Blog post

Would you like that when you do a maneuver at 100Km/h, an attacker compromises your system? Would you like then to have a backup safety mechanism to stop your vehicle? 

Imagine yourself crossing at a crosswalk on a city. How much trust would you have on a vehicle that drives automatically? Would you like that the vehicles are simulated in design time and verified, monitored, and audited all the time? 

What if the system can detect anomalies and it is able to learn how to heal itself? And if all the info is collected, analyzed and visualized in a dashboard? You may feel more secure and safe. 

Connected and Collaborative Automated Mobility (CCAM) depends on interactions between systems. Let’s ROBUSTify them! Let’s verify and validate them on run-time! Let’s provide the system with capabilities to detect anomalies and vulnerabilities! Let’s make the system to self-learn and self-heal. Let’s monitor all this and be aware of its trust levels. 

The SELFY project proposes a Cooperative Reslience and Healing System that provides resilience, robustness, trust levels to the system and a safe degraded mode to the vehicle by means of collaboration of a set of tools orchestrated by a VSOC. It is composed by a set of tools deployed at cloud, infrastructure (RSU) and vehicles that provide resilience and self-healing capabilities to the system.  

However, to reach this goal, it uses the Situational Awareness context information from the Situational Awareness and Collaborative Perception (SACP) macro-tool and the Trust tools from the Trust and Data Management System macro-tool. 


The brain of the CRHS (as also of SELFY) is the Vehicle Security Operations Center (VSOC). The VSOC collects information from all the SELFY tools analyzing possible anomalies, vulnerabilities, jamming situations and assign trust levels to the related CCAM elements.  

It can also trigger actions as triggering audits to a specific vehicle or a Secure Over the Air Update. The VSOC offers a series of dashboards providing real time information of what is happening. 


The robustification tool is composed of several algorithms, each tailored for a particular adversarial model. The robustification tool tackles some of these cybersecurity problems from different angles.  

Depending on the adversarial model (the class of cyberattack, e.g., denial of service, jamming, or false data injections), physical models of the vehicles, and networking scheme, we provide algorithms to make vehicular platooning robust against cyberattacks. 


The Interaction-based Tool (IVT) provides utilities for drawing, manipulating and exploiting interaction models (akin to UML sequence diagrams or Message Sequence Charts). These models are adapted to specify high-level behaviours and V2X communication flows in CCAM systems in which communication networks and information technology provide smarter and more efficient connected and cooperative driving.  

The solution combines interaction models and techniques of Runtime Verification (RV), which can be used to model and analyse the security and robustness of CCAM system at the operation phase. 


In general, Artificial Immune Systems (AIS) are computational models inspired by the biological immune system of living organisms, mainly the human immune system. AIS are used in computer science to solve complex problems and, in particular, we use them within the SELFY project in the area of anomaly detection. The AIS Module can learn and mitigate, handle an attack or detect an anomaly behavior (deviation) by himself, providing characteristics to the vehicles such as self-learning and self-healing. 



The Situational Assessment Module is a system that detects anomalies in the traffic environment. It uses data from roadside units (RSUs), vehicle sensors, vehicle Controller Area Network (CAN) messages to detect anomalies. It applies artificial intelligence techniques to analyse the data and identify any abnormal events, such as misuse, malfunctions, or hazards. Based on the analysis, it assigns a risk level to the current situation.  


The Audit Box is attached to a RSU and audits all the surrounding vehicles passing nearby. It has direct communication with the VSOC, and it can be configured to scan determined vehicles with different parameters. When it detects a vulnerability on a vehicle it informs immediately the VSOC, as well as if it detects jamming situations near the RSU. 

Audit Box Concept


The Safety Operational Tool is a system that evaluates the situational and own risks of the ego-vehicle and modifies the planned trajectory to a minimum risk maneuver when needed.  

The tool provides an alternative solution to localization when the primary source, typically the Global Navigation Satellite System (GNSS), fails. It selects the most suitable operation mode according to the environmental circumstances and the ego-vehicle status and handles Minimum Risk Maneuver calculations. The goal during fail-safe operation is to perform a controlled stop at a suitable location. 

This tool is going to be validated in a real-world environment, but previously it is being validated using CARLA Software.  

Author: Eurecat