Download PDF Spring Bridge on Sustainable Smart Cities March 15, 2023 The world’s cities face increasing threats from natural disasters, aging infrastructure, traffic, and resource constraints. The articles in this issue examine smart infrastructure, sustainability, net zero carbon options, and autonomous driving, among other approaches to smart and sustainable cities. Developing Humanoid Architectural Structures for a Resilient City Friday, March 17, 2023 Author: Xiangsheng Chen, Changqing Xia, Hongzhi Cui, Chengyu Hong, and Min Zhu Advanced materials and intelligent technologies are enabling urban structures that are as resilient, flexible, aware, and self-healing as the human body. Urban resilience has been defined as “the measurable ability of any urban system, with its inhabitants, to maintain continuity through all shocks and stresses, while positively adapting and transforming toward sustainability.” The concept aims to ensure a safe environment for city dwellers, with architectural structures that can endure disasters and recover quickly. A humanoid architectural structure (HAS) synergizes advanced materials, structure, intelligent technology, and real-time holographic sensing and warnings to support such a safe urban environment and contribute to urban resilience. The following “human” characteristics distinguish this new resilient construction: a robust and restorable “body” with powerful and flexible capacities for external load resistance, self-healing, and self-resetting; acute “sensory” perception, enabled by ubiquitous and wide-area self-sensing; an “intelligent brain” with the ability to self-diagnose, make decisions, and learn; and strong “immunity” (i.e., self-protective capability) (figure 1). Robust and Restorable Body Materials and structural engineering research developments enable a physically robust HAS body, and a novel design paradigm integrating material and structure has economic advantages. High-performance fiber-reinforced cement-based composites—such as steel, carbon, ultrastrong tensile botany (Tang et al. 2021), and two-dimensional polymeric fibers (Zeng et al. 2022)—exhibit superior mechanical properties and aging resistance, and they are gaining traction as potential applications in complex and harsh catastrophic environments. When using such novel materials for structural design, the evolution of structure modes (e.g., deformation, damage, and destruction) and critical state health threshold values can be determined by developing a digital twin (DT) (Tao et al. 2019). Use of a DT facilitates alteration and optimization of material characteristics (e.g., quantity and kind of fibers in different positions) and/or structural style through scenario simulations and iterative computation. A significant advantage of the digital twin is that designers may quickly assess and forecast the behavior of structures without conducting extensive research. Another essential feature of the HAS body is its restorative ability. Self-healing technology for concrete is being researched in depth. When tiny cracks or holes in the material occur, self-healing is activated through autonomous or autogenic forms, using either a substance released via internal premade microcapsules or hydration with microorganisms (e.g., bacilli, fungi) that can create calcium carbonate to fill microscopic cracks (Brasileiro et al. 2021). Several novel and more efficient self-healing techniques have been inspired by some biochemical reactions in cells. For example, the enzyme carbonic anhydrase, which is found in red blood cells, catalyzes the reaction between Ca2+ ions and atmospheric CO2 to create calcium carbonate crystals, healing millimeter-scale cracks within 24 hours (Rosewitz et al. 2021). And shape memory alloys and self-centering joints are being studied to enable conversion of a rigid structure to a flexible one and thus increase seismic performance in terms of both rapid deformation recovery and energy consumption (Movaghati and Abdelnaby 2021). In cases of extreme damage or destruction, these technologies will help extend the life of an architectural structure. However, they are insufficient for all circumstances, so HAS needs advanced capabilities for sensing, diagnosing, decision making, response, and learning. Acute Sensing System The next generation of technology in construction and maintenance is characterized by the fusion technologies of Internet of Things (IoT)–based sensors and terminals, edge computing, cloud computing, and 5G wireless technology (Dai et al. 2020). These technologies, when used, for example, in underground construction equipment (e.g., a shield machine), enable automated operation and the rapid synthesis of smart-monitored information (Armaghani and Azizi 2021). Multielement information—about construction activities, subsurface disturbance, ground movement, water level change, and potential impacts of multisource disasters—is monitored using reliable IoT sensors such as microelectromechanical systems, fiber-optic sensors, and machine vision, all of which are low-power, low-cost, wireless, and autonomous. Disasters can be efficiently managed by a system’s integrated capacity for sensing, diagnosis, decision making, response, and learning. For operation and maintenance, holographic sensing provides a function comparable to human sensory organs. Ubiquitous wireless sensors embedded in architectural structures can detect stress, strain, cracks, temperature, humidity, and even ionic concentration with high accuracy (Sofi et al. 2022). In addition, materials such as sensing skins (based on electrical capacitance) and self-sensing concrete (based on electrical resistance) are highly anticipated (Bekzhanova et al. 2021). In particular, sensing skins, such as printable conductive polymer and soft elastomeric capacitive sensors, are being studied for 3D printing on the surface of architectural structures (Laflamme and Ubertini 2020). They could convert stress, strain, and cracks into measurable or observable changes via an electric signal. The Smart Brain and Nervous System Massive data collected by advanced sensors and actuators in the IoT reveal quantitative variation in structures and circumstances. Data from holographic sensing are used both to construct a knowledge graph in the HAS brain and to analyze and update the DT model. Cloud computing is a well-established paradigm for the “brain,” but colossal data, photos, and videos create a significant transfer and computation burden on the limited-capacity internet and even the cloud. Fifth-generation (5G) wireless technology is a good alternative to improve massive data transfer efficiency (Eid et al. 2021). The cloud computing platform will be decentralized and replaced by distributed edge computing (near the end of IoT nodes) to reduce information delay (Fraga-Lamas et al. 2021; Ren et al. 2019). Microcomputers at the edge will evaluate structural mode data in real time (Chen 2018) and provide early warning with analysis and diagnosis. In the event of an emergency, the actuators or triggers will receive immediate autonomously generated commands from the microcomputers (Bai and Scholl 2021). Subsequently, big data streams from the edge will aid in dynamic interaction between the cyber and physical worlds at the cloud layer. DT models will be updated to measure bearing capacity, evaluate resilience, predict imminent mode variation, and perform necessary repairs or replacements in the cyber world to serve as a reference for actual repairs (Jiang et al. 2021). Artificial intelligence (AI) in support of the IoT and big data can be used to enhance decision making about infrastructure maintenance (Zhu et al. 2020). Machine learning (ML) algorithms are critical to AI, whose applications range from robotics and computer vision to autonomous vehicle control and neuroscience research (Jordan and Mitchell 2015). Supervised learning, one type of ML, is used for the detection of infrastructure cracks based on computer vision technology. Deep learning, part of a broader family of ML algorithms, is normally underpinned by artificial neural networks (ANN). A big data center linked with ANN (through a website, software, and mobile app) is required for effective data processing and analysis. The center should be highly integrated and provide relatively high availability, cost savings, and strong adaptability. Such a center can automatically respond through ANN programmed and assigned to DT model components (Huang et al. 2021). Machine learning based on optical neural networks (the physical implementation of ANN with optical components) could aid in systematic analysis from multiple perspectives, dimensions, and granularities. It can thus be used to predict variation of structure modes and circumstances, and create knowledge for scenario innovation (Wang et al. 2022). In summary, AI technologies can enable automated solutions and support decision making and policies for infrastructure maintenance to produce a “nervous system” with the collaborative integration of local and global computing. Immune System Humanoid architectural structures have extraordinary self-defense capacities, similar to the human immune system based on the sensing and nervous systems. For example, a servo system has been used to reduce lateral deformation of the retaining structure during foundation pit excavation, and it can be used in architectural structures to control deformation and settlement. Force and displacement are continuously monitored using pressure sensors and ultrasonic displacement sensors. The system compensates automatically as soon as it detects that the force is less than the designed value, and automatically reduces the force if it is higher. An alert is triggered if the monitoring data exceed the warning range. In the event of a disaster such as a flood, if a sensor detects stagnant water at a certain level, automated floodgates will be activated promptly and water that has entered an underground space will be directed to drainage or storage. Other calamities, such as a poison gas leak, fire, or explosion, can be efficiently managed by the system’s integrated capacity for sensing, diagnosis, decision making, response, and learning. Putting It All Together Infrastructure construction and operation can benefit from the advanced materials, structure, and HAS intelligent technology. Figure 2 shows an underground shield tunnel with rigid-flexible structures and high--performance, self-healing materials that improve robustness and restorability. With sensing, nervous, and immune systems, the tunnel is capable of self-diagnosis, decision making, response, and learning. Areas of Needed Progress Significant efforts are needed in a variety of areas: Upgrades: Hardware and software both need to be significantly enhanced, based on research and development of new materials with higher performance, new structural forms, new long-life sensors, an intelligent structural system, and the establishment of design theory. Digitization and data flow: These are needed to realize real-time multisource heterogeneous data capture and linkage with physical space to achieve -bidirectional mapping, dynamic blending, and real-time coupling. Cognification: Intelligent control and learning evolution rely on explicit knowledge that can provide human-understandable explanations to learning results (Omran et al. 2019), programming and software based on knowledge graphs, and cloud computing and machine learning based on next-generation neural network algorithms. Integration: Multiple elements combine with and integrate innovation, during which HAS components are assembled and debugged. Interdisciplinary technologies, such as those spanning civil engineering, next-generation IT, microelectronics, measurement, and control, among others, should be prioritized. Conclusion Humanoid architectural structures, endowed with the advanced technologies of digital twins, the IoT, big data, and AI, will be able to monitor themselves, provide early warning of potential risks, and withstand external disturbance using automatic and smart immune systems for risk mitigation. They will enable urban buildings and infrastructure for various systems to sense external disturbances through different smart sensors (like human senses). Information and communication technologies that mimic the human nervous system transmit and integrate all sensor information, while big data centers and AI, acting as the brain, process multisource signals. High-performance materials and the flexible structure of infrastructures (comparable to the human skeleton and muscles) actively protect against infrastructure disasters and failures. The performance of city infrastructures will thus change from passive to reactive and then to proactive. Although some HAS components, such as the digital twin, are nearing application maturity, the vast majority, including self-sensing materials and optical neural networks, remain in the research stage. With continued development and improvements in materials research and structural design, it will be possible to scale the HAS economically and technically. We expect that the integration of HAS features in buildings, underground infrastructures, bridges, avenues, and other constructions that serve human habitation can be achieved within one or two decades. Humanoid architectural structure has the potential to revolutionize infrastructure and make a city more resilient and secure during its entire lifespan. Acknowledgments We thank Mingwei Hu, Tian Xie, Wei Liu, Haiyang Zhou, Xuetao Wang, Qian Zhao, Weiyi Zhang, and Liaohan Xie for many helpful discussions and suggestions. This research is supported by the Shenzhen Science and Technology Program (KQTD20200909113951005), National Natural Science Foundation of China (51938008, 52090084, 52108329, L1924061), and China Post-doctoral Science Foundation (2021T140475). References Armaghani DJ, Azizi A. 2021. Applications of Artificial Intelligence in Tunnelling and Underground Space Technology. Springer. Bai H, Scholl B. 2021. Edge Computing and Capability--Oriented Architecture. CRC Press. Bekzhanova Z, Memon SA, Kim JR. 2021. Self-sensing cementitious composites: Review and perspective. Nanomaterials 11(9):2355. Brasileiro PPF, Brandão YB, Sarubbo LA, Benachour M. 2021. Self-healing concrete: Background, development, and market prospects. Biointerface Research in Applied Chemistry 11:14709–25. Chen X. 2018. Research on combined construction technology for cross-subway tunnels in underground spaces. Engineering 4(1):103–11. Dai Y, Zhang K, Maharjan S, Zhang Y. 2020. Edge intelligence for energy-efficient computation offloading and resource allocation in 5G beyond. IEEE Transactions on Vehicular Technology 69(10):12175–86. Eid A, Hester JG, Tentzeris MM. 2021. 5G as a wireless power grid. Scientific Reports 11(1):1–9. Fraga-Lamas P, Lopes SI, Fernández-Caramés TM. 2021. Green IoT and edge AI as key technological enablers for a sustainable digital transition towards a smart circular economy: An industry 5.0 use case. Sensors 21(17):5745. Huang MQ, Ninić J, Zhang QB. 2021. BIM, machine learning and computer vision techniques in underground construction: Current status and future perspectives. Tunnelling & Underground Space Technology 108:103677. Jiang F, Ma L, Broyd T, Chen K. 2021. Digital twin and its implementations in the civil engineering sector. Automation in Construction 130:103838. Jordan MI, Mitchell TM. 2015. Machine learning: Trends, perspectives, and prospects. Science 349(6245):255–60. Kelly K. 2016. The Inevitable: Understanding the 12 Technological Forces That Will Shape Our Future. Penguin. Laflamme S, Ubertini F. 2020. Back-to-basics: Self-sensing materials for nondestructive evaluation. Materials Evaluation 78:526–36. Movaghati S, Abdelnaby AE. 2021. Experimental and analytical investigation of improved behavior of existing steel connections by adding self-centering capabilities. Building Engineering 43:102543. Omran PG, Wang K, Wang Z. 2019. An embedding-based approach to rule learning in knowledge graphs. IEEE Transactions on Knowledge & Data Engineering 33(4):1348–59. Ren J, Zhang D, He S, Zhang Y, Li T. 2019. A survey on end-edge-cloud orchestrated network computing paradigms: Transparent computing, mobile edge computing, fog computing, and cloudlet. ACM Computing Surveys 52(6):1–36. Rosewitz JA, Wang S, Scarlata SF, Rahbar N. 2021. An enzymatic self-healing cementitious material. Applied Materials Today 23:101035. Sofi A, Regita JJ, Rane B, Lau HH. 2022. Structural health monitoring using wireless smart sensor network: An overview. Mechanical Systems & Signal Processing 163:108113. Tang Q, Zou M, Wang S, Fang L, Guo W. 2021. Processing composites reinforced with wood fibers into an ultra-strong structural materials. Polymer Composites 42(6):2872–81. Tao F, Sui F, Liu A, Qi Q, Zhang M, Song B, Guo Z, Lu SY, Nee AYC. 2019. Digital twin-driven product design framework. International Journal of Production Research 57(1):3935–53. Wang T, Ma SY, Wright LG, Onodera T, Richard BC, -McMahon PL. 2022. An optical neural network using less than 1 photon per multiplication. Nature Communications 13(1):1–8. Zeng Y, Gordiichuk P, Ichihara T, Zhang G, Sandoz-Rosado E, Wetzel ED, Tresback J, Yang J, Kozawa D, Yang Z, & 11 others. 2022. Irreversible synthesis of an ultrastrong two-dimensional polymeric material. Nature 602(7895):91–95. Zhu S, Bao H, Zhao Z. 2020. White paper on intelligent highway prospect and technology development. Zhejiang Lab Research Report. Available from the author.  UN-Habitat, What Is Urban Resilience?, https://unhabitat.org/topic/resilience-and-risk-reduction  Cognification is defined as the process of making objects smarter by integrating AI-endowed sensors. See Kelly (2016). About the Author:Xiangsheng Chen (CAE) is a professor, Changqing Xia a postdoc, Hongzhi Cui a professor, Chengyu Hong an associate professor, and Min Zhu an associate research fellow, all in the College of Civil and Transportation Engineering, Shenzhen University.