Web Services Description and Discovery for Mobile Crowdsensing: Survey and Future Guidelines

Introduction

Fueled by the widespread adoption of smartphones as powerful measuring devices, a novel class of mobile Internet of Things applications fall under the category of mobile crowdsensing. By leveraging people mobility and smartphone integrated sensors (GPS, Velocity, Gyroscope, Air quality, Microphone, etc.), this new sensing paradigm allows to continuously monitoring phenomena. In fact, sensed and shared data between mobile and computing devices are processed by urban applications in order to deduce noise level, traffic information, people density and more.

Intrinsically, the mobile crowdsensing allows accessing information that are inherently out of the spatial and/or technological scoop. Information out of the spatial scoop are those related to specific regions or locations different from the consumer place. Citing the example of asthmatic people who would monitor air quality before moving to other places or city planners that would deduce noise level by leveraging smartphones’ microphones and people mobility. Information out of the technological scoop are those that cannot be deduced alone due to technological problem such as smartphone capabilities limitation (e.g. lack of air quality sensor).

However, mobile crowdsensing collected data arise several challenges related to coverage quality, incentive mechanisms, data management and more. In this paper, we pay particular attention to collected data presentation standardization through Web services in an interoperable platform-independent infrastructure. In fact, Web services present an important direction to encapsulate widely detected data in order to be accessed through WebAPI (Sheng et al., 2013). Thus, as shown the Figure 1, we investigate mobile crowdsensing services description, discovery and reasoning techniques from one hand and their immediate provision from different and heterogeneous mobile devices in energy efficient manner from the other hand.

Figure 1.

Mobile crowdsensing studied challenges

Especially in mobile crowdsensing, service discovery among the network is important for enabling devices cooperation. In fact, urban applications are exposed to a large pool of mobile services. Faced with this diversity, discovering services that best meet their requirements is a challenging task. Spatial/temporal correlation between consumers’ requests and providers’ data is needed in most scenarios; therefore not only the provided service data should be highly described but also the context surrounding it. This allows an easier and efficient service discovery required by urban application consumers. Then, faced with the large pool of discovered services, reasoning technique is a salient feature to compare query against offered services and to evaluate system performance in relation to user requirements.

However, as detection tasks and service provision can easily expose participants to a significant drain on intrinsically limited smartphone resources, a rigorous management of their energetic resources is desired and often enforced. This is a challenging task hindered by the intrinsic network heterogeneity. In fact, crowdsensing services delivery and discovery systems best demonstrate their values when integrating heterogeneous systems for exploiting their embedded capabilities regardless of how these systems were built (Mohamed & Wijesekera, 2012).

Recently, the Open Geospatial Consortium (OGC) Sensor Web Enablement (SWE) standards enable heterogeneous sensory readings to be discoverable and accessible via the Web in a well-defined way (Bröring et al., 2011). However, several gaps limit its capabilities to achieve the sensor Web desires. Firstly, SWE does not support the description of complex processing operations within a middleware (Funk et al., 2011). Secondly, it is built upon XML description, which is verbose and creates an overhead for communication establishment. Those facts make the SWE not suitable for low powered devices. More recently, the Machine to Machine (M2M) technology puts this intelligence directly into offered Service Capability Layers (ETSI, 2011), enabling easier and fluent devices cooperation.