Energy-efficient wireless communication for urban physical sensing

Motivation

Sensing and actuation are at the core of smarter cities as they allow the connection of the physical artifacts of our cities with the digital world. As a matter of fact, various sensor & actuator networks are being deployed in cities, as for instance illustrated by systems for lighting control, intelligent transportation or smart metering, to name a few. The Internet of Things vision as well as that of smarter cities are further raising the prospects of open networking environments where the various specialized networks can be interconnected and leveraged for different applications. However, the composed networks have been designed with respect to their target application domains, in silos (vertical approach). In the urban scenario, the heterogeneity of the technologies and the particularity of the urban services bring up new network-dimensioning challenges. The network optimization has to be extended to the inter-technological perspective and to the multi-services standpoint. There is a need to break these silos and to make different applications to access every infrastructure node even if it is was not initially foretold for this. The different technologies that compose the resulting “capillary network” have to inter-operate in a seamless and optimal way so that they can provide user-centric services with the desired quality of experience. Consider, for instance, dimensioning the scheduling mechanism of a mesh network, which has to carry the traffic generated by different Wireless Sensor Networks (WSNs) in the city. Predicting the time and spatial distribution of the traffic generated by the different WSNs are clearly among the key elements that shall be considered. At the same time, from a downlink standpoint, the judicious setting of a WSN aggregation mechanism must be considered according to the time varying capacity of the mesh backbone level. Further, the heterogeneity of the composed networks can conveniently be leveraged to enhance the overall networking performance. An illustrative example is the coupling of sensors and RFID tags. Indeed, an RFID reader may be a sensor in a wireless sensor network and data hold by RFID tags and collected by readers may need to be reported to a sink. This will allow new applications and possibilities such as the localization of a tagged object in an environment covered by sensors.

Goals

Building upon the research of the FUN and URBANET teams, the objective of this effort is to study advanced solutions for city-scale physical sensing that in particular effectively integrate the various wireless technologies. The main goal is to dynamically adapt wireless communications to changes in traffic, environment and applications requirements to offer the best QoS and energy consumption. Indeed, applications that access the devices may not have the same QoS and traffic requirements and can change along time. Some resources may be more or less busy along the day or having more or less energy. All these parameters have to be integrated in the adaptive network self-organization and data routing.

State of the art and challenges

The capillary networks paradigm unifies the wealth of wireless connectivity available in the urban environment [Urbanet12]. However, while the paradigm covers a large panel of technologies and network architectures, common challenges remain. Our record of research on spontaneous and multi-hop networks let us think that autonomic networking appears as the main issue: the connectivity to Internet, to cyber-physical systems, and to Information Systems should be transparent for the user, context-aware and location-aware [Urbanet12].

Urban capillary networks have a number of specific properties: distributed and localized topologies, very high node degree, dynamic network diameter, unstable / asymmetric / non-transitive radio links, concurrent topologies, heterogeneous capabilities (in terms of storage, computing, energy, etc.), etc. These properties are particularly challenging for the communication layers and Quality of Service (QoS) support and should be mitigated by exploiting the structures of the data collected and the devices capacities with regards to applications requirements and expectations.

In this regard, self-configuration, self-organization and self-healing are some of the major concerns within the context of capillary networks. These mechanisms have the capacity to alleviate the deployment and the management of the network, and can provide efficient support to the network dynamics.

In architectures where self-* mechanisms govern the protocol design, both robustness and energy are more than ever essential challenges at the network layer. Energy-harvesting solutions [Mitwol14] can significantly increase the network lifetime but raise new challenges to smartly adapt the protocol behavior to the current energy level of distributed devices. Integrating such approaches within the design of the protocols themselves [Ducrocq13, Mitton13, Radak11] is still new but has shown to be a promising approach.

Because of the urban context, but also because of the wireless media itself, network connectivity is always temporary, while applications require a delivery ratio close to 100% [Mouradian13]. QoS is a major challenge [Gaillard14]. Robust coexistence and cooperation mechanisms among heterogeneous networks and technologies have to be investigated.

One way to cope with these challenges is to leverage the nature and spatial and temporal relevance of the collected data. Data aggregation has a major role to play in data gathering with network robustness issues. In particular, combining data prediction [Ghaddar12], energy prediction [Jumira12], local aggregation and measurement redundancy for improving on data reliability is a promising idea, which can also be important for energy saving purposes [Cui14].

Methodology

The different capillary network technologies have been until now individually studied by the different research communities, without considering the impact for the city and its citizens. We believe that the study of the capillary network paradigm needs to follow three distinct, but tightly connected steps, as outlined below.

  • Understanding network characteristics: Understanding the properties, the constraints and the requirements of capillary networks is essential. We argue that experimental data should go further than classical closed environment laboratory test beds, and need to rely on actual deployment in the studied context. Experimentations should help us to measure the differences between theoretical models and experiment results and to correct our models to make them better fit the reality.
  • Designing capillary solutions: It is essential to go beyond generic, theoretical solutions that failed to result in any major breakthrough in the last decade of wireless networking. We propose to focus on specific issues raised by the urban environment and design our solutions with these specific constraints in mind. These solutions, resulting from a design phase aware of urban specificities, should be applicable to different wireless technologies within the capillary network. We have to keep in mind at every step the applications requirements and the environment and hardware characteristics, in addition to relying on strong theoretical analysis designed in the first step and enriched by the environment impact observed in experimentations.
  • Evaluating solutions: The performance evaluation step plays an important double role. First of all, it allows a feedback for the design phase, either validating the proposed solutions, or uncovering problems initially not taken into account. This feedback loop can be more rapid, when the evaluation is made using a mathematical model or a simulation study, or slower when the analysis goes through an actual deployment. However, in the former case the feedback can be biased by assumptions needed for analytical or simulation purposes. We believe both the faster and the slower loops have their advantages, and we propose to use both of them in the evaluation phase in a complementary way. Moreover, the experimental assessment of the proposed solutions also has a second role, as it provides us with real data issued from the urban environment, data that can be used as an input for the characterization phase. Also, second rounds of loops may integrate additional features in their design itself, potentially issued from other tasks like privacy and security concerns.

References

[Cui14] J. Cui, F. Valois, Data aggregation in Wireless Sensor Networks: Compressing or Forecasting? In Proceedings of IEEE WCNC, 2014.

[Ducrocq13] T. Ducrocq, M. Hauspie, N. Mitton. Balancing energy consumption in clustered wireless sensor networks. ISRN Sensor Networks, Hindawi, 2013.

[Gaillard14] G. Gaillard, D. Barthel, F. Theoleyre, F. Valois, Service Level Agreements for Wireless Sensor Networks: a WSN Operator’s Point of View. In Proceedings of IFIP/IEEE NOMS, 2014.

[Ghaddar12] A. Ghaddar, T. Razafindralambo, I. Simplot-Ryl, D. Simplot-Ryl, S. Tawbi, A. Hijazi. Investigating Data Similarity and Estimation Through Spatio-Temporal Correlation to Enhance Energy Efficiency in WSNs. Adhoc & Sensor Wireless Networks, Old City Publishing, 16 (4). 2012.

[Jumira12] O. Jumira, R. Wolhuter, N. MittonPrediction Model For Solar Energy Harvesting Wireless Sensors.  Fourth International IEEE EAI Conference on e-Infrastructure and e-Services for Developing Countries (Africomm). 2012.

[Mitton13] N. Mitton, E. Natalizio, R. Wolhuter. Beacon-less mobility assisted energy efficient georouting in energy harvesting actuator and sensor networks. The 12th International Conference on Ad Hoc Networks and Wireless (ADHOC-NOW). 2013.

[Mitwol14] N. Mitton, R. Wolhuter. Energy Harvesting in Wireless Sensor Networks. Rechargeable Sensor Networks: Technology, Theory, and Application Introducing Energy Harvesting to Sensor Networks. 2014.

[Mouradian13] A. Mouradian, I. Augé-Blum. On the Reliability of Wireless Sensor Networks Communications. In Proceedings of Adhoc-Now, 2013.

[Radak11] J. Radak, N. Mitton, D. Simplot-Ryl. Using Battery Level as Metric for Graph Planarization.  10th International Conference on Ad Hoc Networks and Wireless (AdHocNow). 2011.

[Urbanet12] I. Augé-Blum, K. Boussetta, H. Rivano, R. Stanica, F. Valois, Capillary Networks: A Novel Networking Paradigm for Urban Environments. In Proceedings of ACM CoNext workshop. 2012.

Permanent link to this article: https://citylab.inria.fr/energy-efficient-wireless-communication/