Introduction
Internet of Things (IoT) solutions rest on the extensive use of sensors, and the deployment of IoT applications to make cities "smart" is no exception. Urban growth and infrastructure replacement cycles provide municipalities the opportunity to introduce new IoT technologies to replace aging and costly equipment not only for cost savings, but also to take advantage of improved quality of life for citizens with a concomitant increase in desirability and revenue for the city.
The promised benefits are enticing, but the realization of those benefits is a challenge, due in part to the reality that the primary driver for design and implementation is a narrowly defined concept of cost. A deeper understanding of comparative costs and benefits is needed to facilitate successful deployment of Smart Cities IoT projects.
Additionally, the wealth and variety of benefits from potential Smart Cities technologies, coupled with the diversity of technical solutions, complicate the evaluation process. Municipalities must target benefits that provide the highest return on investment (ROI) and to evaluate the technical solutions that most effectively enable these benefits. The proposed Smart Cities Decision Aid (SCDA) framework enables municipalities to effectively pursue Smart Cities deployments in lieu of the current ad hoc, cost-focused process.
Industry Needs
Municipalities need a methodology to more fully evaluate and manage the complexities of sensor applications in Smart Cities IoT projects. The methodology should support an analysis of alternatives and interactions between alternatives to model deployments. The methodology should not just model deployment costs, but also model how solutions impact perceived needs, culture, and quality of life for constituents of a municipality.
These additional factors facilitate more comprehensive measures of effectiveness for project proposals, providing a more complete cost evaluation than the current process of evaluating proposals based mainly on deployment cost. The new measures of effectiveness should include reductions in societal costs and other societal benefits achieved from deployment of technology solutions to capture a more complete representation of overall utility.
The proposed SCDA framework offers a way for municipalities to consistently define objectives and preferences for Smart Cities IoT projects, then translate those inputs to a common measure of utility for consistent and complete proposal comparison. The framework enables focused evaluation of feasible technology solutions that align with stated project goals through a model-based relationship between cost, benefit, and utility.
Additionally, the framework enables a more targeted and complete exploration into societal costs and benefits. These societal costs and benefits are more difficult for municipalities to quantify and consequently are often overlooked in project decisions.
Proposed Process for IoT Decisions within a Smart City Environment
A new approach for decision making is needed to address challenges municipalities face in performing three critical tasks related to Smart Cities IoT projects:
1) Understand preferences to focus research on target benefits and technical solutions.
2) Limit model development scope to variables with most impact on project goals.
3) Understand and quantify the full scope of costs and benefits for project proposals.
Next page
Each task should be designed and orchestrated to focus research efforts on IoT project elements that have the most impact on achieving project goals for the municipality. A major problem faced by municipalities is the overwhelming number of decisions involved in evaluating the technical solutions needed to fulfill the IoT-shaped project goals.
By focusing design decisions on related variables that have the most impact on the final expected utility of a project, a municipality can concentrate limited resources on the elements that provide the most impact. This approach offers a more quantifiable way to evaluate solutions that are more aligned with preferences and project goals, because the impact of the design on project goals is more fully understood.
Process Flow
The functional flow presented as figure 1 is a model of the proposed SCDA framework.
Fig. 1: SCDA Functional Flow.
Prior to evaluation of any technical solutions, it is imperative to understand the objectives, constraints, and preferences of the municipality. This information is utilized to identify possible target benefits that are best aligned to produce the desired goals for the IoT project. Once the target benefits are identified, the challenges involved with achieving those benefits are derived from similar reference projects and expanded into possible technology solutions to meet the desired goals.
Instead of allowing the latest technology solutions to dictate goals of the project, the SCDA approach narrows the scope of possible technology solutions, which need to be evaluated. Systems engineering processes can then be applied on a narrower subset of technology solutions, therefore reducing the project scope and increasing the probability that the final solution aligns with stated project goals.
After identifying possible technology solutions in line with municipality goals, deeper analysis of technology solutions is performed through Model Based Systems Engineering. Exploration into benefits achieved and challenges presented by a technology solution provides data to understand the most impactful elements on project goals. To increase the opportunities for success, a project should focus on technology solutions and challenges that have the most impact on achieving the defined target benefits of the municipality. The SCDA framework provides municipalities a methodology to create more properly focused and complete models to increase the opportunity of successful Smart Cities IoT project deployments.
Next page
Model Framework
Model Based Systems Engineering is a powerful process for evaluating technical solutions, but expanding the scope of models to include every design variable of a system can quickly overwhelm the usefulness of any modeling effort. The increased scope drives up complexity and cost exponentially, while simultaneously reducing the usefulness any information derived.
The SCDA approach defines measures of effectiveness for models, based on identified target benefits and challenges, using a common measure of utility. This common measure of utility is expressed in financial terms, so a municipality can quickly understand and maximize the expected utility of a technical solution.
The SCDA framework is based upon an adapted version of Hazelrigg's Rational Design Framework , presented by Lee and Paredis as a Value-driven Design Framework. Initial efforts to focus technology solutions around the objectives, constraints, and preferences of the municipality serve as inputs to models. The models are organized as layers, with each layer optimized individually, and the results of inner layers feed into the optimization of outer layers.
This layered approach allows smaller, more focused models of specific solutions to aggregate into larger models that include more complex configurations and combinations of specific technical solutions in a large project deployment. The chart below illustrates the model layers with specific applied systems engineering methodologies (figure 2).
Fig. 2: Model layers with specific applied systems engineering methodologies.
The separation of complex logic into smaller components is not a new concept to software development, but it is a novel approach for modeling Smart Cities IoT projects. Each layer of the model is designed and validated separately, which provides a tremendous reduction in complexity as compared to modeling the entire solution monolithically.
The Benefits layer utilizes the previously identified target benefits and associated challenges, outside of influence from any specific technical solutions. The purpose of this modeling layer is to understand completely the range of costs and benefits associated with achieving a specific benefit offered by an IoT solution. The current societal and municipality costs are compared to the proposed benefits of the new solution to understand what level of performance is needed to optimize the expected utility from a specific benefit.
The Benefit layer aligns with the objectives, constraints, and preferences of a municipality to provide a quantifiable way to optimize expected utility from a specific benefit offered by the concept of a particular IoT solution. This optimization process allows municipalities to rationally determine which target benefits best align with project goals and provide the maximum expected utility.
After target benefits are identified in the Benefits layer of the model, the Design layer is utilized to understand the technical solutions that support achieving the target benefits. Specific technical solutions are evaluated with cost-effectiveness models to understand the cost required to meet a desired level of performance.
While the Benefits layer of the model is optimized based on the cost-benefit analysis of specific target benefits, the Design layer of the model is optimized based on the ability of a specific design to meet the performance requirements of the target benefits analyzed in the inner Benefits layer of the model. This approach allows a more complete understanding of the costs to achieve a performance target defined by the Benefits layer, which creates a feedback loop for more complete evaluation of proposed target benefits.
The Configuration layer aggregates multiple target benefits, and multiple technical designs aimed at supporting those target benefits, into an aggregated deployment configuration. Models of the Configuration layer explore the deployment densities, technology combinations, and configurations of different technology solutions to optimize overall expected utility for the project.
Without this layered approach, a full evaluation of benefits aligned to preferences, technical solutions aligned to benefits, and configurations aligned to technical solutions would not be feasible for a municipality. The complexity of creating a monolithic model to capture and coordinate many different modeling aspects could possibly overwhelm a municipality and lead to project decisions misaligned with project goals, therefore putting project success at risk.
Next page
Details And Benefits Of Proposed IoT Approach
Total Sensor Cost is determined by relating annual operational costs to an Equivalent Annual Cost. This is multiplied by the sensor's expected lifetime and combined with one-time total lifecycle costs. These calculations should include weightings to account for uncertainties and risk.
Societal Costs, which would be affected by the sensor installation should be identified. These may include such factors as legal personnel, lost productivity, insurance costs, personal injury, fire station density, taxes, property prices, privacy, traffic or health. If gunshot detectors are shown to reduce crime, then it is logical to assume that other societal costs related to crime will also be reduced. This reduction in societal costs from a known baseline is what the SCDA program refers to as the Societal Benefit. The cost savings can be addressed as a benefit in the overall evaluation of how a technology deployment will provide benefit to a municipality.
The proposed modeling method results in higher financial returns on Smart Cities sensor installations than the immediate ROI would indicate. Sensors are used to affect a community's health in a variety of very positive ways.
To realize the total effect of a large sensor installation, the Total Cost is determined by summing the Sensor Cost and the Society Benefit. The final model is the Cost-Benefit Analysis (CBA) model. This model pulls together the previous cost models and benefit models into one final evaluation of total utility as the benefit of various technology solutions in financial terms.
Within the context of a specific target smart cities use case, the modeled utility can be optimized in the benefit layer to understand which related benefits provide the most utility. The utility can also be optimized at the design layer, to understand how different design variables affect performance and cost of a technology solution, which ultimately influences overall utility. It is possible to perform a CBA on each technological solution, or to abstract to an outer model layer and perform a CBA on a combination of technology solutions constructed in a configuration. Different configurations of technologies and deployment options may yield different overall utilities, based on the technology interactions and compatibilities.
Optimizations at the benefit, solution design, and configuration levels are performed through simulations that explore the trade space and account for uncertainty elements in the various models. The optimized solution is then evaluated to understand the overall utility gained, which provides the overall benefit a municipality will receive for the project, with respect to cost to achieve that project. Maximizing that value will allow the municipality to identify the best project options for selection.
Several software products are available to develop supporting models and execute simulations. In a reference project the SCDA team used Phoenix Integration's ModelCenter optimizing algorithm to converge on an optimal solution by seeking the minimum Total Cost for the evaluated sensor configuration.
ModelCenter returned an exact number of gunshot detectors, cameras, RFID sensors, and LEDs to optimize utility. These values represent the point where additional sensors are not worth their cost. While additional gunshot detectors per block may marginally reduce crime, Society Benefit integrated with Sensor Cost can show the diminishing returns economically. The final numbers are then compared to third party data for validation of the proposed solution.
Conclusion
The SCDA framework presents a methodology for municipalities to evaluate potential large scale Smart Cities IoT sensor deployments based on optimized utility. A consistent process for identification and focused evaluation of technical elements that are most aligned to preferences is a key benefit of the SCDA framework.
The layered model approach allows for a more integrated and comprehensive evaluation, while at the same time reducing complexity and focusing on the elements with the greatest impact on project success. The SCDA framework integrates societal costs and benefits in a quantifiable way in contrast to standard deployment costs. This more complete representation of costs and benefits is critical to understand fully the expected utility, and eventually the attractiveness, of a proposed SC IoT project.
About the Author
This article summarizes a Capstone project, centered on the City of Atlanta's Smart Pole Initiative, realized under the direction of Dr. Margaret Loper, Chief Technologist of the Georgia Tech Center for the Development and Application of Internet of Things Technologies (CDAIT), within the Professional Master's Degree in Applied Systems Engineering (PMASE) program at Georgia Tech. The full graduate team participating in the project included authors Christopher Byrd, Ryan O'Keefe, and Cheryl Emerson.