Objectives

General Objective

Developing communication network and Internet of Things technologies to enable smart, efficient, secure and innovative applications and services in strategic national areas.

Specific objectives

1. Coordinate the development of human resources. Train human resources with national and international experience at undergraduate and postgraduate levels. Encourage the qualification of teachers. Train high school and technical students, as well as professionals working in areas related to the project. Promote mobility, innovation, and the exchange of experience between the project’s human capital. Develop skills and competencies in the project area, including entrepreneurship.

2. Coordinate the international activities of the project. Promote the training of human resources with international experience, strengthen joint PhD agreements, discuss project progress and results, publish articles in collaboration with international partners, consolidate international partnership networks, promote courses and subjects taught in English with international support, develop research projects with international groups, and disseminate the project abroad.

3. Coordinate Technology Transfer to the Business Sector and/or the Public Sector. Identify the best strategies and models for the technological transfer of products, services, and technical knowledge from the project. Promote meetings with start-ups, technology parks, and public and private business sectors. Encourage entrepreneurship and innovation among project partners. Elevate Brazil's technological and competitive standards. Register software and patents.

4. Coordinate Technology Transfer to Society. Identify the best strategies and models for transferring the project’s technological products and technical knowledge to society. Promote lectures and meetings at schools, universities, science fairs, scientific events, and government and society sectors. Encourage social and sustainable entrepreneurship. Create secure and efficient digital services. Encourage the use of the project’s technological products and results for the benefit of society.

5. Coordinate Scientific Dissemination and Popularization of Science. Present the project’s results and innovative products to the general public in an accessible way. Publish articles in scientific events and qualified journals. Promote workshops, meetings, lectures, panels, and courses for different audiences. Foster science, research, technology, and innovation, including in interdisciplinary and cross-cutting ways.

6. The popularization of new technologies and Internet-connected applications has led to a significant increase in data transmission demands on the optical infrastructure of 5G/6G networks within the project's context. The use of artificial intelligence and load-aware routing helps overcome existing limitations in the optical core. This objective aims to investigate and propose intelligent and efficient solutions for routing and protection in elastic optical networks with spatial division multiplexing.

7. The spread of intelligent IoT devices presents challenges for protecting user privacy in light of new laws (e.g., LGPD) and verticals such as health and industry. The lack of robustness in monitoring traffic entering and leaving devices makes them attractive targets for attackers, either being attacked or becoming bots. This objective will develop lightweight and intelligent mechanisms based on technologies that enhance privacy in intrusion detection systems in IoT environments.

8. Smart environments composed of IoT devices and users generate data that can be used as a basis for AI services and data analysis. However, this data may contain sensitive user information, directly affecting privacy laws. This objective aims to develop security solutions for communication and data protection in smart IoT environments capable of securely analyzing, transmitting, and storing data.

9. Define, implement, and evaluate mechanisms to protect IoT network traffic against information leaks. The use of encryption will be considered, but other forms of traffic behavior obfuscation will also be proposed.

10. Define, implement, and evaluate algorithms and protocols for selecting IoT clients in federated learning, considering the client's capacity and available energy, as well as their impact on improving the accuracy of the global federated model, in order to keep client devices operational and accelerate the convergence of global model parameter values.

11. To cope with high dynamism, IoT applications and services require quick adaptations to changes, usually based on the detection of relevant contexts. These adaptations encompass actions across all system layers, from the network, through the middleware, to the application. Moreover, to handle ultra-large scale, such systems must operate autonomously. This objective proposes and evaluates support models and mechanisms for performing context-aware adaptation actions in IoT systems.

12. Define, implement, and evaluate algorithms and protocols for selecting IoT clients in federated learning, considering the client's capacity and available energy, as well as their impact on improving the accuracy of the global federated model, in order to keep client devices operational and accelerate the convergence of global model parameter values.

13. Smart IoT applications deal with the distributed nature of the edge and network core, in a computational continuum where computing can happen anywhere, involving requirements such as privacy, latency, efficiency, dynamism, and mobility. This objective aims to design an architectural framework, use cases, and algorithms, and evaluate smart applications using various models, such as federated learning.

14. Energy consumption is a limiting factor in IoT environments. Therefore, it is important to consider the balance between energy and performance in IoT ecosystems. This objective aims to propose intelligent task schedulers at the edges that are energy-aware and meet the quality requirements of IoT applications. Different energy management mechanisms, such as Dynamic Voltage and Frequency Scaling, Energy Harvesting, and Internet of Energy, can be investigated to identify advantageous approaches.

15. Develop early exit strategies in Deep Neural Networks (DNNs) in an IoT edge computing scenario for different verticals to classify more samples at the network edge in less time and with lower energy consumption.

16. Data generated by IoT devices, including social and vehicular networks, have different forms, meanings, and privacy requirements. In this context, federated learning techniques can be used to provide information security for applications that require privacy while still generating knowledge. This objective aims to propose and evaluate a framework and services using federated learning techniques in the IoT context for smart city applications.

17. Design, develop, and evaluate an intelligent resource allocation architecture for the computing continuum encompassing IoT edge, fog, and cloud devices, considering them as a computing continuum. Study, develop, and evaluate distributed learning models focused on the efficient management of computing resources and optimization of applications for different smart city and industrial verticals in the computing continuum.

18. Study, propose, and evaluate distributed machine learning solutions using smart devices at the edge of the network that consider user context, data privacy, device computational and communication constraints, as well as the computing and communication requirements of applications for different project verticals.

19. The premise of distributed learning is the convergence of a global learning model through the aggregation of parameters from locally trained models. This specific objective envisions that participants will train models with privacy and rapid convergence based on distributed and/or federated learning premises, including participant evaluation/reputation, collaborative training, adaptive participant selection, and intermediate information exchange with privacy.

20. Conventional federated learning models are susceptible to attacks where malicious nodes can inject false parameters to divert the model from its intended objective. With this attack, the global model may lose accuracy or even deliberately make wrong inferences. This objective aims to propose and evaluate federated learning models that are robust to the main adversarial attacks typical of distributed edge environments.

21. The number of applications and services available to users has increased considerably with the adoption of cloud computing in 5G/6G and IoT environments. Communication will be information-oriented, and user and data security and privacy will be ensured across all communication and computing layers. This objective aims to propose and evaluate mechanisms and technologies for secure and private communication in heterogeneous multi-layer cloud scenarios, namely Edge, Fog, and Cloud.

22. Develop architectures, mechanisms, and tools to provide service orchestration in the multi-layer Edge, Fog, and Cloud scenario, as well as communication solutions for integrating the Edge-Fog-Cloud Continuum. The goal is to also add Intelligence at the network Edge, where collected data can be pre-processed at the network edge and generate knowledge about specific regions, allowing for a faster, more efficient, intelligent, and personalized response.

23. This objective aims to develop strategies for intelligent and efficient resource allocation in the context of Multi-Access Edge Computing (MEC) at the edge of 5G/6G networks. Strategies should consider factors such as user mobility, available information for operators, application requirements, and user privacy.

24. Develop and evaluate models based on deep reinforcement learning to improve the use of caches in edge computing environments with the ability to generate knowledge about the spatiotemporal dynamics of content popularity through user requests. The proposal allows the model to alter the cache states to represent the local and momentary popularity of requests and satisfy them without needing to resort to the content provider.

25. Ultra-low latency applications (e.g., holograms in healthcare and autonomous vehicle driving) are part of the smart cities portfolio. To ensure the application requirements, the efficient use of intelligent edge and extreme edge computing is essential. This objective will study, develop, analyze, and evaluate algorithms and protocols capable of managing edge resources efficiently and intelligently in smart cities for ultra-low latency applications.

26. Unmanned Aerial Vehicles (UAVs) equipped with processing, computing, and network resources can be used by cloud/edge providers and cellular operators to reduce latency in service access, as they have the ability to fly very close to end users. This objective proposes and evaluates efficient and secure scheduling algorithms, protocols, and mechanisms that use UAVs to provide connectivity and services in smart cities.

27. Edge computing can use resources efficiently and improve data analysis performance. The in-network computing paradigm allows the use of programmable switches to perform computing tasks offloaded from servers to within the network itself. This objective proposes and evaluates mechanisms that use the programmable data plane at the edge to integrate strategies that improve communication, analysis, and data fusion in IoT environments.

28. Intelligent routing in 5G/6G software-defined networks with multiple controllers is essential for improving network resource utilization and application quality. This objective aims to develop and evaluate intelligent routing techniques based on reinforcement learning for networks with flat topologies and hierarchical SDN controller topologies for different verticals.

29. The correct and efficient operation of the application scenarios foreseen in the project and the 5G/6G infrastructure itself depends on Machine Learning and AI mechanisms contributing to tasks, e.g., classification, data analysis, etc. These tasks need to be performed on large volumes of data and at very high speed. This objective explores and evaluates the programmability of 5G/6G network devices to implement ML and AI in-network and at line speed.

30. Attacks on critical systems supported by 5G/6G networks have the potential to cause injury or loss of human lives, material or environmental damage, and financial losses. This scenario is challenging due to the inherent complexity and broad attack surface in systems and networks. This objective aims to develop and evaluate technologies to promote resilience in critical systems over 5G/6G networks, even in the face of cyberattacks.

31. The goal is to use the centralized view of the 5G/6G SDN controller to implement elastic defense concepts. This approach allows network elements to defend themselves by diverting attacks while luring the attacker into wasting time and resources attacking resilient (or fake) servers, avoiding or delaying the attack. Additionally, this strategy provides behavioral data for new insights into the attacker's modus operandi and can be applied in data centers or IoT systems.

32. Conventional vulnerability detection techniques lose performance when applied to IoT due to the diversity of devices and manufacturers and the behaviors generated in network traffic. This objective aims to model and investigate automated ways of identifying security vulnerabilities in 5G network communication with IoT environments. Protection mechanisms for communication networks and IoT environments against information leaks will also be proposed.

33. Develop emerging distributed system models and applications that enable optimized resource management in a 5G and 6G evolution horizon. Identify applications that benefit from distributed computing in the computing continuum (IoT-edge-cloud), identifying their requirements for modeling optimization problems in computing and network resource allocation. Design and develop algorithms for intelligent resource allocation.

34. This objective aims to develop and evaluate new technologies for resource and service orchestration in core and radio infrastructures in 5G/6G networks to meet the requirements of the project's verticals. Ensure that resources are efficiently provisioned and elastically throughout their entire lifecycle.

35. The integration of new resource-constrained devices can be improved by considering the programmability of 5G/6G networks. Such integration allows devices and gateways to mutually authenticate and establish secure communication channels that are resistant to tampering and eavesdropping. This objective aims to develop robust mechanisms for secure and independent integration of resource-constrained devices to provide a secure transport platform in constrained networks.

36. The programmability paradigm in networks has been gaining increasing interest not only from academia but also from industry. Various solutions and approaches have been proposed to reduce the complexity of these networks while making them more dynamic and scalable. This objective aims to develop and evaluate solutions that promote fully automated and intelligent management and control of programmable networks, with minimal (or even no) human intervention.

37. Caching and off-loading mechanisms in edge networks are pillars for immersive applications in the context of 5G/6G networks. These applications will have prohibitive costs associated with processing, storing, and transmitting data to provide individualized experiences. This objective aims to propose and evaluate intelligent caching and off-loading mechanisms for immersive virtual environments with stringent temporal restrictions in 5G/6G networks.

38. This objective aims to develop and evaluate network security solutions against Web server attacks, such as Distributed Denial of Service (DDoS), using network function virtualization and software-defined network technologies. These solutions are intended not to significantly affect benign traffic and avoid overloading SDN controllers to prevent the network from collapsing during attacks.

39. NFV has the potential to increase network agility and reduce infrastructure costs. Deploying VNFs in edge environments is challenging. This objective aims to propose and evaluate an architecture and mechanisms to determine which nodes to deploy VNFs, how many resources to allocate (and how to resize them), and whether to move VNFs (to balance load, save energy, ensure QoS in the presence of mobility) for different verticals.

40. Virtualization has been incorporated into the programmability domain of the data plane. However, current solutions for virtualizing programmable switches lack guarantees of secure and independent management of virtual instances and do not allow resource sharing between virtual instances to reduce memory demand. This objective proposes and evaluates robust and secure management mechanisms for lightweight virtualization in programmable data planes.

41. This objective aims to design and evaluate new architectures, techniques, and mechanisms that enable the autonomous and proactive management of resources and services in virtualized and orchestrated slices in 5G/6G networks. To this end, it intends to explore advanced AI and Machine Learning techniques for the intelligent management of 5G/6G network slices across the IoT-edge-cloud continuum, considering URLLC use cases.

42. This objective aims to study, develop, and evaluate Intrusion Detection Systems (IDSs) for Next-Generation Automotive Networks using artificial intelligence. This research aligns with recent regulations, such as UN Regulation No. 155, which establishes, among other measures, that automakers must be able to detect and respond to security incidents across their entire fleet of vehicles. IDSs will be developed with a focus on intra-vehicular Ethernet networks.

43. Factors such as urban mobility prediction, knowledge extraction, and efficient use of computational resources play an important role in the Intelligent Transportation System. In this context, the federated learning paradigm is emerging as a trend to address challenges in vehicular networks. This objective aims to develop and evaluate federated protocols that combine topological and social properties of urban computing to develop federated vehicular networks.

44. Efficient mobility management is fundamental to the success of vehicular networks and autonomous vehicles. This objective aims to propose and analyze protocols, algorithms, and mechanisms for intelligent and efficient mobility management (including with predictive support) to address application and service challenges and keep vehicles better connected in 5G/6G vehicular and autonomous vehicle networks.

45. Model, investigate, and evaluate automated and intelligent ways of identifying security vulnerabilities in vehicular inter-network communication quickly.

46. This objective aims to investigate, implement, optimize, and analyze aspects related to intra- and inter-vehicular communication for vehicular applications in the context of smart cities. These aspects include issues with network topologies, security, communication protocols, energy consumption, performance, and hardware infrastructure for autonomous vehicle applications, considering aspects of active safety functions, predictive maintenance, and driver assistance.

47. Smart cities have various types of sensors that can be used as a data source for urban sensing. With the collected data, relevant information can be extracted and used in various types of application domains. This objective aims to propose intelligent mechanisms for collecting, analyzing, fusing, and disseminating heterogeneous urban data to enable the popularization and conception of applications and services for smart cities.

48. Smart cities have various types of sensors that can be used as a data source for urban sensing. With the collected data, relevant information can be extracted and used in various types of application domains. This objective aims to propose intelligent mechanisms for collecting, analyzing, fusing, and disseminating heterogeneous urban data to enable the popularization and conception of applications and services for smart cities.

49. Improving traffic management remains a challenge in smart cities. This objective aims to develop and evaluate machine learning-based solutions for optimizing traffic signal timing in smart cities, considering different actors (vehicles and pedestrians), local and global views of the problem, and various aspects, from travel time to fuel consumption optimization.

50. 5G surveillance cameras are becoming popular and monitoring various spaces in urban centers. Strengthening device security and data privacy is vital for a more digital society. This objective aims to propose and evaluate efficient security protocols and deep learning algorithms to improve the security of remotely updatable 5G cameras and user privacy in smart cities.

51. This objective aims to develop and evaluate secure solutions for smart cities using security properties offered by blockchain technologies and blockchain-based systems combined with modern cryptography techniques to ensure the privacy of users and data. They should also consider the scalability requirement of the scenario, evaluating scalable technologies in blockchain systems, such as payment channel networks and side channels.

52. Participatory sensing networks (PSNs) are sources of social sensing on an unprecedented scale. This objective aims to characterize and model social behavior patterns and city dynamics with the help of different PSNs, explore the identified patterns to propose new services and applications to improve services for urban societies, and evaluate their benefits in smart cities.

53. This objective aims to propose and evaluate secure and efficient federated learning for digital health applications and services with IoT devices, including wearables. Federated learning is an alternative to sharing and transferring large volumes of data to central servers. The proposed algorithms will consider horizontal and vertical data partitioning as well as the security of training and inference procedures.

54. Channel State Information (CSI) data capture information about how the Wi-Fi signal behaves around and through individuals, considering time, frequency, and space domains. This information can be used for various human detection applications, including falls, movements, and health monitoring. This objective aims to develop and evaluate intelligent techniques for remote patient monitoring using CSI data - Ethics Committee #54359221.4.0000.5243.

55. 5G/6G networks will drive digital health services and rely on cloudification, MEC, AI, and resource slicing to improve the efficiency of available resources even in mobile and dynamic scenarios. This objective aims to develop and evaluate 5G/6G network solutions to ensure the quality and connectivity requirements of digital health services even in mobile and dynamic scenarios.

56. This objective aims to define, develop, implement, and evaluate machine learning models capable of analyzing data and making intelligent and autonomous decisions at the edge and extreme edge of rural networks for bioeconomy, agriculture, and environmental monitoring verticals. The models should be optimized to operate on devices with low processing power, storage, and connectivity, as well as with heterogeneous data.

57. The development of connected and smart rural areas through their synchronous or asynchronous connection to the Internet and support for intelligent IoT has the potential to transform Brazil's bioeconomy, agriculture, and environmental preservation and enable producers to improve their operations. This objective aims to define, test, prototype, and evaluate network, cloud, and IoT devices for use in rural areas with intelligence at the edge and synchronous or asynchronous connectivity.

58. This objective aims to define, develop, implement, and evaluate communication network and IoT technologies at the edge and extreme edge of rural networks to enable intelligent applications and services with support for intelligence, autonomy (reducing human dependence), scalability, resilience, security, energy efficiency, heterogeneity, and differences in the physical, environmental, social, and economic characteristics of rural properties.

59. Digital Twin (DT) is the name given to a type of system that integrates a physical part and its digital replica, and its use in industry contributes to cost reduction and increased productivity. This objective aims to propose and evaluate intelligent, efficient, and scalable mechanisms to implement communication infrastructure for digital twins in industry verticals with real-time requirements based on edge computing.

60. The digitization process of the industry contributes to increasing the productivity and accuracy of services available in smart factories, including applications in challenging scenarios such as mobile IoT or robots. This objective aims to define, develop, implement, and evaluate intelligent IoT technologies for challenging fixed and mobile scenarios of digital industries with support for sustainability, energy efficiency, and application quality requirements.