Crowdsourcing using mobile devices, known as mobile crowdsourcing, is a powerful approach incorporating human wisdom into mobile computations to solve problems while exploiting the advantages of mobility and context-awareness. The problems that can be tackled include the use of geographically distributed tasks, and mobile sensing using the collective wisdom of the crowd. However, implementation of mobile crowdsourcing applications has been found to be challenging to users due to the nature of dynamic sensing, crowd engagement with data distribution, and a process of data verification. In this paper, we provide an extensive survey of the literature on mobile crowdsourcing research, highlighting aspects of particular concerns in terms of implementation needs during the development, architectures, and key considerations for their development. We present a taxonomy based on key issues in mobile crowdsourcing and discuss the different approaches applied to these issues. We also provide a critical analysis of some challenges and suggest directions for future work. In particular, with the future Internet-of-Things in view, we generalize the notion of mobile crowdsourcing to thing crowdsourcing, where crowdsourcing can be issued from smart Internet-connected things that need to harness human resources to solve problems.

Currently, many governments are gearing up to promote the development of smart cities in their countries. A smart city is an urban area using different types of sensors to collect data, which will then be used to manage assets and resources efficiently. Through smart technology, the quality of living and performance of urban services are enhanced. Recent works addressed a set of platforms aimed to support the development of smart city applications. It seems that most of them involved dealing with collecting, managing, analyzing, and correlating data to extract new information useful to a city, but they do not integrate a diversified set of services and react to events on the fly. Moreover, the application development facilities provided by them seem to be limited and might even increase the complexity of this task. We propose an event-based architecture with components that meet important requirements for smart city platforms, supporting increased demand for scalability, flexibility, and heterogeneity in event processing. We implement such architecture and data representation models, handling different data formats, and supporting a semantics-based data model. Finally, we discuss the effectiveness of a S mart Event-based Middleware (SEMi) and present empirical results regarding a performance evaluation of SEMi.

The rapid adoption of mobile devices enables capture and transmission of a variety of sensor and user-contributed data, creating a new data collection paradigm and a wide range of services, often termed as mobile crowdsourcing. However, we are still investigating factors contributing towards the effectiveness of such systems. One key factor to succeed in mobile crowdsourcing applications is the incentive mechanism, which motivates people to contribute to a crowdsourcing effort. In this research, we conducted field experiments that compared the effectiveness of non-monetary and monetary incentive mechanisms, using both quantitative and qualitative methods. The focus was on an exploration of how these mechanisms motivate users’ performance in a mobile crowdsourcing environment. In the experiment, we developed a smart zoning application that allows users to share areas of the cities they perceived as safe or unsafe. The results from this study contribute to research in mobile crowdsourcing for urban understanding. Taken together, these results suggest that payment-related incentives can not only engage people’s interest in project participation but also help improve their work performance in terms of productivity and quality. Notably, a fixed-price financial reward mechanism is best suited for short period data collection and achieving data quality. Also, referral incentive mechanisms, if properly designed, have the potential to extend user coverage, both spatially and temporally. These results can be helpful with regard to the formulation of guidelines on how to create and organize effective payment-based incentives for crowd involvement in cities.

Crowdsourcing has become an important problem-solving technique, for both traditional and mobile applications. There have been work on crowdsourcing particular database operations to humans, and work on spatial crowdsourcing or mobile crowdsourcing, where the workers are mobile device users. This paper presents LogicCrowd which extends logic programming, commonly used for knowledge-based applications, with crowdsourcing capabilities by adding operators to connect to crowdsourcing platforms, thereby enabling crowdsourcing via logic programs and crowd-powered knowledge-based applications. In LogicCrowd, we also introduce a novel unification approach called crowd unification that automatically leverages human knowledge for comparisons through the crowdsourcing paradigm. In addition, LogicCrowd is built on a Prolog platform running on Android mobile devices, thereby enabling logic-based spatial and mobile crowdsourcing. Our proposed method for combining rule-based reasoning and crowdsourcing via LogicCrowd programs is demonstrated in a range of scenarios. Because energy is an important consideration for mobile platforms, we also investigate the energy characteristics of the crowdsourcing operators in our LogicCrowd prototype. Our experiments show the relationships between crowdsourcing operations and energy consumption, and illustrate the factors influencing energy consumption when using crowdsourcing.

Email spam and phishing detection is typically evaluated using accuracy-centric metrics under implicitly unconstrained computational settings. However, in practical deployment scenarios—particularly in real-time and resource-constrained environments—models with comparable predictive performance may differ substantially in inference latency and resource usage, directly affecting their operational feasibility. This paper introduces TERA, a deployment-aware evaluation framework that formulates model assessment as a constraint-aware decision problem. Instead of aggregating performance and efficiency into a single objective, TERA treats predictive performance as a feasibility requirement that defines an admissible set of models. Within this feasible region, operational factors such as latency and resource usage are used to differentiate among candidates through structured, multi-dimensional analysis. Experiments on benchmark email datasets show that multiple models achieve comparable detection performance, forming a region of predictive equivalence. Within this region, significant variations in latency and resource consumption are observed, indicating that predictive equivalence does not imply deployment equivalence. These findings demonstrate that accuracy-based evaluation alone may provide limited guidance for deployment-oriented model selection. By explicitly separating feasibility constraints from preference-based trade-offs, TERA enables transparent and deployment-aligned model evaluation. The framework supports consistent comparison and selection among accuracy-comparable models without altering the role of detection effectiveness as a primary requirement, thereby complementing existing evaluation practices with a structured decision-oriented perspective.

Spatial crowdsourcing is a potentially powerful method for incorporating human wisdom into mobile computations to solve problems while exploiting the advantages of mobility and context-awareness. This paper proposes and investigates task assignments and recruitment in iterative spatial crowdsourcing processes to find regions of particular interest among a collection of regions. We consider cases where associations between regions can be exploited to reduce costs and increase efficiency in crowdsourcing. We describe five approaches, incorporated into crowdsourcing algorithms, for reducing the cost (the number of queries required) and increasing the efficiency (reducing the number of rounds of querying required) in using such spatial crowdsourcing. We demonstrate the performance improvements gained using these approaches based on simulation scenarios. The findings show the interplay and relationships among our proposed approaches using a range of metrics including responses, energy consumption, costs, and time usage. These metrics are demonstrated via a range of scenarios, showing that our proposed approaches can lead to improved performance over randomly choosing regions for inquiry.