We are

University of New Brunswick

The 

The People in Motion Lab is located at the Department of Geodesy and Geomatics.

 

More information here.

University of New Brunswick

Department of Geodesy & Geomatics Engineering

Head Hall, E-50
P.O. Box 4400
Fredericton, N.B. E3B 5A3

Canada

Research Projects

Our research projects are being funded by:

  • Cisco Canada 

  • NSERC IRC, NSERC Discovery Grant, NSERC Engage 

  • NBIF RA

  • Mitacs Globalink Research Internship, MITACS Accelerate

  • SSHRC Partnership Development Grant

  • EU FP7 PEOPLE Marie Curie Action

  • US National Science Foundation (NSF), Office of Cyberinfrastructure (OCI)

  Investigations and Analysis of Industrial Internet of Things Ecosystems  

UNB has partnered with Rimot to undertake this project over the next year to (1) generate a five-year artificial dataset of internal RF site sensor data based on actual external weather data (2012-2017) and (2) to use that data to determine the best methods of developing models that can predict site degradation and failure. Over the same time period, Rimot will continue to collect sensor data from customer sites for future testing using the methods discovered.

UNB will determine the optimal mix of computer power availability with edge analytics with local machine learning algorithms. Towards this end, the main objectives are

 

  1. Select the appropriate metrics and three types of analytics capabilities (i.e. descriptive, diagnostic, and predictive) in order to determine which one(s) would be most beneficial for analysing machine generated data.

  2. Analyze existing data streams available from all the sensors within the network and determine a short list of most applicable algorithms using different metrics (e.g. performance metrics, algorithm transparency metrics, data quality metrics).

  3. Evaluate the results taking into account the trade-offs of computing power of edge device(s) versus the insight capability of the system, as well as data latency and data rate requirements.  

  4. Test these algorithms in a closed environment against a small subset of the data.

  5. Build a prototype as a proof-of-concept.

  Developing machine learning classifiers for general-purpose sensing in IoT  

Our role was to develop an edge-fog-cloud architecture to support a general-sensing approach that is minimally social obtrusive (i.e. the sensors will be better integrated into the environment); capable of indirectly sensing and be easily plugged-in to a power outlet. This architecture is of paramount importance to support adaptive machine learning classifiers that will be used to answer questions of interest to citizens to improve their quality of life. We are working on developing a pilot prototype in Fredericton (UNB) and Madrid (UPM) to evaluate the flexibility of general-purpose sensing in reducing the costs of having thousands of physical sensors, including even greater costs of deployment and maintenance. The outcomes of the project will be used to enhance the human experience, especially the one related to different physical contexts (e.g. home, office, hospital) and the human activities taken place within.

  Smart Connected Community Strategy  

The projects were developed in partnership with the community in Saint John to begin to explore the insights, efficiencies, and competitive edge that Big Data and IoT can bring to their organizations. We worked in teams on the following projects:


Smart Parking:  

A team of 3 undergraduate students from UNB have developed an approach for predicting parking availability using machine learning techniques. Drivers who use Hotspot smart parking services were able to receive information about park spot availability that can shorten their parking search time, reduce environmental pollution, reduce costs with less fuel consumption, and alleviate traffic congestion. For the Saint John Transit, the project provided a cost benefit analysis about the benefits and drawbacks of expanding the smart parking services in the city

 

Tourism Intelligence:

A team of 2 graduate students and 1 undergraduate student developed a machine learning algorithm to build smart tourist profiles which are dynamic and context-based using tourists’ travel values, social values, and attractions of interest. Our first prototype was evaluated in the city of Saint John using beacon technology.

 

Smart Water - Can Big Data save our water infrastructure?:

One undergraduate student has investigated how IoT strategies can be used for analyzing water infrastructure. New sensor technologies have arrived on the market to help Saint John Water survey their underground pipes and detect real and potential leaks. This is an exploratory project looking for descriptive and prescriptive analytics to detect real and potential leaks.

  Real-time Social Signalling and Collective Goods  

The Internet of Things (IoT) will have ten times greater impact on Geomatics than the Internet. It was critical for our research group to engage a multidisciplinary team between UNB and McGill University. This team is capturing value from big data, deploying new IoT technologies & techniques, and show the way to mobile solutions by encouraging the creation of new spatio-temporal and socially-inspired innovations. In this project, we are working on collecting data using wearables, mobile applications, GNSS receivers, and generic sensing platforms. The range of new methods and innovations from our research is contributing to the conception of the next generation of IoT-GIS, especially those providing information concerning privacy-aware mobility patterns of things over time. The disciplines of law, computer science, engineering, management, behavioral economics, communications studies, and biology are represented among the researchers collaborating in this project and cover all the dimensions of these key research areas.

  Design of new mobile applications for the Internet of Things  

Our role was to integrate a mobile application with an IoT platform for a real-time recommender system that enables streaming data collection from smartphones in order to recommend new items on the fly using geofencing as a user context. Geofencing is a virtual circle defined by a centre point and a radius. The geofencing was generated by using two approaches: (a) without IoT devices: using the Google Location API for defining a-priori coordinates of a point of interest and its radius; or (b) with IoT devices: using the physical position of IoT devices and their signal range as the radius. We have selected for our experiment the city of Saint John in New Brunswick, Canada, mainly because tourism is critical to Saint John’s culture, heritage, arts, recreation, and entertainment industries, and it also contributes significantly to city’s service industries including transportation and travelling services, accommodations, and food and beverage services. It is a 20 billion industry, with the Port Authority expecting over 60 cruises with a total of 140,000 passengers and 50,700 crew members in 2016.

  Developing a space-time synchronization process in complex road transportation networks  

Our role in this project is to explore big data that will be produced by smart-connected vehicles in the future. Towards this end, our aim was to study how a synchronization process in complex networks can be used in road transportation systems for minimizing delays and fleet management. The unique IoT data provided by Google® and Telit® was used in the project. One of the main outcomes of this project was to demonstrate the full potential of the Internet of Things and enrich education in Big Data at UNB in order to generate business value and produce real-life solutions in the Geomatics sector.

  Discovering flow patterns across spatial and temporal scales  

Our research was focussed on developing a scale-free geographic knowledge process for uncovering the formation, the structure and the dynamics of flow patterns. Flow patterns are inherently not random as they have a scale-free structure in space (e.g. from vehicles, crossings, highways, to networks) as well as time (e.g. from minutes to an hour time). New mobility measurement and flow pattern techniques were developed to mitigate the scale issues due to the use of different data sources. The outcome of this research contributed to narrow the knowledge gap between the vastly collected vehicle mobility data and very limited capacity of flow pattern information extracted from this data today. It has also helped traffic managers to effectively utilize the discovered flow patterns within their applications such as Intelligent Transportation Systems and Emergency Evacuation Planning in smart cities.

  Playful Planning: Citizens making sense of transit systems and impact  

The Black Arcs (TBA) anticipates civic engagement as a key component to the planning process in smart cities. The problem TBA was facing was due to the difficulty of having citizens connecting to planning and understanding how smart cities will affect their lives each day. Our role in this project was (1) develop a suitable graph data structure to allow access and query transit data of spatial transit network; (2) conceptualize the game scenario for modelling a citizen transit decision process based on how denser land use around transit stimulates the demand for public transport services and how citizens can influence in a positive way to promote sustainability; (3) develop a user-friendly, intuitive game interface to support civic engagement.

  Utilizing social media to support real-time (Re) Insurance Catastrophe Response  

The Insurance Bureau of Canada has reported that the December 2013 ice storm in southern Ontario and eastern Canada resulted in $200 million in insured losses and pushed the year-end severe weather insured loss total to $3.2 billion, which is the highest in Canadian history. Analyze Re expects a huge demand for new risk assessment products among medium-sized companies in the $500-billion reinsurance market. Our role in this project was to develop algorithms and a Hadoop cloud to specifically facilitate an insurer's need to access factual data about ongoing or recent catastrophe events in real-time by examining geotagged Twitter feeds. Analyze Re have incorporated the algorithms and technology discovered by the project to expand their existing product suite and offer solutions for managing catastrophe response scenarios to customers in the insurance and reinsurance sector.

  Master Vehicle Data for Supporting Advanced Learning in Big Data Analytics  

Traffic data were the first data to be automatically sensed in cities but databases go back centuries. The new wave of traffic data that has emerged during the last decade and it was fostered by the widespread diffusion of wireless technologies, GNSS, and mobile phone networks. Our role was to develop a set of automated algorithms for disaggregating macro scale traffic data, and aggregating micro scale traffic data, in ways that can handle the problems of traffic data acquisition at different geographical scales. New sampling techniques were devised for measuring vehicle mobility in cities using four types of technologies: high resolution satellite images, video cameras, mobile phone networks and GPS. The outcomes improve our understanding on the effective and feasible ways to coordinate different technologies that could provide master vehicle data in the near future.

  MODAP – Mobility, Data Mining and Privacy  

MODAP was a Coordinating action (CA) project of the ICT FET, FP7-ICT-2009-C that aimed to stimulate an interdisciplinary research area combining a variety of disciplines such as data mining, geography, Geomatics, visualization, data/knowledge representation, and transforming them into a new context of mobility while considering privacy which is the social aspect of this project. Our roles were to continue the efforts on Privacy Observatory in Europe, organize interdisciplinary workshops to bring together the technical and non-technical experts; develop training activities in the form of summer schools and tutorials to disseminate the research on privacy and to initiate PhD ideas in the area. 

  DYNCOOPNET  

Our role in this project was to improve the awareness of the need for the integration of different historical data sources and making them available through a DynCoopNet Portal, which has led towards the creation of a Cartographic Archive at the Technical University of Madrid similar to the ones developed for the Cartoteca de Canarias and the Cartoteca del Servicio Hidrográfico de la Marina.  The data sources were published and archival sources about the commerce and routes maintained between European cities, ports and islands; access database of personal information of financiers, tax system, and commercial world in Europe, and the Cartographic collection of Naval Museum, the National Cartographic Library, and others.