Mobile Participatory Sensing Systems: A Comprehensive Review

Mobile Participatory Sensing Systems have the potential to improve different services through monitoring of the urban landscape using mobile devices based on the collaboration of thousands of mobile users. Many articles have been recently published related to mobile participatory systems where data collected from thousands of mobile users are analyzed to extract vital community information and a spatiotemporal interpretation of the phenomenon of interest is built. The purpose of this paper is to assess the state-of-the-art mobile participatory sensing systems to classify key practical requirements of such systems and related challenges. The Kitchenham method has been used to conduct this review. A selection of 24 articles out of 590 articles related to mobile participatory sensing frameworks have been made between the period of 2013 to 2018 from the IEEE Xplore Digital Library and ACM library for assessment. A detailed review has been conducted through the classification of mobile participatory sensing systems and a critical evaluation is carried out. Potential opportunities and challenges for modern mobile participatory sensing systems are also discussed. This paper provides researchers in the field with a comprehensive and up-to-date review of mobile participatory sensing systems


Introduction
Several participatory systems have emerged in recent years for different domains such as traffic monitoring and air pollution monitoring in urban spaces [1].There are also trustworthiness and reliability are important challenges among others to be considered.
In recent times, mobile phones have certainly shaped various breakthroughs in different domains of socio-economic activities.This modern wave of Internet-enabled, sensor-rich, smart mobile devices such as Apple, Samsung, and many more have unlocked the door for an innovative paradigm known as participatory sensing.Participatory sensing allows monitoring the urban landscape using mobile devices based on the collaboration of thousands of mobile users.Using this paradigm, mobile users gather multi-level statistics information from the external surrounding environment using their mobile devices.Therefore, they can share the data through an existing communication infrastructure.Nowadays most of the mobile phones are equipped with sensors and various applications are dependent on the data being generated by these sensors.This data collection from thousands of mobile users is commonly known as mobile crowdsourcing.The collected data are analyzed to extract vital community information and a spatiotemporal assessment of the phenomenon of interest can be constructed [6,22].Given the widespread of smartphones and the mass of people in the urban area, a participatory sensing application can accomplish an extraordinary level of granularity both in time and space for events of interest in metropolitan spaces [12].
The aim of the paper is to provide a systematic literature review in Mobile participatory sensing systems to classify key practical requirements of such systems and related challenges using the Kitchenham method.It consist of three stages namely, planning, conducting, and reporting.Therefore, it provides researchers in the field with a comprehensive and up-to-date review of mobile participatory sensing systems.The relevance of the paper is to conduct an in-depth investigation of the limitations and benefits of mobile participatory crowdsourcing.A detailed review has been conducted through the classification of mobile participatory sensing systems and a critical evaluation is carried out.This paper is organized as follows, first, the research method used to conduct this study is presented.Secondly, the outcomes are reported after grouping the 24 selected frameworks and thirdly, the result are discussed.Thus, the paper is planned as follows: Section 1 presents the research methodology, based on Kitchenham, which was used to review articles from the IEEE and ACM library.Section 2 presents the inclusion and exclusion criteria.Section 3 is an overview of the steps conducted to perform the review.Section 4 discusses the 24 Mobile Participatory Sensing Systems using five criteria namely accuracy, reliability, realtime, trustworthiness, and scalability.Section 5 presents possible research challenges and opportunities.Last, but not the least, section 6 presents a summary and conclusions of the paper.

Research Method
A systematic detailed literature review has been directed from 2013 to 2018 to evaluate, assessed, and interpret all existing research related to the subject area "Mobile Participatory Sensing Systems".Thus, the literature review is an important step to evaluates and to better understand the challenges that exist in this area.Additionally, a research gap analysis has been conducted to have a clear and precise direction for future research work.The paper is based on the Kitchenham guidelines and is conducted using three stages namely, planning, conducting, and reporting.Therefore, the systematic review aims to direct future research by identifying gaps and open challenges in the mobile participatory sensing domain.

Planning the Review
A well-established procedure is used to define a suitable approach for the review, consisting of three phases.Firstly, the data source (inclusive of conference and journal events).Secondly, the domain area dependency (expressions to search).Thirdly, the inclusion and exclusion criteria consisting of the information sources to be investigated (counting databases, specific conferences, and journal proceedings).

Data Sources
Mobile participatory systems have gained a high level of popularity and interest in the research community over the past decade.Several databases were explored to conduct searches where relevant papers were found in IEEE Explore Digital Library and ACM Library.Thus, these two mentioned libraries were selected for the systematic review.The most noteworthy importance of indexed papers was found in IEEE Xplore Digital Library after exploring some databases.Thus, the selected library for this review is IEEE Xplore Digital Library.

Search Terms
To narrow down the selection process of the publications, a custom range is applied to journals.The exploration term (mobile participatory) AND (sensing applications) were used.The review was conducted over the last 5 years, ranging from 2013 to 2018.A complete sketch of the approach for mining articles is presented in section 2.4.

Inclusion and Exclusion Standards
The inclusion and exclusion criteria that were used to accept the relevant articles to assess the state-of-the-art mobile participatory sensing systems are presented in this section.
Mobile Participatory Sensing Systems: A Comprehensive Review 3 For an article to be reviewed, the following features and characteristics need to be demonstrated:

Conducting the Review
In this section, the planning phase is discussed.The planning phase consists mainly of a selection and analysis stage.Mendeley along with Excel was used for classifying the documents.Additionally, data mining and analysis were conducted on the selected works.

Selection Process
The Kitchenham multistage procedure model was used as the selection process.A total of 590 non-copied articles were recovered and retained.Once the exclusion criteria were applied, only 35 articles were retained.As a final point, after combining the exclusion criteria to that of the inclusion measures, 24 relevant research articles were retained.Table 1 shows an overview of the 24 selected papers.

Data Analysis
In this section, the results of the data analysis are displayed.The results were useful to spot the requirements and core features of participatory sensing systems.The systems were reviewed into three major categories of participatory sensing systems.They are as follows, environmental participatory sensing systems, people-oriented participatory sensing systems, and task-oriented participatory sensing systems.The second segment consists of a comprehensive analysis of the systems based on the following criteria: accuracy, reliability, real-time, trustworthiness, and scalability within each of the three categories.
The above-stated categories are identified for data analysis as per the number of applications being developed, contributions of researchers, and the widespread of these applications.Nowadays there are special reasons for thinking about the problems of cities in terms of creativity or lack of it.Factors that once shaped city development were transport, infrastructure, and the different types of pollution.However, there is a need to give special attention to elder people.Common users of these systems are aged care service providers with an exceptional focus on health-related issues, which is the people-oriented part.Additionally, using a proper platform, people can report environmental issues or provide appropriate services on environmental changes.On the other hand, access to resources, data, and information is restricted due to the geographical landscape.Thus, in a disaster situation, authorities may be helpless even possessing the best hardware and manpower.Therefore, there is a need for a task-oriented analysis.

Reporting the Review
After investigation of the designated articles, the mobile participatory sensing systems are grouped as per three categories namely environmental, people-oriented, and taskoriented.The following sections present the outcomes and classification of the mobile participatory sensing systems by analysing the 24 papers in Table 1.

Environmental Participatory Sensing Systems
The main aim of an environmental system is to provide realtime information to the authorities.Services provided can be air monitoring, water monitoring, or sound monitoring.Some frameworks send instructions to the nodes to extract data, whereas others allow participants to send information in a participative way.The mobility of participants can largely affect the server's decision or process.Nowadays environmental processes vary dramatically in time and space.Thus, the system needs to work with a wide range of data.
The environment is a shared space for resources and services which activate entities to provide functionalities such as traffic monitoring, among others.When participants are collecting data and send them to the server, the system should be able to classify the data to the corresponding services.In some systems, different rules can be defined in entities.It can be in the form of a visual check.In other systems, the services provided can be directly linked to the format of data being collected.One example could be that all traffic services can be stored in a picture format.Therefore, the server can track the format of data collected to provide adequate services.
Other systems can define a communication protocol such that the different types and formats can be forwarded to the server by one participant.Data and resources can be identified, generated, modified, or used by the participants.The extent to which the participants can access a specific service may depend on several factors as the system should have a proper level of abstraction.If traffic checking is considered, the street quality will, in general, be adaptable, with potholes and uneven streets being natural.Vehicles types are also very diverse, going from 2-wheelers (e.g., bikes and motorbikes) and 3-wheelers (e.g., auto-carts) to 4-wheelers (e.g., vehicles) and bigger vehicles (e.g., transports).

People-Oriented Participatory Sensing Systems
The smart city is a label where the city is aggressively chasing the use of technology to upturn the quality of life.Some systems are helping to reduce the C02 emissions, waste, improving public and private sector services while others are optimizing energy consumptions by building efficiency and renewable energy productions.Technology is not the only driving force in a smart city, citizen's experiences can help to develop better services and solutions.To reduce failure, one can listen to citizens' potential problems at an early stage.
People can help in collecting data that detected noise and ozone pollution as they lived their regular lives, and the result can be displayed over a mapping engine.These frameworks can help in decision-making processes and also build sustainable local communities where individuals care for one another.Having a people-oriented approach can aim to foster more informed, educated, and socialized citizens.It is also an initiative to allow followers of a city to contribute to the management and governance of a city.It also helps to become an active user.Interest and coordinated effort between government, residents, and associations are viewed as fundamental in the improvement of smart communities.Some of the activities (parks and recreation, community development, and planning) typically involved in smart city projects can benefit greatly from citizen participation.A recent study [13] demonstrates an association between cities' acceptance and application of sustainability strategies as well as public contribution in policy formulation.
Some cities have already developed participatory sensing systems that help to better understand human behavior and to obtain real-time information about environmental conditions and traffic.Counties such as Chicago who has been implementing the Array of [15] since 2014 to monitor air quality component.For example, in urban spaces, many participants can collect and transmit information about air quality.Thus, people and local authorities can be aware of the level of pollution at different locations.This information can be vital for those who have health issues such as asthma and therefore avoiding any risky areas.Louisville [28] has proposed an innovative approach since 2012 and uses GPS gadgets installed in inhalers to gather information on where and when individuals with asthma are especially affected.The concept of this pattern is to allow the city to forecast possible events and take precautionary measures.Each city's community requests the implementation of such a system that permits residents to report problems that they are facing.It can be in the form of broken streetlights, unlawful stopping, water spill, and so on.Participatory sensing tools will allow these organizations to initiate data collection that likewise connect people to the preparation of their environments.
Researchers [21] defines a GIS-based noise preparation tool made for the city of Belo Horizonte in Brazil.Such systems method allows citizens to collect and share basic information on ambient sound at regular intervals.This application could join a data-collection campaign to document noise levels in a community.Citizense -A generic useroriented participatory sensing framework A generic system that can be implemented in the different domain area

Data accuracy has been addressed
Trustworthiness among data collected has not been addressed

Task-Oriented Participatory Sensing system
Task-Oriented participatory sensing applications can be defined as an application area for multi-robot teams.It consists of task-oriented missions, in which possibly heterogeneous robots must solve several individual tasks.
There is a high demand for these types of applications.The arranged robot needs to adapt to non-critical failure.Taskoriented can also be applicable to help users with a disability.For example, a task-oriented system could assist a blind person to follow a specific route.Such a person must have some perception or plan of that route.Thus, the traveler can learn the direction while being directed by a sighted escort or may only have verbal instructions.When a path has been practiced, a user can distinguish and avoid obstacles.Additionally, the traveler can follow the course to know their position, direction, and make fundamental revisions.In disaster situations or mountainous landscapes, a versatile participatory framework should be able to process complex resource-intensive tasks.Due to the restricted access to resources, services, and multimedia, acquiring distributed information for emergency responses is a critical factor.
Relevant research disciplines such as in-network processing, service discovery, and service composition help to better manage restricted areas.In a disaster situation, authorities may be helpless, even having the best hardware and manpower.When the participants have already collected and sent data, the server will be able to monitor the changes or the frequency of data being pushed from a region or a service.The server can then apply a priority mechanism to task processing.Thus, a task-oriented will help to give better information on how to plan, manage, and execute an action before, during, and after a disaster situation.
In complex systems such as robotics, human errors can be disastrous and are beyond our expectations.Thus, to avoid this, a task-oriented system can be used.Many structures, such as a crawler, a legged robot, a manipulator, or other robotic shapes, can be assembled by the arrangement of (identical) modules.This kind of rigidity is extremely desired for robotic systems used in unpredictable and unstructured

EAI Endorsed Transactions on Mobile Communications and Applications
Online First Mobile Participatory Sensing Systems: A Comprehensive Review environments.It can operate as a rescue operation in earthquake-stricken areas or deep-sea and space exploration.An alternative advantage of self-reconfiguration is self-repair by changing broken modules.Thus, the system is not down waiting to be repaired where human intervention latency could prove to be disastrous.An example would be in a nuclear plant where task-orientation is extremely important.

General Results
In this section, the 24 mobile participatory sensing systems are evaluated against the five criteria which are accuracy, reliability, real-time, trustworthiness, and scalability.

Accuracy
In a participatory sensing application, providing a service will largely depend on how accurate the data has been captured.Accuracy is a requirement that is desired before any processing to provide adequate services in a participatory application.Over the years, many researchers [1,12] have largely contributed to the participatory sensing systems.At the end of this section is an analysis of the impact of "accuracy" in those systems.Accuracy can be defined as the acceptable level of measurement with an accepted standard.

Real-time Latency
A machine cannot act on something, such as providing a service instantaneously, and the amount of waiting time between an input and its output is called latency.In a participatory system, to preserve delay between input and output such that it is non-existent, the latency must be low, that is the system must react in real-time.Consistent latency is the most desirable requirement in a system.The amount of data being collected or pushed from participants, and the constantly changing environment data should be relatively low.

Reliability
In a participatory system, when data is sent for processing, it is important to check if the captured data is still relevant.For example, if a participant has captured data and is disconnected from the network, it is important to check the reliability of the data.When the user will connect to communicate the data to the server, the relapsing of the data has changed.As such reliability can be defined as the ratio of data being collected that produces correct output to the total number of participants in a participatory sensing system.

Trustworthiness
It will check whether a malicious person will send irrelevant data to the server.Often while capturing data, several types of noise adversely affect the collected data and when this data is sent to the server to map an environmental change, the system reports inaccurate data to the users.Thus, the application server must have the necessary mechanisms to assess the trustworthiness of the collected data so that malicious contributions, corrupted, and incomplete data can be detected.

Scalability
It is described as the capability of a process, network, software, or association to develop and manage increased demand.A system or software that is labeled as accessible has a benefit because it is more flexible to the varying needs or demands of its users.Scalability is the property of systems to deal with a developing measure of work by adding assets to the systems.In a participatory system, the number of participants is not fixed, instead, it varies, and to handle the constant change of participants, the system must be scalable.
Table 5 presents the evaluation of the 24 mobile participatory sensing systems against the five criteria.Three benchmarks are used as the level of acceptance which is: low, average, and critical.A rating analysis is performed based on the following details:

EAI Endorsed Transactions on Mobile Communications and Applications
Online First Mobile Participatory Sensing Systems: A Comprehensive Review

Discussion
In this section, the 24 mobile participatory sensing systems are discussed as per the categories identified in section 4 and section 5. Potential challenges are identified and discussed.

Environmental Participatory Sensing System Challenges
The environmental participatory sensing system has a high impact on our daily life.From the 24 systems stated above, 11 systems have been identified as an environmental participatory sensing system and the overall rating is 157.Different systems such as traffic monitoring, air monitoring, water monitoring, or sound monitoring have contributed A. Rajoo et al.

EAI Endorsed Transactions on Mobile Communications and Applications
Online First hugely to this category and there is an increasing demand for such systems.Some researchers [27]  represent task-oriented systems, with a rating of 109.These structures can function in a multidisciplinary location and be technology autonomous.Some researchers [33] have shown the interoperability and the immense opportunities of such systems.There is an apprehension among employees of being replaced by these systems to cut costs.This is due because researchers are focusing on how technology can increase productivity while decreasing cost and death rather than proposing how humans can interact with these systems to increase efficiency.Below is a representation of the five criteria's assessment: From this review, it has been noticed there is an important consideration for environmental systems.The reliability and trustworthiness of data remain critical criteria for the participatory sensing systems.One criterion that demand further attention is scalability.Most of the systems identified have either a limited number of participants during testing or participants are not willing to join and share information over the network.Table 6 shows a classification of the five criteria as per the priority of the systems.

Interpretation of Results
From table 6, a detailed analysis has been performed from the 5 criteria which showed where there is a need for future research and what is the ongoing domain interest.Scalability is a domain where improvement is required due to the increasing number of devices and users.On the other hand, trustworthiness and reliability are the areas where researchers are exploring and contributing heavily.

Scalability
Mobile participatory sensing systems should be able to define the maximum number of users and devices.In an environmental system, the targeted audience should be on a larger scale due to the constant changes of data.Therefore, scalability will focus mainly on the number of users participating in the system as it is difficult to give an exact amount of users in such systems.For a task-oriented system, it must harness a greater set of processors.Therefore, it must scale to a higher amount of processors compatred to the existing multiprocessor systems, such as networks of workstations.Thus, scalability will have a different behavior depending on the category targeted.

User Participation
The number of participants is a critical characteristic to ensure proper service.Users are expecting to have accurate and reliable services from the system, else participants will not be willing to use and participate.Thus, there should be a proper mechanism, architecture, or protocol to increase the participation of users.Some researchers [16] have proposed incentives depending on the user's participation while others have proposed a reputation model [16].Nevertheless, there is a big challenge to encourage the users to capture, send data, and to use the services.

Efficiency/Power Consumption
Data should be continuously sensed and captured to prove a real-time service.This addresses the limited power resources which mobile devices and sensor nodes possesses.Futher research should be conducted to provide a "dormant-active" state of sensing devices or optimization algorithms [8].The in depth research could apply to a limited degree to report mobile participatory sensing systems topics such as node localization, optimal deployment, and data aggregation.Interoperability A mobile participatory system is highly desirable to interact with different technologies.To guarantee a general data format, interoperability, and conventions can be used.In a mobile participatory sensing architecture, different types of devices and different types of OS are communicating.Diverse protocols and software are also used to address the topic of interoperability.Therefore, the widespread use of mobile can be used to report the characteristic of interoperability in such systems.

Conclusion
This paper presented a review in the state-of-art mobile participatory sensing systems and categorized the systems as environmental, people, and task-oriented with a specific focus on topics and experiments related to mobile participatory sensing systems.Nowadays most of the mobile phones are equipped with sensors and various applications are dependent on the data being generated by these sensors.
Researchers have started and are exploring this technological solution to enhance the provision in the different categories.
In this paper, 24 mobile participatory sensing systems from 2013 to 2018 are evaluated to improve ongoing research/systems and various prominent problems are defined.From the results, it is seen that there is a demand for mobile participatory sensing systems, especially environmental systems.Moreover, people-oriented, and taskoriented participatory sensing systems appear to be gaining in popularity and there are a number of pressing challenges that needs to be overcome.Researchers have shown how critical data reliability and trustworthiness of data is, without neglecting the other criteria identified in this paper.However, as pointed out in this paper, several challenges remain to be addressed to attain an almost flawless mobile participatory sensing systems.In the near future, the evolution of mobile participatory sensing systems can significantly improve the day to day life of people.The main challenges in such systems were outlined, and areas that researchers should investigate futher for motivating users to participate were discussed.

Table 1 .
An overview of the 24 papers identified

Table 2
is derived from table 1 where all environmental-oriented participatory sensing systems are analyzed.A. Rajoo et al.EAI Endorsed Transactions on Mobile Communications and ApplicationsOnline First

Table 2 .
Environmental Participatory Sensing Systems EAI Endorsed Transactions on Mobile Communications and ApplicationsOnline FirstMobile Participatory Sensing Systems: A Comprehensive Review Table 3 is derived from table 1 where all peopleoriented participatory sensing systems are analyzed.
A. Rajoo et al.EAI Endorsed Transactions on Mobile Communications and ApplicationsOnline First

Table 3 .
People-Oriented Participatory Sensing Systems

Table 4
is derived from table 1 where all task-oriented participatory sensing systems are analyzed.
EAI Endorsed Transactions on Mobile Communications and ApplicationsOnline First

Table 5 .
Evaluating the Participatory Sensing Systems have shown how critical such applications are now.It has also been noticed that these systems are limited in terms of users' participation, that is, the rate at which users are entering or leaving the networks.

Table 6 .
General evaluation of Participatory Sensing Systems