Components for Secure, Safe and Reliable Wireless Systems

Artificial intelligence applied to wireless IOT applications is expected to boost the number of automated industrial applications in Europe. This means an explosion in the use of digital infrastructure for the delivery of wireline-like tunnelling trustable, safe and secure transmission systems. Trusted systems can be built on top of the highly adaptable and intelligent wireless transmission layer, with reduced interference, minimized energy expenditure, reduce carbon footprint, improve quality of service, reduced emissions, improved spectral efficiency, etc. In addition, wireless technologies enable a wide set of flexible mobility scenarios that reduce the need of fixed infrastructure, , plus provide dynamic configurability over the air, and allow for, troubleshooting and management in a much more flexible than cable infrastructure. To make this more palpable, InSecTT looks into the possibility of highly available intelligent wireless services on the fly for a wide set of traffic types, available over large footprints to a larger set of stakeholders that can use intelligent services in a flexible trustable, secure and safe manner connecting directly objects, actuators, or other critical infrastructure to edge/cloud applications.  This has large repercussions in the economy of the countries and societies. Industrial processes will become more intelligent, flexible, available, and affordable than ever before. The cloudification of object via intelligent wireless ultrareliable and ultra-low latency links means the explosion of real time remote applications, leading to an economic growth without precedent and emerging issues that can be only handled by a trustable verifiable and accountable AI layer proposed by InSecTT.

AI-enabled wireless networks are expected to have a huge impact on new applications such as automated driving vehicular networks. Therefore, 5G is directly related to mobility aspects considering different mobility models for terminals, relays, and clusters of terminals (which arise in self organized networks)

AI-wireless IoT will generate research and industrial efforts around the world. Human resources and specialized services will be created around a technology that is expected to revolutionize the industrial services providing a higher level of automation as ever seen today. This human capital will get lots of benefits from the expertise generated by this proposal.

Autonomous driving vehicles will enable huge savings in fuel consumption and emissions. Road safety will be improved. Traffic control can be used to reduce emissions based on environmental information.

AI-enabled IoT represents a new paradigm in communication between machines and industrial applications. This opens also issues in security, trustiness, safety and privacy. This project will also explore these issues and attempt to evaluate existing solution that will minimize the vulnerabilities of the new communication paradigm.

Hidden

Your content goes here. Edit or remove this text inline or in the module Content settings. You can also style every aspect of this content in the module Design settings and even apply custom CSS to this text in the module Advanced settings.

Impact in the PHY layer of wireless systems:
  1. More accurate and dynamic propagation modelling and prediction. This is particularly important for autonomous driving applications, where context can be inferred from different measures (on and out of band) to improve the prediction of the signals and obstacles where waves will propagate. The use of artificial intelligence in this particular aspect can lead to a reduction of the probabilistic nature of wireless communication, thus contributing to create an almost wireline-like performance with deterministic or quasi deterministic bounds and performance. Quantifiable metrics: New generation of supervised channel learning algorithms for wireless IoT in different industrial domains. This new generation will be tested, optimized and disseminated. This will lead to improvements in interference, spectrum efficiency, and energy consumption by several orders of magnitude considering large MIMO operation. At least one new supervised algorithm for each use cas propagation scenario.
  2. AI can be used together with software defined and cognitive radios to provide a higher level of adaptivity according to the measurements, predictions and learned lessons from the algorithms. This adaptivity means that better exploitation of channel resources will be achieved, improved energy consumption, reduce power and carbon emission, safer operation for humans (radiation exposure), controlled electromagnetic interference, reduce jamming attacks and eavesdropping passive attacks, and in general boosting of capacity of wireless radio resources. This will contribute to a more efficient use of radio spectrum and to a more flexible paradigm for opportunistic access that will allow different stakeholder to improve their business models and budget plans. In the end, the final user is expected to be offered with a set of wireless over-the-air services, management, commissioning and troubleshooting that will enhance its expertise to a quasi-deterministic secure and trustable medium.
  3. Modulation schemes and information can become more secret and complex than ever before. AI modulation can lead to multiple formats for communication that can be adapted or improved on the fly and lead to an adaptation PHY layer according to the end use requirements.
  4. Signal processing with multi-objective evolutionary AI algorithms is regarded as one of the major breakthroughs for future 5G systems. AI is therefore essential for the operation, optimization, management troubleshooting, and commission IoT future wireless networks. Quantifiable metrics: Reduction of convergence time for Pareto frontier by 50% using AI algorithms.
  5. The use AI algorithms in blind and semi-blind wireless transmission is regarded as a Pandora box, once this aspect is opened, the PHY-layer performance will be considerably improved. The large number of training sequences currently used to estimate e and equalize channel in wireless networks will see the efficiency reduce when large number of objects need to be located with such orthogonal sequences. The use of non-orthogonal semi-blind algorithms promises to reduce by several orders of magnitude those training sequences and there improves the efficiency of channel occupation. Spectral efficiency improvement 10x.
At the MAC layer
  1. AI algorithms will be used to bring new distributed intelligence that will reduce conflicts between different entities, will improve resource allocation in semi or fully distributed or decentralized scenarios, in resource constrained networks such as RFID and tags, and will improve channel utilization by several orders of magnitude. Channel occupancy efficiency nearly 100 percent.
  2. One of the main issues in future wireless IoT in industrial applications is how to achieve ultra-low latency and real time operation. Artificial intelligence can be used to achieve the so-called multi-packet non-orthogonal reception, and thus help alleviate the require number of retransmissions to receive packets at the MAC layer. It will also alleviate or completely eliminate the need of coordination and resource reservation improving algorithms such as the grant-free and non-orthogonal multiple access. Latency target less than 5ms end to end.
At the context aware management

InSecTT aims to development a framework for trustable certifiable and verifiable AI-IoT wireless services. The AI framework allows context information to be transported safely and in a private manner across domains, enabling the flexible application development. This aims to propel European industry to enable industry4.0 and other IoT smart city applications under verifiable trustable AI platforms. This will enable mission-critical and other high-dependability services in the physical and digital world.

Adding trustworthy ranging and localization information to the framework described above – by employing a broadband radio technology such as UWB – will bring additional benefits to a wide range of applications where mobile devices are interconnected with each other as well as with different infrastructure components of (industrial) IoT applications. UWB-based secure wireless localization and ranging technology is currently one of the fastest growing technologies because it allows to combine sensor data with the sensors location which enables multiple use cases that are depending on location-aware data. Secure ranging schemes are essential enablers for all systems where secure location is a critical parameter (e.g. in healthcare, red zone access and secure incident prevention. In parallel to organisations such as the FiRa consortium, which is defining protocols for having interoperability between different UWB-based devices used in different use cases such as smart access, or IEEE, where the UWB 802.15.4z standard with focus on secure ranging technology and replay/relay attack prevention is currently defined, InSecTT will show the benefits of employing such secure ranging technology in several industrial, cross-domain use cases and thereby further contribute to fostering interoperability within industrial IoT applications.

At the security, safety, and trustiness level

InSecTT will propose AI algorithms to improve many different aspects of wireless European industry connectivity and therefore enable flexible, dynamic, secure, trustable services over wider geographical areas for multi-domain and multicultural societal scenarios. AI is regarded as a key enabler of security breach analysis, hazard identification, dynamic countermeasure, threat avoidance, troubleshooting, and management of privacy in IoT systems. In InSecTT we aim to enable this prominent feature though a set of AI algorithms according to European standards related to safety, security and trust. Target metrics for AI algorithms: convergence time, stability, trust framework for Ai algorithms in cyberspace as extension of security metrics. When combined with highly dependable wireless (SW, HW) communication solutions, it will be possible to to offer mission-critical services in the market. Therefore, in InSecTT also the solutions are analysed and pre-developed to bring the enabling wireless communication technology platforms towards high automotive (ASIL-B) and industrial integrity levels.

At the context aware management

InSecTT aims to development a framework for trustable certifiable and verifiable AI-IoT wireless services. The AI framework allows context information to be transported safely and in a private manner across domains, enabling the flexible application development. This aims to propel European industry to enable industry4.0 and other IoT smart city applications under verifiable trustable AI platforms. This will enable mission-critical and other high-dependability services in the physical and digital world.

Adding trustworthy ranging and localization information to the framework described above – by employing a broadband radio technology such as UWB – will bring additional benefits to a wide range of applications where mobile devices are interconnected with each other as well as with different infrastructure components of (industrial) IoT applications. UWB-based secure wireless localization and ranging technology is currently one of the fastest growing technologies because it allows to combine sensor data with the sensors location which enables multiple use cases that are depending on location-aware data. Secure ranging schemes are essential enablers for all systems where secure location is a critical parameter (e.g. in healthcare, red zone access and secure incident prevention. In parallel to organisations such as the FiRa consortium, which is defining protocols for having interoperability between different UWB-based devices used in different use cases such as smart access, or IEEE, where the UWB 802.15.4z standard with focus on secure ranging technology and replay/relay attack prevention is currently defined, InSecTT will show the benefits of employing such secure ranging technology in several industrial, cross-domain use cases and thereby further contribute to fostering interoperability within industrial IoT applications.

At the security, safety, and trustiness level

InSecTT will propose AI algorithms to improve many different aspects of wireless European industry connectivity and therefore enable flexible, dynamic, secure, trustable services over wider geographical areas for multi-domain and multicultural societal scenarios. AI is regarded as a key enabler of security breach analysis, hazard identification, dynamic countermeasure, threat avoidance, troubleshooting, and management of privacy in IoT systems. In InSecTT we aim to enable this prominent feature though a set of AI algorithms according to European standards related to safety, security and trust. Target metrics for AI algorithms: convergence time, stability, trust framework for Ai algorithms in cyberspace as extension of security metrics. When combined with highly dependable wireless (SW, HW) communication solutions, it will be possible to to offer mission-critical services in the market. Therefore, in InSecTT also the solutions are analysed and pre-developed to bring the enabling wireless communication technology platforms towards high automotive (ASIL-B) and industrial integrity levels.

At the IoT level

AI is widely accepted as the boosting element of future IoT, industry4.0 and smart city applications. Yet the AI boom has not yet been full adopted in Europe. InSecTT aims to change this paradigm, enabling a set of use cases of IoT in industrial domains with the new AI layer, solving issue of compatibility, stability, verification, validation, security, trustiness, etc. in this convergence that seems strategical for all the countries and blocks in the world.

At the edge/fog/computing level

AI is closely connected with edge/cloud and fog computing concepts. AI algorithms aim to exploit the different level of complexity and available context information at different levels of the architecture making clear the advantages/disadvantages of centralized or decentralized information processing. InSecTT will provide guideline of implementation of AI algorithms and the trade-offs in different locations or levels of the Bubble architecture with the main objectives of each use case, namely: latency, deadlines, fairness, privacy, security, safety, trustiness etc.

In the EU economy

InSecTT aims to lead the main players of the EU economy to a new understanding and acceptance of AI-enabled IoT applications, thus leveraging the role of technical development and standardization, regulation to create value among European users. This will create competitiveness, catching up with other superpowers of the orb, and providing not only European citizens, but worldwide citizens with an example of how to create value from artificial intelligence in a decentralized, secure, standardized, robust and reliable manner.

In other stakeholders

Civil society, human rights, cybersecurity agencies, privacy sealing efforts will get benefit form a, EU-based artificial intelligence effort to propel regional industrial trustable services and validate global valuable imported services in the IoT future digital economy.