Skip to main content

An experimental characterization of reservoir computing in ambient assisted living applications

Abstract

In this paper, we present an introduction and critical experimental evaluation of a reservoir computing (RC) approach for ambient assisted living (AAL) applications. Such an empirical analysis jointly addresses the issues of efficiency, by analyzing different system configurations toward the embedding into computationally constrained wireless sensor devices, and of efficacy, by analyzing the predictive performance on real-world applications. First, the approach is assessed on a validation scheme where training, validation and test data are sampled in homogeneous ambient conditions, i.e., from the same set of rooms. Then, it is introduced an external test set involving a new setting, i.e., a novel ambient, which was not available in the first phase of model training and validation. The specific test-bed considered in the paper allows us to investigate the capability of the RC approach to discriminate among user movement trajectories from received signal strength indicator sensor signals. This capability can be exploited in various AAL applications targeted at learning user indoor habits, such as in the proposed indoor movement forecasting task. Such a joint analysis of the efficiency/efficacy trade-off provides novel insight in the concrete successful exploitation of RC for AAL tasks and for their distributed implementation into wireless sensor networks.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6

Notes

  1. This paper, based on ex-novo conducted experiments, extends the preliminary investigations presented in the conference papers [18], proposing a comprehensive vision of the costs that can practically occur for real deployment of the RC models into mote devices. In particular, previous works did not include an exhaustive comparative analysis of different network sizes and configurations [4, 18], or the adoption of a reduced reservoir weight-encoding scheme [4, 18] or the external performance evaluation [18].

  2. Here, we refer to the general usage of the leaking parameter a in the RC literature. Note, however, that Eq. 1 represents a reduced form of the complete leaky integrator dynamics [22, 24, 30, 41]. In this regard, the characterization of Eq. 1 is that of a low-pass filter with smoothing parameter a.

  3. Crossbow Technology Inc., http://www.xbow.com.

  4. Note that for this task, learning has the general aim of distinguishing among the different types of user trajectories (straight or curved in our scenario), based only on the history of noisy RSS signals. On the other hand, the specific application presented in this paper, consisting in the prediction of the user localization (room change or not), represents a concrete example of real-life RC application in the field of AAL.

  5. http://wnlab.isti.cnr.it/paolo/index.php/dataset/6rooms.

  6. Note that the choice of the connectivity, leaking rate and spectral radius values is not critical for the task. The experiments in this section use the same values identified in previous preliminary works, e.g., Gallicchio et al. [18].

  7. Note that the results reported for the non-local 2 and non-local 3 settings are averaged over all the possible configurations of the available anchors (i.e., 6 configurations for non-local 2 and 4 configurations for non-local 3).

  8. http://wnlab.isti.cnr.it/paolo/index.php/dataset/6rooms.

References

  1. AAL (2009) Ambient assisted living roadmap. http://www.aaliance.eu/public/documents/aaliance-roadmap/

  2. Antonelo EA, Schrauwen B, Campenhout JMV (2007) Generative modeling of autonomous robots and their environments using reservoir computing. Neural Proc Lett 26(3):233–249

    Article  Google Scholar 

  3. Antonelo EA, Schrauwen B, Stroobandt D (2008) Event detection and localization for small mobile robots using reservoir computing. Neural Netw 21(6):862–871

    Article  Google Scholar 

  4. Bacciu D, Gallicchio C, Micheli A, Chessa S, Barsocchi P (2011) Predicting user movements in heterogeneous indoor environments by reservoir computing. In: Bhatt M, Guesgen HW, Augusto JC (eds) Proceedings of the IJCAI workshop on space, time and ambient intelligence (STAMI) 2011, pp 1–6

  5. Bacciu D, Chessa S, Gallicchio C, Lenzi A, Micheli A, Pelagatti S (2012) A general purpose distributed learning model for robotic ecologies. In: Robot Control—10th IFAC symposium on robot control, vol 10(1), pp 435–440. doi:10.3182/20120905-3-HR-2030.00178

  6. Bahl P, Padmanabhan V (2000) RADAR: an in-building Rf-based user location and tracking system. In: Proceedings of INFOCOM 2000. 19th Annual joint conference of the IEEE Comp. and Comm. Soc. vol 2, pp 775–784. doi:10.1109/INFOCOM.2000.832252

  7. Barsocchi P, Lenzi S, Chessa S, Giunta G (2009) A novel approach to indoor RSSI localization by automatic calibration of the wireless propagation model. In: IEEE 69th vehicular Tech. Conf., pp 1–5. doi:10.1109/VETECS.2009.5073315

  8. Barsocchi P, Lenzi S, Chessa S, Giunta G (2009) Virtual calibration for RSSI-based indoor localization with IEEE 802.15.4. In: 2009 IEEE international conference on Comm. ICC ’09, pp 1–5. doi:10.1109/ICC.2009.5199566

  9. Battiti R, Villani A, Le Nhat T (2002) Neural network models for intelligent networks: deriving the location from signal patterns. In: Proceedings of AINS2002, Citeseer

  10. Boedecker J, Obst O, Mayer N, Asada M (2009) Initialization and self-organized optimization of recurrent neural network connectivity. HFSP J 3(5):340–349

    Article  Google Scholar 

  11. Brunato M, Battiti R (2005) Statistical learning theory for location fingerprinting in wireless LANs. Comput Netw ISDN Syst 47(6):825–845. doi:10.1016/j.comnet.2004.09.004

  12. Büsing L, Schrauwen B, Legenstein R (2009) Connectivity, dynamics, and memory in reservoir computing with binary and analog neurons. Neural Comput Appl 22(5):1272–1311

    Article  Google Scholar 

  13. Chang M, Terzis A, Bonnet P (2009) Mote-based online anomaly detection using echo state networks. In: Krishnamachari B, Suri S, Heinzelman W, Mitra U (eds) Distributed computing in sensor systems, lecture notes in computer science, vol 5516. Springer, Berlin, pp 72–86

  14. EvAAL (2011) The 1st evaal competition, evaluating aal systems through competitive benchmarking. Special theme on Indoor Localization and Tracking, AAL Forum proceedings, valencia, ES

  15. Fette G, Eggert J (2005) Short term memory and pattern matching with simple echo state networks. In: Proceedings of the international conference on artificial neural networks (ICANN), Springer, Lecture Notes in Computer Science, vol 3696, pp 13–18

  16. Gallicchio C, Micheli A (2010) A markovian characterization of redundancy in echo state networks by PCA. In: Proceedings of the European symposium on artificial neural networks (ESANN) 2010, d-side, pp 321–326

  17. Gallicchio C, Micheli A (2011) Architectural and markovian factors of echo state networks. Neural Netw 24(5):440–456. doi:10.1016/j.neunet.2011.02.002

    Google Scholar 

  18. Gallicchio C, Micheli A, Barsocchi P, Chessa S (2012) User movements forecasting by reservoir computing using signal streams produced by mote-class sensors. In: Mobile lightweight wireless systems (Mobilight 2011), lecture notes of the institute for computer sciences, social informatics and telecommunications engineering, vol 81. Springer, Berlin, pp 151–168. doi:10.1007/978-3-642-29479-2_12

  19. Gärtner T (2003) A survey of kernels for structured data. SIGKDD Explor Newsl 5:49–58

    Article  Google Scholar 

  20. Hajnal M, Lorincz A (2006) Critical echo state networks. In: Proceedings of the international conference on artificial neural networks (ICANN) 2006, pp 658–667

  21. Hartland C, Bredeche N (2007) Using echo state networks for robot navigation behavior acquisition. In: IEEE international conference on robotics and biomimetics, IEEE, pp 201–206

  22. Jaeger H (2001) The “echo state” approach to analysing and training recurrent neural networks. Tech. rep., GMD—German National Research Institute for Computer Science

  23. Jaeger H, Haas H (2004) Harnessing nonlinearity: predicting chaotic systems and saving energy in wireless communication. Sci Agric 304(5667):78–80

    Article  Google Scholar 

  24. Jaeger H, Lukosevicius M, Popovici D, Siewert U (2007) Optimization and applications of echo state networks with leaky-integrator neurons. Neural Netw 20(3):335–352

    Article  MATH  Google Scholar 

  25. Kolen J, Kremer S (eds) (2001) A field guide to dynamical recurrent networks. IEEE Press, Parsippany

    Google Scholar 

  26. Kontkanen P, Myllymaki P, Roos T, Tirri H, Valtonen K, Wettig H (2004) Topics in probabilistic location estimation in wireless networks. In: Personal, indoor and mobile radio communications, 2004. PIMRC 2004. 15th IEEE International Symposium on, vol 2, pp 1052–1056. doi:10.1109/PIMRC.2004.1373859

  27. Kushki A, Plataniotis K, Venetsanopoulos AN (2007) Kernel-based positioning in wireless local area networks. IEEE Trans Mobile Comp 6(6):689–705. doi:10.1109/TMC.2007.1017

    Google Scholar 

  28. Legenstein RA, Maass W (2007) Edge of chaos and prediction of computational performance for neural circuit models. Neural Netw 20(3):323–334

    Article  MATH  Google Scholar 

  29. Liu H, Darabi H, Banerjee P, Liu J (2007) Survey of wireless indoor positioning techniques and systems. Systems, man, and cybernetics, part C: applications and reviews. IEEE Trans 37(6):1067–1080. doi:10.1109/TSMCC.2007.905750

    Google Scholar 

  30. Lukosevicius M, Jaeger H (2009) Reservoir computing approaches to recurrent neural network training. Comput Sci Rev 3(3):127–149

    Article  Google Scholar 

  31. Luo X, OBrien WJ, Julien CL (2011) Comparative evaluation of received signal-strength index (RSSI) based indoor localization techniques for construction jobsites. Adv Eng Inf 25(2):355–363

    Article  Google Scholar 

  32. Maass W, Natschlager T, Markram H (2002) Real-time computing without stable states: a new framework for neural computation based on perturbations. Neural Comput 14(11):2531–2560. doi:10.1162/089976602760407955

    Google Scholar 

  33. Obst O (2009) Distributed fault detection in sensor networks using a recurrent neural network. Tech. rep., CoRR, abs/0906.4154

  34. Oubbati M, Kord B, Palm G (2010) Learning robot-environment interaction using echo state networks. In: Proceedings of the 11th international conference on Simul. of adapt. behavior, Springer-Verlag, SAB’10, pp 501–510

  35. Pan JJ, Kwok J, Yang Q, Chen Y (2006) Multidimensional vector regression for accurate and low-cost location estimation in pervasive computing. IEEE Tran Knowl Data Eng 18(9):1181–1193. doi:10.1109/TKDE.2006.145

    Google Scholar 

  36. Rodan A, Tiňo P (2011) Minimum complexity echo state network. IEEE Trans Neural Netw 22(1):131 –144

    Article  Google Scholar 

  37. Steil JJ (2004) Backpropagation-decorrelation: online recurrent learning with O(N) complexity. In: Proceedings of the IEEE international joint conference on neural networks (IJCNN) 2004, vol 2, pp 843–848

  38. Steil JJ (2006) Online stability of backpropagation-decorrelation recurrent learning. Neurocomputing 69(7–9):642–650

    Article  Google Scholar 

  39. Tiňo P, Dorffner G (2001) Predicting the future of discrete sequences from fractal representations of the past. Mach Learn 45(2):187–217

    Article  MATH  Google Scholar 

  40. Tiňo P, Cernanský M, Benusková L (2004) Markovian architectural bias of recurrent neural networks. IEEE Trans Neural Netw 15(1):6–15

    Article  Google Scholar 

  41. Verstraeten D, Schrauwen B, D’Haene M, Stroobandt D (2007) An experimental unification of reservoir computing methods. Neural Netw 20(3):391–403

    Article  MATH  Google Scholar 

  42. Waegeman T, Antonelo E, Wyffels F, Schrauwen B (2009) Modular reservoir computing networks for imitation learning of multiple robot behaviors. In: 8th IEEE international symposium on comput. intelligence in robotics and autom., IEEE, pp 27–32

  43. Wu CL, Fu LC, Lian FL (2004) Wlan location determination in e-home via support vector classification. In: Networking, sensing and control, 2004 IEEE international conference on, vol 2, pp 1026–1031

  44. Youssef M, Agrawala A (2005) The horus WLAN location determination system. In: MobiSys’05: proceedings of the 3rd international conference on mobile Sys. appl. and services. ACM, New York, pp 205–218. doi:10.1145/1067170.1067193

  45. Zàruba GV, Huber M, Kamangar FA, Chlamtac I (2007) Indoor location tracking using RSSI readings from a single wi-fi access point. Wireless Netw 13:221,235

    Google Scholar 

Download references

Acknowledgments

This work is partially supported by the EU FP7 RUBICON project (http://www.fp7rubicon.eu/), Contract no. 269914.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Claudio Gallicchio.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Bacciu, D., Barsocchi, P., Chessa, S. et al. An experimental characterization of reservoir computing in ambient assisted living applications. Neural Comput & Applic 24, 1451–1464 (2014). https://doi.org/10.1007/s00521-013-1364-4

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00521-013-1364-4

Keywords

  • Ambient assisted living
  • Reservoir computing
  • Wireless sensor networks
  • Indoor user movement forecasting