**1. Introduction**

Within the context of the artificial intelligence, big data, and cloud computing (ABC) era, 2019 is the first year to have fifth generation (5G) mobile Internet service, meaning big data based on the cloud computing environment must rely on artificial intelligence technology for processing. Without solutions for the volume, velocity, variety, and veracity (4V) data characteristics, the problem of garbage in, garbage out (GIGO), which is often mentioned in computer science, would continue to occur. Therefore, open data can be a valuable tool to solve this problem. Taiwan is a densely-populated country with heavy traffic, especially in metropolitan areas; thus, an open traffic database provided by government units is very important. This study aims to solve the problem of "true" and "false" in big data floods, which are commonly found in cloud information systems, by using the linked open data (LOD) technology of government units.

Smart mobile devices have become prevalent in recent years. According to the eMarketer survey data of 2016 [1], Taiwan has the highest penetration rate of smartphones, as compared with Singapore and South Korea. According to the 2018 Household Digital Opportunity Survey Report of the National Development Council [2], the mobile Internet usage rate among Taiwanese grew from 41.9% in 2009 to

98.2% in 2018, and mobile Internet usage has long been the main trend of Taiwanese Internet usage. The survey also pointed out that 84.9% of Taiwanese people over 12 years old use mobile Internet, while more than 60% of Taiwanese people under sixty-five years old use mobile Internet; therefore, the use of information systems or applications on various mobile devices by Taiwanese people via the Internet is undoubtedly a fairly mature digital skill. Thus, appropriate, convenient, and correct mobile or cloud information systems or applications are in great demand.

Taiwan is a highly informational country, and transportation is one of the most important issues related to people on a daily basis. The transportation network is critical to the national economy, and is the benchmark index of public construction. Building a sound, smooth traffic network is an important infrastructure for the country's economic development, while traffic has a negative association with a large influx of vehicles, road construction, and traffic accidents [3–6], which are the reasons for road congestion. However, Taiwan is a country with dense population and heavy traffic, which is one of the main causes of road congestion. Therefore, there is a growing demand for information systems that can facilitate real-time and dynamic queries, and present current traffic information, especially for mobile application users.

With the prevalence of smartphones, people have easy access to communication applications, such as LINE, WeChat, Telegram, and Skype. As of March 9, 2018, the monthly active users of LINE reached nineteen million in Taiwan, with an opening rate of about twenty times a day [7], which also represents the dependence of users on smartphones. In addition, the Taiwanese have a large number of private cars [8], and self-help travel is a common mode of transport [9]. Therefore, most of them rely on car navigation systems, resulting in a higher probability of road congestion [10], such as the problem mentioned by the Waze system—because there is no car owner feedback mechanism, this leads to a dilemma in avoiding the "A" road section, crowding drivers into the "B" road section, just like the problem caused by using only Google Maps [11]. For this reason, if users can use mobile devices to access the real-time traffic information on communication applications, they can be updated on the latest road conditions, and thus avoid traffic congestions. This study aims to develop a traffic assistant agent with a simplified reporting procedure on an intuition-based graphical interface for smartphone use.

Therefore, this study used LOD and global positioning system (GPS) technology to develop an intelligent traffic assistant agent with user-friendly interfaces and a user reporting mechanism to instantly check traffic flow. This study further explored the feasibility of developing a traffic assistant agent that does not require users to download the application, provides cross-platform services, saves system development costs, and has practicability, accessibility of information, excellent user experience, and approximately optimal balance of the important interface design.

#### **2. Literature Review and Development Technologies**

This study first reviewed the various traffic assistant information systems available in Taiwan, including Road Condition Autotoll [12], Real-Time Traffic Image—RoadCam [13], Police Broadcast Real-time Traffic [14], and New Taipei City Advanced Traveler Information System [15], as shown in Table 1. In summary, although those systems can use the responsive web design (RWD) technology to present the information page with the most suitable size and can automatically retrieve the users' position, they still need improvement on providing the correct corresponding information, and most of them lack the user reporting function.


4

Information System (capture time:

December 24, 2018)

New Taipei City Advanced Traveler

The website uses responsive web design

longitude and latitude positions originally set by the map are presented.

Furthermore,

needs to be checked manually. Lack of user reporting function in this

study.

 entering the website still

technology and can view relevant

information pages in an optimal size on

an intelligent handheld device.

Open data, especially government open data, is a huge resource that has not been fully developed. Since 2011, Taiwan's government has also started to actively promote open government and open data platforms to provide public access. The JavaScript Object Notation (JSON) format is most commonly used on the government open data platform, which can store and transmit data in plain text. For more complex types of data, transmission can be achieved through objects or arrays [16]. LOD is an applied open database that organizes the data set according to the principle of linked data, and there are many research studies related to open data or LOD in the literature. For example, Wang [17] combined common classification methods, K nearest neighbor (KNN), support vector machines (SVM), and decision trees to explore and link open data in order to achieve automatic classification of news articles. Yan [18] took the government open data platform as the database, and explored the technologies related to building the resource description framework (RDF) based on a prototype platform. Yu [19] studied Taipei City's open data platform's traffic parking lot, smiling bicycle (a public bicycle rental system in Taipei), and mass rapid transit (MRT, a rapid transit system in Taipei) data to explore relevant technologies for LOD applications. Albert et al. [20] presented the research project census data open linked (CEDAR) data set, which is the historical census of the Netherlands in LOD format, to explore a more accessible, better connected, and history-related integration technology. Alobaidi et al. [21] used LOD to enrich the query content to improve search effectiveness and ranking. Selvam et al. [22] proposed a systematic approach using ontology and LOD with significance to semantic links in the social event detection (SED) task. Pourhomayoun et al. [23] proposed an effective end-to-end system for traffic vision, detection, and counting on real-time traffic open data. Most of the above literature used LOD support to add and evaluate subsequent information services. This study used the Azure cloud platform support to extract appropriate corresponding LOD information (as detailed in Sections 3.2 and 3.3) after referencing GPS position capture and conversion comparison, which supports the operation of the overall information system and enhances the correctness, authenticity, and integrity of location-based service (LBS) information service provision.

GPS is a satellite network that can measure satellite distance to accurately capture the position on earth by continuously transmitting coded information to satellites, thus providing corresponding information services. The GPS information is free and open to users around the world. There are many research studies on the GPS architecture and information systems in literature. For example, Juan [24] quoted GPS signals from mobile devices to explore relevant technologies to analyze the space–time data of office workers' lives. Liou [25] integrated the GPS and Beidou satellite navigation system, put forward an adaptive multitime algorithm, and explored relevant technologies to achieve single-frequency, real-time, precise single-point positioning. Tu [26] connected the electronic collars of pets to smartphones as an invisible dog leash, and used the BLE (Bluetooth Low Energy) received signal strength indicator with GPS to explore the relevant technologies of real-time pet tracking. Fridman et al. [27] used styles, applications, web browsing, and GPS positioning to realize the related technologies of active authentication on mobile devices. Based on the survey of GPS-assisted travel in Beijing suburbs, Ta et al. [28] studied the related technologies of personal commuting efficiency according to the difference in commuting distance and route selection efficiency between the morning and evening peak traffic hours. Aliprantis et al. [29] described a concept with image identification and matching from the Europeana platform, which can link the LOD cloud from cultural institutes around Europe and mobile augmented reality applications for cultural heritage without accurate geo-based locations. Khaghani et al. [30] proposed a platform for dynamic performance assessment of roadway networks, leveraging coarse GPS data from probe vehicles, such as taxis, to quantify the resilience of road networks using a multidimensional approach. As discussed above, it is a development trend of contemporary information systems to quote GPS to explore corresponding information services (as detailed in Section 3.2). This study used GPS location acquisition and conversion to explore how to quote appropriate LOD and related technologies to provide corresponding value-added LBS information services.

In the age of the Internet, mobile devices have been widely used to deliver personalized services. For example, LBS can be integrated with various technologies, such as GPS positioning and mobile data, to provide comprehensive applications for services related to spatial location. In fact, LBS was first provided as a rescue service by an American operator in 1996, then called E911 [31]. Syu [32] used the message queuing telemetry transport protocol (MQTT) to carry out data transmission and realize mobile phone application push and broadcast function. They applied relevant technologies to calculate the value of taxi drivers' LBS in the Taoyuan area of Taiwan. Chou [33] combined the Internet of Things (IoT) and LBS, using Arduino and mobile devices to identify the common causes of accidents in China, including not paying attention while driving, not keeping a safe distance, careless driving, drunk driving, etc. They also implemented various technologies to integrate sensors with government public data to provide suitable services. Sun et al. [34] used the location label to mark the sensitive and normal locations of mobile users, and designed an algorithm based on the location label to explore relevant technologies to protect users' location privacy. Wang et al. [35] discussed the issues of location awareness and privacy protection, and explored technologies related to location-based services according to users' requirements, as based on different locations. Ukrit et al. [36] proposed an LBS architecture PROFILER (a framework for constructing location centric profiles), which was centered on the discrete locations visited by users, and explored the related application technologies. Lin et al. [37] described the methods for efficiently finding the links across maps, converting the data into RDF, and querying the resulting knowledge graphs to solve the problem of how to convert vector data extracted from multiple historical maps into linked data. Sansonetti et al. [38] proposed a research study on integrating the recommendation process of nearby points of artistic and cultural interest (POIs) with related multimedia content, exploiting the potential offered by LOD by following semantic links in the LOD graph with the specific POI to provide personalized suggestions. As discussed above, location-based services, mobile positioning services, and location services can obtain the location information (geographic coordinates) of mobile users through the mobile operator's radio communication network (e.g., GSM—Global System for Mobile Communications, network, CDMA—Code Division Multiple Access, network) or external location method (e.g., GPS). With the support of the back-end geographic information system (GIS) platform, it provides corresponding value-added information services. This study utilized the Azure cloud platform to explore how to intercept appropriate LOD information after referencing local GPS location acquisition and conversion to support the overall system operation, thus effectively adding value to the quality of cloud information consultation and the sharing function of this system.

Microsoft Azure [39,40] is Microsoft's public cloud service platform. At present, Microsoft Azure can support up to 30 kinds of service contents, including computing, storage, analysis, network, management, and identification services. In addition, Microsoft Azure has data centers around the world, which have tens of thousands of servers to provide users with applications and research. If one server suddenly fails, another server can take its place to operate in real time to ensure the sustainability of website services. Many studies have used Azure as the research tool. Liou [41] proposed a set of air quality index (AQI) deterioration real-time early warning systems, as based on the Azure cloud computing platform, including a prediction model, evaluation model, and other system operation modules. Ho [42] used the decision tree model on the Microsoft Azure platform to analyze data, and produced a customer recommendation list with high purchase probability. Lin [43] developed an integrated entrance guard system using a face recognition system and Raspberry Pi 3 in combination with Microsoft Azure cloud services. Richard [44] introduced all information services provided on Microsoft Azure. Diaz and Freato [45] mentioned that Microsoft Azure has supported data administrators and developers to provide a rich platform for big data workloads, such as linked services with Azure data and Azure storage. Färber [46] presented the Microsoft Academic Knowledge Graph (MAKG), a large RDF data set with over eight billion triples with information about scientific publications and related entities, used to solve problems in the LOD cloud. Based on the above systems, and with the support of the Azure cloud platform, this study explored and developed relevant

technologies for cloud information consultation and sharing in order to automatically balance system operation efficiency and corresponding system stability.

The Google Maps API (Application Programming Interface) is a copyrighted and chargeable development kit developed by the Google company. Users can use the application functions provided by Google Maps (e.g., JavaScript API) by typing the kit and serial key into the webpage program. The users receive the data through the mobile client or the webpage, which presents the map and positioning information. The use of Google Maps API as a map in the literature is quite extensive. Chang [47] combined Google Maps API with a generic algorithm to generate an approximate optimal solution, explored the mathematical vehicle route model, and provided an effective route guidance plan, thereby greatly shortening drive time. Wang [48] analyzed the data of the old 119 system and the new 119 system, explored the related issues and research, and developed technologies for shortening emergency rescue response time after the emergency rescue service data was imported into the geographic information system. Rahmi et al. [49] consolidated MySQL as the main data storage space, and used Firebase to store additional data, where Firebase's real-time database processes chat data to provide corresponding notifications, along with Google Maps API to support GIS to explore and construct relevant technologies to meet the needs of mobile Android and web applications between doctors and patients. Xia et al. [50] introduced population grid data into a new gravity radiation model, used Google Maps API to obtain grid-level travel costs, and explored relevant technologies for population flow estimation. Tan et al. [51] adopted SPARQL (SPARQL Protocol and RDF Query Language) to query useful data from the DBpedia LOD database to acquire related data nodes and used the page rank algorithm to calculate the importance of each data node in order to build a concept map for awareness training in cybersecurity. Nyo and Hein [52] presented a technique for guiding and controlling autonomous vehicles by using the Google Maps API with GPS for localization of the vehicle on the Google Maps application via WiFi module. As mentioned above, the system also used Google Geolocation API to display the converted address information on the system page according to the location reported by the users' device and WiFi signal. Furthermore, Google Traffic [53] was used to indicate color stratification as the congestion degree of road sections, thus presenting the authenticity of the proposed system by verifying road condition information in multiple ways.

In summary, this study used a service-oriented information service system, which integrates LOD and GPS to achieve a mobile information service system, with the aim of solving all of the aforementioned problems. The proposed system serves no borders and uses a cross-platform approach, meaning users do not have to repeat learning for different programs. With the integration of Google Maps, the proposed system allows users to obtain the current real-time traffic information according to their current location. In the case of traffic emergencies, they can also inform rear road users to pay special attention, suggesting diversions through the system's registration form. The real-time information can be sent to the police units so the appropriate manpower can be deployed or other measures could be taken to reduce road congestion. In addition, the proposed system can facilitate user reporting of traffic conditions and can quickly check the nearest intersection monitor and sensors, so as to present the information on the most suitable page. Moreover, it can access cloud resources more quickly, and use and combine the open data provided by the government units to obtain the users' first-hand information regarding traffic conditions. In summary, the proposed system has the advantages of real-time information, simple and convenient usage, cross-platform use-related information services, and optimal balance of the important items of the interface design, without the need to update applications,

#### **3. Proposed System Architecture**

This study used Visual Studio and ASP.NET to develop the web applications. The Microsoft SQL server was used as the system data storage space, and Microsoft Azure was especially used to set up the cloud servers. Government open data was screened and analyzed according to specific conditions by using the circulation method of R language, and then traffic monitoring data content was built in the corresponding cloud database. The R & D of hardware equipment included an ASUS MD710 computer mainframe and ACER E5-572G notebook computer, which were used for information system development and testing. Finally, the receiver was tested on an ASUS ZenFone 3 intelligent handheld device.

#### *3.1. Overall System Architecture*

The intelligent traffic assistant agent is a comprehensive application with a location-based service as the main body, which includes a front-end agent web application (web app), a back-end database (SQL Server), and a corresponding cloud database (Azure SQL Database). The system architecture is shown in detail in Figure 1. The web application can present all of the operating functions of the proposed system: real-time information, endpoint information, and historical analysis. When a user opens the system through an intelligent networking handheld device, the system reads the longitude and latitude of the user's location through the GPS function of the device and presents it on the map, which can display the current road condition data in color; for example, green means smooth and red means congestion. Drivers are reminded to obtain the latest road condition information through the graphical interface in advance. The user reporting subsystem provides up-to-date road condition information, including member verification, addition of reports, and report records.

**Figure 1.** Overall system architecture.
