Innoguard

Hybrid and Generative intelligence for a Safe and Autonomous Future

An international project that will train the next generation of researchers to ensure reliability and trust in Artificial Intelligence (AI)-driven Autonomous Cyber-Physical Systems (ACPSs)

When autonomy becomes a challenge

Automated public transport is set to grow from 30% to 70% by 2030. The autonomous systems powering this transformation are becoming increasingly complex — integrating constant connectivity, diverse hardware, and advanced software. With such complexity, even a small failure can have serious consequences.

InnoGuard exists to ensure these systems are safe and reliable.
We combine engineering with advanced artificial intelligence to verify, debug, and build trust in autonomous systems.

Our approach is responsible and precise — continually evaluating safety, reliability, and potential bias in applied AI. By doing so, we not only protect today’s systems but also train and empower the next generation of researchers to lead the future of trustworthy technology.

What are ACPSs?

Autonomous Cyber-Physical Systems (ACPSs) are intelligent systems where AI and software directly interact with hardware and the real world to make automatic, data-driven decisions.

Common examples include autonomous vehicles, automated trains, and smart-drones.

These systems combine powerful computing, high connectivity, and diverse hardware – making them incredibly capable, but also critical. Any failure can lead to serious consequences; hence, ensuring their safety, reliability, and performance is vital for building trust in autonomy.

What do we aim to achieve?

At InnoGuard, our goal is to develop new techniques and methods that apply Artificial Intelligence–such as deep learning, evolutionary algorithms, reinforcement learning, and advanced language models— to the development of ACPSs.

We focus on ensuring these systems meet the highest standards of quality, safety, and reliability.

Developing ACPs is a complex process that requires automation across multiple phases– from requirements analysis to design and coding. Our priority is quality assurance, ensuring every system performs dependably under the demanding conditions autonomy requires.

Train disruptive talent

Design Innovative Training Program and Development Activities tailored to the ACPS context.

Automate trust

Automate quality assessment and evolution of ACPS behavior to ensure high ACPS trustworthiness and reliability.

Real-time safety

Enhance system dependability by improving security, privacy, and uncertainty management during operation.

More reliable AI

Strengthen AI trustworthiness and quality improvement approaches through novel quality engineering methods.

Digital sustainability

Reduce the environmental footprint of both ACPSs and the AI models that power them through efficient, sustainable design.

Open validation

Validate the relevance and cost-effectiveness of the project services in open source contexts.

Legal impact

Ensure compliance and foster innovation by developing a robust constitutional and legal framework for trustworthy AI and CPS, guiding policymakers, developers, and end-users toward responsible adoption aligned with EU fundamental rights and regulations.

Knowledge dissemination

Effectively communicate and share project results with the scientific and industrial communities.

How do we work?

To achieve our objectives, InnoGuard is organized into six Work Packages (WPs) that combine research, training, technological development, and results dissemination. Each package addresses a fundamental aspect to ensure Autonomous Cyber-Physical Systems that are safe, reliable, and sustainable.

Work Packages (WP):

Network and individual training

We provide tailored training programs for Doctoral Researchers that combine advanced academic expertise with hands-on industrial experience across Europe.

Project Management and Coordination

We ensure all project activities are executed efficiently, transparently, and collaboratively across all network members — fostering a cohesive and results-driven research environment.

Quality engineering in ACPS design

We develop and integrate innovative methods to embed quality, safety, and reliability into AI-enabled Cyber-Physical Systems (ACPS) from the earliest stages of design.

Monitoring and self-adaptability

We design methods that enable systems to self-monitor, self-heal, and adapt dynamically to changing contexts, ensuring continuous dependability and resilience.

Reliability of quality methods

We assess and validate the effectiveness, safety, and robustness of applied engineering methods to ensure trustworthy AI-driven systems.

Dissemination and exploitation of results

We actively communicate and transfer project outcomes to industry, academia, and policymakers — maximizing InnoGuard’s scientific and societal impact.

WP1: Project Management and Coordination

We ensure all project activities are executed efficiently, transparently, and collaboratively across all network members — fostering a cohesive and results-driven research environment.

WP2: Network and Individual Training

We provide tailored training programs for Doctoral Researchers that combine advanced academic expertise with hands-on industrial experience across Europe.

WP3: Quality Engineering in ACPS Design

We develop and integrate innovative methods to embed quality, safety, and reliability into AI-enabled Cyber-Physical Systems (ACPS) from the earliest stages of design.

WP4: Monitoring and Self-adaptability

We design methods that enable systems to self-monitor, self-heal, and adapt dynamically to changing contexts, ensuring continuous dependability and resilience.

WP5: Reliability of Quality Methods

We assess and validate the effectiveness, safety, and robustness of applied engineering methods to ensure trustworthy AI-driven systems.

WP6: Dissemination and Exploitation of Results

We actively communicate and transfer project outcomes to industry, academia, and policymakers — maximizing InnoGuard’s scientific and societal impact.