New wireless network architectures such as Ultra-Dense Networks (UDNs) and Ultra-Dense Heterogeneous Networks (UDHetNets) are enabling technologies to meet increasing demands and achieve the required performance of 5G cellular networks. Recent research considered employing the radio signals transmitted by LTE-A cellular networks for localization. The infrastructure of such systems can provide accurate results due to the wide spread of mobile devices and the ability to detect them, which can provide accurate estimation of occupants’ headcount. The advantage of a cellular-based system is its wide availability and ability to cover areas where Wi-Fi access points or Bluetooth devices are not available. The geographical area covered by cellular networks can be used to track occupants over the area of interest (e.g., at the building or university campus). This can allow analyzing occupants’ movement and understand individual as well as emergent behavioral patterns.
In this research, we propose and evaluate the performance of fingerprinting-based indoor localization approaches with LTE-A UDNs. Furthermore, we study the application of such localization approaches in different areas such as building occupancy count estimation, optimization of building operation, and estimation of COVID-19 spread based on building occupancy count. 
Click here to watch video

Inhabitants play a crucial role in the sustainable outcome of green buildings. Due to privacy, interactive behavior of occupants is not apparent to the designers. Inaccessibility to this information leads the designers, engineers to making assumptions of occupant behavior in building designs. Often, the predicted consumption tends to be significantly lower than actual usage. Nowadays, sensors and IoT enables designers to have information in buildings that may give knowledge on occupants’ interactive behavior. The goal is to identify typical interactive behavior (Personas) in buildings based on measured data and to evaluate building performance using the persona models for better sustainable outcomes.

Sustainable cities are complex systems. Designing an urban space requires a planner to consider an ever-expanding number of factors. Multi-modal transit systems, increasing population density, responsible water, and power management or air quality control are only a fraction of challenges that must be faced. Geographic Information Systems (GIS) are well-established tools used to tackle design and planning at a large urban scale. Discrete Event System Specification (DEVS) based modeling and simulation is a technique well suited to the virtual representation of the complex sub-systems that inhabit urban space. This research proposes the development of an automated DEVS based methodology aimed at assisting users in the design of GIS models of urban systems through a constraint-based optimization (Generative design) process. The generative design will allow users to generate multiple potential designs, simulate the target systems, and analyze their strengths and weaknesses. The evolution develops new and novel designs that build upon the previous iterations. The research aims at providing support for urban planners in designing spaces that optimally fulfill a set of constraints.

The School of Architecture at Carleton University has a new neighbor and, as in many other cases before in the history of university, the users of the existing buildings must contend with and adapt to new environmental conditions. The idea of this study is to propose ways to mitigate the negative impact of new buildings towards existing buildings on a university campus. The data collection consists of a qualitative survey given to students/users of the existing Architecture building and qualitative analysis of daylight and massing simulations. For this project, the use of Generative Design (GD) helped to analyze the data and to propose solutions to the current situation. Besides, the project will try to inform future Campus Master Plans in order to reduce possible negative externalities that new developments may produce towards adjacent existing buildings/spaces. This research focuses specifically on the relationship between the Architecture Building and the recently raised Nicol Building (Sprott School of Business) as a case study.

The goal of this project is to develop a collection of metrics that can be used independently or collectively to evaluate design proposals at a campus scale (working towards an urban scale). For example, flexibility metric (e.g. Olympic village or adaptive classrooms), qualitative / subjective parameters (e.g. human comfort), health factors (e.g. exposure over time), amenity locations: WiFi coverage / restrooms / study areas / quiet areas.

Research proved that CO2 sensors are among the most accurate in detecting the presence of occupants. In addition, CO2 sensors are non-intrusive, require no motion or special action to detect the occupant’s presence based on CO2 levels, and do not have extra cost since they are installed as part of the building code in Ontario since 2014. However, CO2 sensors are highly sensitive to the configuration, and the relationship between CO2 levels and occupancy information is not always linear. It varies case by case. (e.g., what will the effect of HVAC do on measured CO2 levels?). This motivates us to research the relationship between the closed space configuration and CO2 based occupancy detection. The goal is to determine the best placement of CO2 sensors for the most accurate occupancy detection based on the characteristics of the occupied space (e.g., dimensions, vents locations, etc.). Besides, we aim to calculate the latency between the arrival/departure of an occupant and the detection of an increase/decrease in CO2 levels, respectively.
Click here to watch videos

Energy management and conservation in buildings are facilitated by the use of equipment that monitor changes to the environment and ensure that appliances are controlled accordingly. These equipments can achieve this by the use of sensors, actuators, and control systems. The design, development, and operation of sensors, actuators, and control systems are governed by the theory of Cyber-Physical Systems (CPS). Recent trends in CPS design have imposed the need for increased performance demands and complex usage, this has significantly changed the manner of interaction between the cyber and the physical components. They are now tightly coupled and integrated at every level, making it challenging to design cyber components separately from the physical components. In addition to design complexity, another very important issue to be considered is reliability.
The main goal of this work is to present a robust discrete event approach for the design of cyber-physical systems (CPS). This method is built on an existing methodology called Discrete Events Methodology for Embedded Systems (DEMES). Starting with DEMES, we implement certain changes to accommodate the complexity and tight inter-connectivity of the physical and cyber aspects of a CPS. The various aspects of the method considered for improvement include: the development of a formalism for real-time fault definition in cyber-physical systems at the design stage using DEVS; a method for real-time fault detection and diagnosis of CPS using the DEVS formalism; and the development of techniques to guarantee fault-tolerance in CPS.

Self-driving vehicle technology could revolutionize small scale public transit systems through increasing efficiency and sustainability. This project focused on the development of a Carleton campus digital twin prototype to study cyber-physical, embedded systems for self-driving vehicles. A control system running on Real-Time Cadmium was developed to move a robot along a track representing the Carleton campus road network. As the vehicle moves along the road network, it communicates its position to a web GIS-based monitoring system. Antennas were used to establish communication between the monitoring system and robot, RFID chips on the track were used to emulate updates to the vehicle’s GPS location. The monitoring system allowed users to visualize the self-driving vehicles on the network and obtain information regarding their state through simple and intuitive interaction with the digital twin.
Click here to watch video

In recent years, our lab The Advanced-Real Time Simulation laboratory at Carleton University has built various simulation models based on the traditional Susceptible-Infected-Recovered (SIR) equations. These models are being used worldwide to predict pandemic dynamics. Our lab is experienced in integrating simulations and complex spatial visualization engines to allow decision-makers to study how the virus spreads. This research uses that expertise to study how viruses expand indoors and outdoors for a campus model. The study will also consider how the virus spread over a small area or a large area. It will examine different spatial requirements for the pandemic model. A detailed 3D models of the buildings on campus and adaptations of the SIR model for indoor environments will be used.  Geographical Information Systems and Carleton’s Digital Campus map will be used to build spatial versions of the SIR model with adaptations and improvements. The models will be available through the web for remote collaboration and for the use of decision-makers. This tool will permit studying different strategies for returning to campus at the end of the pandemic, and to deal with future outbreaks.
Click here to watch video

Large BIM models like campus or cities need a methodological approach to extract and run complex simulations. Modeling for simulation does not require detailed BIM models but an abstraction of the model. Often, these abstractions vary depending on the type and scale of the simulation. We aim to create an API to extract data from a large BIM model like campus or cities to run complex simulations. On the other hand, contextual visualization of the simulation results in BIM models will enhance the design. This project aims to have an effortless end-user interaction to extract the data and run simulations for complex models.

Large scale geospatial simulations at the municipal, provincial or higher levels typically generate massive volumes of data. Presenting these data in a comprehensive, intuitive way for non-expert users requires adequate visual support. In geospatial data, web-based geographic information systems (GIS) have become well-established platforms to do so. The DEVS-GIS Simulation Explorer we developed relies on OpenLayers, an Open Source Web GIS libraries, and the OpenStreetMap database to contextualize simulation results. As a case study, we use the disease spread simulation results for the City of Ottawa at the Dissemination Area (DA) level and Ontario’s province at the Census Subdivision (CSD) level. The application allows users to build choropleth maps that display the boundary polygons classified by color according to different variables produced by the simulation (infected, susceptible, deaths, recovered, etc.) Users can also animate the map to visualize each time step of the simulation and interact with individual geometries to explore the simulation’s detailed results.