Special issue on security, usability and sustainability of smart cities

Jia Hu, Hissam Tawfik, Jia Hu 0001, Kevin Kam Fung Yuen, Mohammed M. Alani, Nektarios Georgalas

Introduction

The transition of urban centers into "smart cities" relies on the deployment of intelligent environments—systems that synthesize real-time and historical data to facilitate autonomous decision-making. While these systems promise enhanced service efficiency and resource management, they also introduce significant risks. Traditional urban infrastructure was not designed to meet the rigorous performance, reliability, and security requirements of interconnected intelligent systems. This special issue presents multi-disciplinary research addressing the critical need for technologies that are secure against cyber threats, usable for citizens, and sustainable for long-term urban growth.

Core Research Contributions

1. Lightweight Security for IoT Healthcare

In the context of smart city healthcare, data privacy is paramount. However, the high computational overhead of traditional security protocols often exceeds the capabilities of resource-constrained IoT devices. Research in this issue proposes a lightweight mutual authentication scheme specifically designed for IoT-based healthcare. By balancing high-level security with device performance, this scheme provides a viable path for protecting patient data in real-time monitoring environments.

2. Automated Verification of Robotic Autonomy

As autonomous robots transition from controlled laboratory settings to complex smart buildings, ensuring their reliability over long durations is a significant challenge. The TestIt tool addresses this by providing automated model-based testing for multi-robot systems. A case study involving smart building security demonstrated that TestIt can effectively verify robotic autonomy, ensuring these machines can operate independently without human intervention.

3. Achieving Determinism in Deep Learning

A major barrier to the adoption of AI in safety-critical smart city roles is the "black box" nature of standard deep learning. Stochastic weight initialization leads to unpredictable outcomes, which complicates public liability and functional safety. To address this, researchers explored non-random weight initialization schemes. By moving toward deterministic learning sessions, deep learning models become more predictable, meeting the legal and functional requirements of urban safety systems.

4. AI-Driven Environmental Monitoring

Sustainability in smart cities can be enhanced by repurposing existing infrastructure. One study utilizes a hybrid deep learning approach combined with image cropping to transform standard IoT security cameras into environmental monitoring tools. By classifying images of urban drainage systems (gullies and drains), the system provides a scalable method for flood prevention and infrastructure maintenance.

Key Insights and Results

The research presented highlights several critical themes:

  • The Determinism Requirement: Repeatable determinism is essential for AI applications where safety and liability are paramount.
  • Efficiency-Security Trade-offs: Lightweight protocols are necessary for the practical deployment of security in the IoT ecosystem.
  • Cross-Domain Utility: Software-driven repurposing of hardware (e.g., security cameras for flood monitoring) offers a cost-effective path to urban sustainability.

Evaluation of these methods across various contexts—including smart building case studies and comparative security analyses—confirms their efficacy. For instance, the proposed healthcare authentication scheme outperformed existing models in efficiency, and the TestIt tool successfully validated robotic autonomy in real-world scenarios.

Future Directions

Despite these advancements, scaling these solutions to metropolitan levels remains a significant hurdle. Future work must focus on the interoperability of intelligent systems with legacy urban infrastructure. Furthermore, as autonomous systems become more prevalent, the development of robust legal and ethical frameworks will be necessary to address the complexities of public liability in the smart city.

Summary updated

BibTeX
@article{hu2020special,
  author = {Jia Hu and Hissam Tawfik and Jia Hu and Kevin Kam Fung Yuen and Mohammed M. Alani and Nektarios Georgalas},
  title = {{Special issue on security, usability and sustainability of smart cities}},
  journal = {J. Reliab. Intell. Environ},
  year = {2020},
  doi = {10.1007/s40860-020-00101-6},
}
Figure