro: Find features independent of data domain
    • Cons: Device not considered.

Ref.

[SenSys12-CrowdMon] Pushp, S., Min, C., Lee, Y., Liu, C. H., & Song, J. (2012, November). Towards crowd-aware sensing platform for metropolitan environments. In SenSys (pp. 335-336). ACM.

[SenSys12-IODetector] Zhou, P., Zheng, Y., Li, Z., Li, M., & Shen, G. (2012, November). IODetector: A generic service for indoor outdoor detection. In SenSys (pp. 113-126). ACM.

[UbiComp13] Blunck, H., Bouvin, N. O., Franke, T., Grønbæk, K., Kjaergaard, M. B., Lukowicz, P., & Wüstenberg, M. (2013, September). On heterogeneity in mobile sensing applications aiming at representative data collection. In UbiComp (pp. 1087-1098).

[MobCom14] Li, L., Shen, G., Zhao, C., Moscibroda, T., Lin, J. H., & Zhao, F. (2014, September). Experiencing and handling the diversity in data density and environmental locality in an indoor positioning service. In MobiCom (pp. 459-470). ACM.

[TMC14] Sigg, S., Scholz, M., Shi, S., Ji, Y., & Beigl, M. (2014). RF-sensing of activities from non-cooperative subjects in device-free recognition systems using ambient and local signals. IEEE TMC, 13(4), 907-920.

[SenSys15] Stisen, A., Blunck, H., Bhattacharya, S., Prentow, T. S., Kjærgaard, M. B., Dey, A., … & Jensen, M. M. (2015, November). Smart devices are different: Assessing and mitigatingmobile sensing heterogeneities for activity recognition. In SenSys (pp. 127-140). ACM.

[iiWAS16] Inomoto, H., Saiki, S., Nakamura, M., & Matsumoto, S. (2016, November). Mission-oriented large-scale environment sensing based on analogy of military system. In iiWAS (pp. 414-421). ACM.

[WWW17] Yao, S., Hu, S., Zhao, Y., Zhang, A., & Abdelzaher, T. (2017, April). Deepsense: A unified deep learning framework for time-series mobile sensing data processing. In WWW (pp. 351-360). International World Wide Web Conferences Steering Committee.

[MobiCom18-CrossSense] Zhang, J., Tang, Z., Li, M., Fang, D., Nurmi, P., & Wang, Z. (2018, October). CrossSense: Towards Cross-Site and Large-Scale WiFi Sensing. In MobiCom (pp. 305-320). ACM.

[MobiCom18-EI] Jiang, W., Miao, C., Ma, F., Yao, S., Wang, Y., Yuan, Y., … & Xu, W. (2018, October). Towards environment independent device free human activity recognition. In MobiCom (pp. 289-304). ACM.

[MobiSys19] Zheng, Y., Zhang, Y., Qian, K., Zhang, G., Liu, Y., Wu, C., & Yang, Z. (2019, June). Zero-Effort Cross-Domain Gesture Recognition with Wi-Fi. In MobiSys (pp. 313-325). ACM.