Artificial Intelligence, Engineering Systems and Sustainable Development

In this chapter, a general introduction on artificial intelligence (AI) is given as well as an overview of the advances of AI in different engineering disciplines, including its effectiveness in driving the United Nations Sustainable Development Goals (UN SDGs). This chapter begins with some fundamental definitions and concepts on AI and machine learning (ML) followed by a classification of the different categories of ML algorithms. After that, a general overview of the impact which different engineering disciplines such as Civil, Chemical, Mechanical, Electrical and Telecommunications Engineering have on the UN SDGs is given. The application of AI and ML to enhance the processes in these different engineering disciplines is also briefly explained. This chapter concludes with a brief description of the UN SDGs and how AI can positively impact the attainment of these goals by the target year of 2030.

Currently, Mauritius is adopting landfilling as the main waste management method, which makes the waste sector the second biggest emitter of greenhouse gas (GHG) in the country. This presents a challenge for the island to attain its commitments to reduce its GHG emissions to 30% by 2030 to cater for SDG 13 (Climate Action). Moreover, issues like eyesores caused by littering and overflowing of bins and low recycling rates due to low levels of waste segregation are adding to the obstacles for Mauritius to attain other SDGs like SDG 11 (Make Cities & Human Settlements Inclusive, Safe, Resilient & Sustainable) and SDG 12 (Guarantee Sustainable Consumption & Production Patterns). Therefore, together with an optimisation of waste collection, transportation and sorting processes, it is important to establish a solid waste characterisation to determine more sustainable waste management options for Mauritius to divert waste from the landfill. However, traditional waste characterisation is time consuming and costly. Thus, this chapter consists of looking at the feasibility of adopting machine learning to forecast the solid waste characteristics and to improve the solid waste management processes as per the concept of smart waste management for the island of Mauritius in line with reducing the current challenges being faced to attain SDGs 11, 12 and 13.

There is an urgent need to develop climate-smart agrosystems capable of mitigating climate change and adapting to its effects. Conventional agricultural practices prevail in Mauritius, whereby synthetic chemical fertilizers, pesticides and insecticides are used. It should be noted that Mauritius remains a net-food importing developing country of staple food such as cereals and products, roots and tubers, pulses, oil crops, vegetables, fruits and meat (FAO, 2011). In Mauritius, the agricultural sector faces extreme weather conditions like drought or heavy rainfall. Moreover, to increase the crop yields, farmers tend to use 2.5 times the prescribed amount of fertilizers in their fields. These excess fertilizers are washed away during heavy rainfall and contaminate lakes and river waters. By using smart irrigation and fertilization system, a better management of soil water reserves for improved agricultural production can be implemented. Soil Nitrogen, Phosphorus and Potassium (NPK) content, humidity, pH, conductivity and moisture data can be monitored through the cloud platform. The data will be processed at the level of the cloud and an appropriate mix of NPK and irrigation will be used to optimise the growth of the crops. Machine learning algorithms will be used for the control of the land drainage, fertilization and irrigation systems and real time data will be available through a mobile application for the whole system. This will contribute towards the Sustainable Development Goals (SDGs): 2 (Zero Hunger), 11 (Sustainable cities and communities), 12 (Responsible consumption and production) and 15 (Life on Land). With this project, the yield of crops will be boosted, thus reducing the hunger rate (SDG 2). On top of that, this will encourage farmers to collect the waters and reduce fertilizer consumption thereafter sustaining the quality of the soil on which they are cultivating the crops, thereby increasing their yields (SDG 15).

Mauritius is a Small Island Development State (SIDS) with limited resources, and it has been witnessed that many containers used for storing household and industrial products are made from plastic. When discarded as waste, those plastic containers pose a serious environmental and economic challenge for Mauritius. Moreover, landfill space is getting increasingly scarce, and plastic waste is contaminating both land and water. Therefore, it is of the utmost necessity to develop solutions for Mauritius' plastic wastes. Due to its abundance and accessibility, plastic waste is a promising material for recycling and energy production. One potential solution is the use of machine learning and artificial intelligence (AI) to predict household plastic consumption, allowing policymakers to design effective strategies and initiatives to reduce plastic waste. Such information is a critical component to be able to efficiently plan for the collection and routing of trucks when collecting recyclable plastics. The development of new strategies for the recycling of plastic waste and development of new industry can address the import and export potential of the country to achieve self-sustainability as well as contribute to reduction in plastic pollution and amount of waste landfilled. These plastics can thereafter be used effectively for recycling and for the making of 3D printing filaments which fall under the SDGs 9 (Industry, Innovation and Infrastructure) and 12 (Responsible consumption and production).

Adsorption parameters (e.g. Langmuir constant, mass transfer coefficient and Thomas rate constant) are involved in the design of aqueous-media adsorption treatment units. However, the classic approach to estimating such parameters is perceived to be imprecise. Herein, the essential features and performances of the ant colony, bee colony and elephant herd optimisation approaches are introduced to the experimental chemist and chemical engineer engaged in adsorption research for aqueous systems. Key research and development directions, believed to harness these algorithms for real-scale water treatment (which falls within the wide-ranging coverage of the Sustainable Development Goal 6 (SDG 6) ‘Clean Water and Sanitation for All’), are also proposed. The ant colony, bee colony and elephant herd optimisations have higher precision and accuracy, and are particularly efficient in finding the global optimum solution. It is hoped that the discussions can stimulate both the experimental chemist and chemical engineer to delineate the progress achieved so far and collaborate further to devise strategies for integrating these intelligent optimisations in the design and operation of real multicomponent multi-complexity adsorption systems for water purification.

It is now well-established that good water quality is associated with economic prosperity, reduced incidence on public health and the good functioning of the various ecosystems found in our environment. Water contamination is mostly related to both diffused (agricultural lands and geologic rock degradations) and point sources of pollution. Mauritius has many water resources which depend solely on precipitation for their replenishment. Water parameters which are of relevance include total dissolved solids (TDS), temperature, pH, electrical conductivity, turbidity, dissolved oxygen, dissolved and particulate organic carbon and major cations and anions. The traditional methods of analysis for these parameters are mostly using electrical and optical methods (probes and sensors in the field), while chemical titrations, Flame AAS and High-Performance Liquid Chromatography techniques are carried out in the laboratory. Image Classification techniques using neural networks can also be used to detect the presence of contaminants in water. In addition to basic water quality parameters, the field sensors range have been extended to cover important major ions and can now be integrated with Artificial Intelligence (AI)-based models for the prediction of variations in water quality to better protect human health and the environment, reduce operation costs of water and wastewater treatment plant unit processes.

This chapter provides an overview of the potential use of Intelligent Transport Systems (ITS) and associated artificial intelligence (AI) techniques in the land transport sector in an attempt to achieve related United Nations Sustainable Development Goals (SDGs) targets. ITS applications that have now been extensively tested worldwide and have become part of the everyday transport toolkit available to practitioners have been discussed. AI techniques applied successfully in specific ITS applications such as automatic traffic control systems, real-time image processing, automatic incident detection, safety management, road condition assessment, asset management and traffic enforcement systems have been identified. These methods have helped to provide traffic engineers and transport planners with novel ways to improve safety, mobility, accessibility and efficiency in the sector and thus move closer to achieving the various SDG targets pertaining to transportation.

Energy production and distribution is undergoing a revolutionary transition with the advent of disruptive technologies such as the Internet of Energy (IoE), 5G and artificial intelligence (AI). IoE essentially involves automating and enhancing the energy infrastructure: the power grid from grid operators to energy generators and distribution utilities. The IoE also relies on powerful connectivity networks such as 5G, big data analytics and AI to optimise its operation. By incorporating the technology that employs ubiquitous devices such as smartphones, tablets or smart electric vehicles, it will be possible to fully exploit the potential of IoE using 5G networks. 5G networks will provide high speed connections between devices such as drones, tractors and cloud networks, to transfer huge amounts of sensor data. Additionally, there are many sources of isolated data across the main energy production units (generation, transmission and distribution), and the data is increasing at phenomenal rates. By applying AI to these data, major improvements can be brought at each stage of the energy production chain. Tying renewable energy to the telecommunications sector and leveraging on the potential of data analytics is something which is gaining major attention among researchers and industry experts. This chapter therefore explores the combination of three of the most promising technologies i.e. IoE, 5G and AI for achieving affordable and clean energy, which is SDG 7 in the UN Sustainable Development Goals (SDGs).

Intelligent real-time systems are significantly impacting several of the UN Sustainable Development Goals (SDGs) by revolutionising processes in several areas such as Industry 4.0, smart cities, transportation, agriculture, renewable energy, climate change and other economic activities. Given that much of the work to achieve the SDGs relies on information and communication technology, cybersecurity has a potentially immense role to play towards achieving these outcomes. Moreover, cyberattacks have emerged as a new functional threat for interconnected, smart manufacturers and digital supply networks, employed in intelligent real-time systems for the Fourth Industrial Revolution. The effects of cyberattacks can be much more widespread than ever before due to the interconnected nature of Industry 4.0-driven operations. Blockchain can be really useful in such situations as it provides edge protection and allows authentication of the machine-to-machine and human–machine operations, stable data share, life cycle management, access control compliance of devices and self-sustaining operations. Moreover, blockchain can be applied for tracking and tracing transactions through devices, which are performed during the operation, as well as to encrypt and transmit data securely. It is vital to establish complete trust in a technology that is being adopted so that its full potential can be exploited. It is consequently critical that the organisational and information technology strategy fully integrates secure, vigilant and resilient cybersecurity strategies such as blockchain. This will ensure that cyber risks are properly managed in the age of Industry 4.0. This chapter, therefore, analyses the application of blockchain in intelligent real-time systems such as Industry 4.0 so that the opportunities these systems present for the SDGs can be exploited safely with minimum risks to society.

To operate with high efficiency and minimise the risks of power failures, power systems require careful monitoring. The availability of real-time data is crucial for assessing the performance of the grid and assisting operators in gauging the present security of the grid. Traditional supervisory control and data acquisition (SCADA)-based systems actually employed provides steady-state measurement values which are the calculation premise of State Estimation. More often, however, the power grid operates under dynamic state and SCADA measurements can lead to erroneous and inaccurate calculation results. The introduction of the phasor measurement unit (PMU) which provides real-time synchronised voltage and current phasors with very high accuracy is universally recognised as an important aspect of delivering a secure and sustainable power system. PMUs are a relatively new technology and because of their high procurement and installation costs, it is imperative to develop appropriate methodologies to determine the minimum number of PMUs as well as their strategic placements to guarantee full observability of a power system. Thus, the problem of the optimal PMU placement (OPP) is formulated as an optimisation problem subject to various constraints to minimise the number of PMUs while ensuring complete observability of the grid. In this chapter, integer linear programming (ILP), genetic algorithm (GA) and non-linear programming (NLP) constrained models of the OPP problem are presented. A new methodology is proposed to incorporate several constraints using the NLP. The optimisation methods have been written in Matlab software and verified on the standard Institute of Electrical and Electronics Engineers (IEEE) 14-bus test system to authenticate their effectiveness. This chapter targets United Nations Sustainable Development Goal 7.

The performance of thermal comfort utilising machine learning and its acceptability by students and other users at the Professor Sir Edouard Lim Fat Engineering Tower at the University of Mauritius are evaluated in this study. Students and building occupants were asked to fill out surveys on-site as data was gathered from sensors throughout the structure. The Thermal Sensation Vote (TSV) and other important data were collected through the surveys, including the effect of wind on thermal comfort. An adaptive model incorporating solar and wind effects was evaluated using multiple linear regression techniques and RStudio. Three models were used to evaluate thermal comfort, including the adaptive one. Numerous models were compared and evaluated in order to select the best one. It was found that the adaptive model (Model 1) was deemed to be the best model for its application. It was also found that Fanger's PMV/PPD (Model 2) was a very good approach to determining thermal comfort. Through thorough analysis, it was concluded that the range of air temperature and wind speed for thermal comfort was 25.830°C–28.0°C and 0.26 m/s to 0.42 m/s, respectively. In order for cities to remain secure, resilient and sustainable, it will be important to manage thermal comfort and reduce populations' exposure to heat stress (SDG 11). The achievement of income and productivity goals will be hampered if measures to protect populations from heat stress are not taken (SDG 8). Thermal regulation is also necessary for the provision of numerous health services (SDG 3).

In this chapter, we investigate the role of the Internet of Things (IoT) for a more sustainable future. The IoT is an umbrella term that refers to an interrelated network of devices connected to the internet. It also encompasses the technology that enables communication between these devices as well as between the devices and the cloud. The emergence of low-cost microprocessors, sensors and actuators, as well as access to high bandwidth internet connectivity, has led to the massive adoption of IoT systems in everyday life. IoT systems include connected vehicles, connected homes, smart cities, smart buildings, precision agriculture, among others. During the last decade, they have been impacting human activities in an unprecedented way. In essence, IoT technology contributes to the improvement of citizens' quality of life and companies' competitiveness. In doing so, IoT is also contributing to achieve the Sustainable Development Goals (SDGs) that were adopted by the United Nations in 2015 as an urgent call to action by all countries to eradicate poverty, tackle climate change and ensure that no one is left behind by 2030. The World Economic Forum (WEF) recognises that IoT is undeniably one of the major facilitators for responsible digital transformation, and one of its reports revealed that 84% of IoT deployments are presently addressing, or can potentially address the SDGs. IoT is closely interlinked with other emerging technologies such as Artificial Intelligence (AI) and Cloud Computing, for the delivery of enhanced and value-added services. In recent years, there has been a push from the IoT research and industry community together with international stakeholders, for supporting the deployment and adoption of IoT and AI technologies to overcome some of the major challenges facing mankind in terms of protecting the environment, fostering sustainable development, improving safety and enhancing the agriculture supply chain, among others.

The need to design buildings with due consideration for bioclimatic and passive design is central to promoting sustainability in the built environment from an energy perspective. Indeed, the energy and atmosphere considerations in building design, construction and operation have received the highest consideration in green building frameworks such as LEED and BREEAM to promote SDG 9: Industry, Innovation and Infrastructure and SDG 11: Sustainable Cities and Communities and contributing directly to support SDG 13: Climate Action. The research literature is rich of findings on the efficacy of passive measures in different climate contexts, but given that these measures are highly dependent on the prevailing weather conditions, which is constantly in evolution, disturbed by the climate change phenomenon, there is pressing need to be able to accurately predict such changes in the short (to the minute) and medium (to the hour and day) terms, where AI algorithms can be effectively applied. The dynamics of the weather patterns over seasons, but more crucially over a given season means that optimum response of building envelope elements, specifically through the passive elements, can be reaped if these passive measures can be adapted according to the ambient weather conditions. The use of representative mechatronics systems to intelligently control certain passive measures is presented, together with the potential use of artificial intelligence (AI) algorithms to capture the complex building physics involved to predict the expected effect of weather conditions on the indoor environmental conditions.

Climate change has been identified as a pressing social, environmental and economical challenge that has been unequivocally linked to human activity through latest Intergovernmental Panel on Climate Change (IPCC) reports. It is here to stay with us for generations to come and is already causing severe tribulations across the world. As nations devise policies to mitigate to climate change to stay within the 1.5 degrees Celsius target and small island developing states (SIDS) like Mauritius and the developing world in general find means to adapt to its consequences, a core shortcoming highlighted is the lack of community engagement and grassroots action so that policies permeate to concrete action. Of prime importance for this to happen is raising awareness on the climate change phenomenon, which has so far been a topic deemed complex for the general public, hence creating systemic barriers for climate action. The use of artificial intelligence (AI) can play a significant role in designing such community outreach programmes based on outcomes reported in literature in the educational sector in support of Sustainable Development Goal (SDG) 4: Quality Education. There is growing interest for a green lifestyle in the world population, and this chapter shows how the home can be used as a basic building block for allowing each household to contribute to climate action, while offering an effective case study to raise awareness on climate change through practical examples and demonstration, in support for SDG 11: Sustainable Cities and Communities. Based on an energy-water-materials nexus, the circular home concept is a clear contribution to SDG 13: Climate Action, with huge potential to use AI techniques and underpinning technologies to implement and optimise the efficacy of the proposed measures.

Industry 4.0 has been identified as a key cornerstone to modernise economies where man and machines complement each other seamlessly to achieve synergies in decision-making and productivity for contributing to SDG 8: Decent Work and Economic Growth and SDG 9: Industry, Innovation and Infrastructure. The integration of Industry 4.0 remains a challenge for the developing world, depending on their current status in the industrial revolution journey from its predecessors 1.0, 2.0 and 3.0. This chapter reviews reported findings in literature to highlight how robotics and automated systems can pave the way to implementing and applying the principles of Industry 4.0 for developing countries like Mauritius, where data collection, processing and analysis for decision-making and prediction are key components to be integrated or designed into industrial processes centred heavily on the use of artificial intelligence (AI) and machine learning techniques. Robotics has not yet found its way into the various industrial sectors in Mauritius, although it has been an important driver for Industry 4.0 across the world. The inherent barriers and transformations needed as well as the potential application scenarios are discussed.

The complexity of atmospheric corrosion, further compounded by the effects of climate change, makes existing models inappropriate for corrosion prediction. The commonly used kinetic model and dose-response functions are restricted in their capacity to represent the non-linear behaviour of corrosion phenomena. The application of artificial intelligence (AI)-driven machine learning algorithms to corrosion data can better represent the corrosion mechanism by considering the dynamic behaviour due to changing climatic conditions. Effective use of materials, coating systems and maintenance strategies can then be made with such a corrosivity model. Accurate corrosion prediction will help to improve climate change resilience of the social, economic and energy infrastructure in line with the UN Sustainable Development Goals (SDGs) 7 (Affordable and Clean Energy), 9 (Industry, Innovation and Infrastructure) and 13 (Climate Action). This chapter discusses atmospheric corrosion prediction in relation to the SDGs and the influence of AI in overcoming the challenges.

The consideration of alternative sources of material for construction is imperative to reduce the environmental impacts as two-fifths of the carbon footprint of materials is attributed to the construction industry. One alternative material with improved biodegradable attributes which can contribute to carbon offset is bamboo. The commercialisation of bamboo in modern infrastructures has significant potential to address few of the Sustainable Development Goals (SDGs) itemised by the United Nations, namely SDG 9 about industry, innovation and infrastructure. Other SDGs covering sustainable cities and communities, responsible consumption and production and climate action are also indirectly addressed when utilising sustainable construction materials. Being a natural material however, the full commercialisation of materials such as bamboo is constrained by a lack of durability. Besides fracture mechanisms arising from load-induced cracks and thermal modification, the durability of bamboo material is greatly impaired by biotic and abiotic factors, which equally affect its natural rate of degradation, hence fracture behaviour. In first instance, this chapter outlines the various factors leading to the durability limitations in bamboo material due to load-induced cracks and natural degradation based on recent findings in this field from the author's own work and from past literature. Secondly, part of this chapter is devoted to a new approach of processing the surge of information about the varied aspects of bamboo durability by considering the powerful technique of artificial intelligence (AI), specifically the artificial neural network (ANN) for prediction modelling. Further use of AI-enabled technologies could have an impactful outcome on the life cycle assessment of bamboo-based structures to address the growing challenges outlined by the United Nations.

Manufacturing in Mauritius is mostly export-oriented. Any supply chain (SC) failure or resilience deficit may result in cancellation of orders and loss of customers, market share and revenue and reduce capability to compete globally. Addressing this challenge is complex, although digital technologies and artificial intelligence (AI) models can improve resilience by assisting decision-making and mitigate risks, thus infusing greater predictability across the SC.

Supply chains are facing increasing disruptions and uncertainties owing to extreme weather events, the war in Ukraine, market volatility and the ongoing COVID-19 pandemic, among other factors. Manufacturing industries and their supply chains essentially create thousands of jobs that enable economic growth and sustain export capability. In addition, they need to maintain or increase both productivity and efficiency and recover quickly from unforeseen or unexpected challenges – that is they need to be resilient. Transformation initiatives, whether in production or supply chain management (SCM), are never easy. Process changes not supported by data or hurried human decisions can sometimes have unintended consequences, mainly adverse. However, in times of greater uncertainty (war and pandemic), setbacks can have greater consequences on the business. Manufacturers are already apprehensive and report slowing exports as recession concerns have caused consumers and businesses to pull back on spending. There is therefore a need to reduce uncertainty and augment resilience by unlocking and synthesising insights that emanate from the power of data analytics, AI and machine learning to improve the resilience efficiency balance.

This chapter will discuss the opportunities arising from the adoption and implementation of digital technologies and AI in SCM, leading to better value creation, less greenhouse gas emissions and resilience. The hurdles that enterprises are facing to integrate AI in their logistics and SCs will also be highlighted. This work comments on initiatives that uphold the objectives of SDG 8 – decent work and economic growth, SDG 9 – industry, innovation & infrastructure and SDG 13 – climate action.

The age of artificial intelligence (AI) is already upon us. The rapid development of AI tools is facilitating sustainable development and its corollary social good. For AI dedicated to social good to be impactful, it has to be human-centred, striving to achieve inclusiveness, sustainable livelihoods and community well-being. In short, it offers major opportunities to holistically enhance peoples' lives in diverse areas: education, health care, food security, disaster reduction, smart cities, etc. However, ethical, unbiased and ‘secure-by-design’ algorithms that power AI are crucial to building trust in this technology. Civil society's engagement can hopefully drive the features and values that should be embedded in AI.

This chapter focuses on the societal benefits that AI can deliver. Our initiatives and decisions of today will fashion the ‘Social Good’ AI applications of tomorrow. Sustainable Development Goals (SDGs) being addressed are 2–4 and 10–11.

This chapter offers an overview of the applications of artificial intelligence (AI) in the textile industry and in particular, the textile colouration and finishing industry. The advent of new technologies such as AI and the Internet of Things (IoT) has changed many businesses and one area AI is seeing growth in is the textile industry. It is estimated that the AI software market shall reach a new high of over US$60 billion by 2022, and the largest increase is projected to be in the area of machine learning (ML). This is the area of AI where machines process and analyse vast amount of data they collect to perform tasks and processes. In the textile manufacturing industry, AI is applied to various areas such as colour matching, colour recipe formulation, pattern recognition, garment manufacture, process optimisation, quality control and supply chain management for enhanced productivity, product quality and competitiveness, reduced environmental impact and overall improved customer experience. The importance and success of AI is set to grow as ML algorithms become more sophisticated and smarter, and computing power increases.