SafeComp 2026
SafeComp 2026
Logo_v4

The use of neural networks has expanded into areas as diverse as medical systems, industrial devices and space structures. In these cases, it is essential to balance performance, power consumption, and silicon area. Furthermore, in critical environments, it is necessary to ensure high dependability.

This workshop will provide a focused forum on how to design, verify, and certify AI-enabled embedded systems while preserving their overall dependability, including safety, reliability, availability, and security, under tight resource and real-time constraints. It targets AI-enabled embedded and cyber-physical systems deployed in safety-critical domains such as automotive, rail, aerospace, robotics, medical devices, and industrial control. The workshop will emphasize how functional safety, cybersecurity, and sustainability jointly contribute to system dependability when AI models, often opaque and non-deterministic, are deployed on constrained embedded platforms exposed to both accidental faults and malicious attacks.

Author Notification:  1 June 2026

Call for Papers
Submission

  • Architectures and patterns for safe deployment of AI in embedded and cyber-physical systems (runtime monitors, safety envelopes, redundancy, graceful degradation).

  • Methods for verification, validation, and testing of AI components in real-time and resource-constrained embedded environments, including scenario-based testing and robustness analysis.

  • Fault-tolerant and resilient hardware and software for AI in embedded systems. Reliability analysis and test of AI-enabled embedded systems.

  • Integration of AI components with safety and security standards and assurance frameworks (e.g., ISO 26262, IEC 61508, DO-178C, IEC 62304, security co-engineering).

  • Safety- and security-aware design of AI models for embedded devices (compression, quantization, scheduling, resource management) with guarantees on timing and performance.

  • Safety and security co-engineering for AI-enabled embedded and IoT devices, including attack surfaces, threat modeling, and defenses (e.g., model poisoning, adversarial inputs, backdoors, physical tampering).

  • Lifecycle management, monitoring, and update strategies for AI models in the field, including continuous assurance, re-certification, and data governance.

  • Out-of-Distribution (OOD) detection and Operational Design Domain (ODD) monitoring, including anomaly detection, uncertainty quantification, and mechanisms for safe fallback when AI systems encounter unforeseen conditions.

  • Case studies and lessons learned from deploying AI in safety-critical embedded systems (automotive, rail, aerospace, robotics, healthcare, energy, industrial automation).

  • Tools, benchmarks, and open datasets for assessing safety, security, and sustainability of AI-based embedded systems.

 

  • Paper submission: 4 May 2026  11 May 2026.

  • Notification of acceptance: 18 May 2026 (EXTENDED) 1 June 2026.

  • Camera-ready papers: 8 June 2026.

  • Workshop date: 22 September 2026 (co-located with SafeComp 2026).

DAIES Workshop tentative Program

Dependable AI in Embedded Systems

08:30 – 09:00

Registration

09:00 – 10:30

Session 1

  • MargisBench: A Framework to Assess AI Models Optimizations and Isolation at the Edge
    Marcello Cinque, Francesco Brunello and Salvatore Cangiano
  • Separation of Concerns for Early-Stage Engineering and Assessment of Dependable CNN Models
    Joan Ferreres Expósito, Juan Carlos Ruiz García and David De Andrés
  • I’m Sorry Driver, I’m Afraid I Can’t Do That: Appraising the Functional Safety of LLMs within Automotive Contexts
    Shaun Feakins, Ibrahim Habli, Kim Littler and Robert Palin
10:30 – 11:00

Coffee Break ☕

11:00 – 12:30

Session 2

  • Characterizing Single Event Functional Interrupts in Jetson Orin Nano SoC under Proton Irradiation
    Lester Frias, German Leon, Jose Antonio Belloch and Jose Manuel Badia
  • Dependability Analysis of a Spiking Convolutional Neural Network (S-CNN)
    Joaquín Gracia-Morán, Juan C. Baraza, Juan Carlos Ruiz García, David De Andrés, Daniel Gil and Luis-J. Saiz-Adalid
  • Radiation Testing Method for Observation of Soft Errors in Smart Sensors
    Matheus Minelli de Carvalho, Benjamin Cheymol, Lirida Naviner and Rodrigo Possamai Bastos
12:30 – 13:30

🎤 Keynote

From Silicon to Outer Space: Deploying Neuromorphic AI and SWaP Optimization in Critical Embedded Environments
Daniel Gutiérrez
CEO & Founder, Intigia

Abstract

The adoption of Artificial Intelligence (AI) in safety- and mission-critical embedded systems—such as autonomous driving, industrial robotics, and the aerospace sector (New Space)—demands a major paradigm shift. Traditional deep learning architectures suffer from unpredictable latencies, high power consumption, and intrinsic certification hurdles under strict functional safety standards such as DO-254, DO-178C and ECSS.

This keynote explores how the new frontiers of Edge AI are addressing these challenges through the convergence of hardware-level optimization, hardware/software co-design and bio-inspired computing. Drawing on Intigia's experience in cutting-edge European and regional projects, the presentation discusses how neuromorphic computing and SWaP (Size, Weight and Power) optimization are enabling a new generation of dependable intelligent embedded systems for critical environments.

Speaker Biography

Daniel Gutiérrez is the CEO and founder of Intigia, an engineering company specialized in SoC and IP core development for safety-critical applications in the aviation, space, automotive and industrial sectors.

With more than twenty years of experience in embedded electronic product development at technology companies such as DS2, Embention and Aerialtronics, he is an expert in Edge AI, semiconductor technologies, FPGA architectures and functional safety.

He currently leads pioneering initiatives in embedded and sustainable AI hardware, including the AceleradorSNN and NeuroAI4Space neuromorphic cognitive systems.

13:30 – 14:30

Lunch

14:30 – 16:00

Session 3

  • AI Safety Engineering Across Automotive and Industrial Robotics Domains
    Majed Mohammed
  • HOLD: A Deterministic Safety Monitor for Mission Critical AI Systems
    Omar Lone and Hans Dermot Doran
  • Towards Safety-Centric AI-Enabled Autonomous Driving Systems: a Proof of Concept Implementation
    Eduardo Lopes, André Santos, João Almeida, Jatin Arora, David Pereira and Joaquim Ferreira
16:00 – 16:30

Coffee Break ☕

16:30 – 18:00

Closing Session

We invite the submission of papers with high quality research contributions, work in progress, experimental and ongoing projects results. Therefore, the following types of submission are accepted:

  • Short papers: max. 6 pages, including references. These can be on new and emerging results, describing challenging problems, tool demonstrations, work in progress or industrial experiences.

  • Research papers: max. 12 pages, including references. Reporting substantial, completed, and previously unpublished research.

Workshop papers will be reviewed by at least three independent reviewers. 

Accepted full workshop research papers will be included in the complementary book of the SafeComp 2026 Proceedings.

Templates for paper preparation can be downloaded from: https://www.springer.com/gp/computer-science/lncs/conference-proceedings-guidelines

We invite the submission of papers with high quality research contributions, work in progress, experimental and ongoing projects results. Therefore, the following types of submission are accepted:

  • Short papers: max. 6 pages, including references. These can be on new and emerging results, describing challenging problems, tool demonstrations, work in progress or industrial experiences.
  • Research papers: max. 12 pages, including references. Reporting substantial, completed, and previously unpublished research.

Submission will be via EasyChair: https://easychair.org/my/conference?conf=daies2026 

ONLINE REGISTRATION IS NOW OPEN
Registration Info

Organizers

Joaquín Gracia-Morán, Universitat Politècnica de València, Spain (jgracia @ itaca.upv.es)

Sergio Cuenca-Asensi, Universitat d'Alacant, Spain (sergio @ dtic.ua.es)

Program Committee

Marcello Cinque, University of Naples Federico II, Italy

Carlos Cruz de la Torre, Universidad de Alcalá, Spain

Daniel Gil-Tomás, Universitat Politècnica de València, Spain

Izan Catalán-Gallach, Universitat Politècnica de València, Spain

Almudena Lindoso-Muñoz, Universidad Carlos III de Madrid, Spain

José Manuel Palomares-Muñoz, Universidad de Córdoba, Spain

José Antonio Pascual-Saiz, Universidad del País Vasco\Euskal Herriko Unibertsitatea, Spain

Karthik Pattabiraman, University of British Columbia, Canada

Behrooz Sangchoolie, RISE Research Institutes of Sweden, Sweden

Horst Schirmeier, TU Dresden, Germany

Alejandro Serrano-Cases, Universitat d'Alacant, Spain

NeuroAI4Space

This workshop is organized under the umbrella of the NeuroAI4Space project, co-financed by the European Union within the framework of the European Regional Development Fund (ERDF) Comunitat Valenciana 2021-2027 Programme, through the IVACE+i Innovation calls, “Strategic Cooperation Projects” (Reference INNEST/2025/339), Project "Neuromorphic Artificial Intelligence Processor for Artificial Vision in aerospace applications" (NeuroAI4Space).

Logo FEDER 2025 ES