**3. Cybersecurity and real-time**

In the context of autonomous driving, ensuring cybersecurity and real-time capability is crucial. With the increasing networking and automation of vehicles, new challenges and questions arise that will be discussed in this section.

A central problem is ensuring end-to-end cybersecurity under real-time conditions. To do this, security measures must be implemented at all levels of the system, starting with the sensors and extending to communications and the cloud.

Examples include authentication and key exchange under real-time conditions. The typical use of asymmetric crypto methods is problematic for key renewal during

#### **Figure 2.**

*Challenges of parallelizability in key exchange with asymmetric cryptography.*

*Autonomous Driving and Cybersecurity by Design DOI: http://dx.doi.org/10.5772/intechopen.112042*

#### **Figure 3.**

*Example of parallelizing a MAC calculation.*

runtime due to their slow runtime. So, if you want to still have real-time capability, you must consider parallel key renewal during runtime (**Figure 2**). In addition, the use of parallelizable crypto algorithms can be an important building block; for example, authentication procedures, such as the Message Authentication Code (MAC) procedure, can be parallelized to guarantee real-time capability (**Figure 3**).

Another component is edge computing, where data processing and analysis take place in the vehicle instead of in the cloud, which can help optimize latency and data rates. This supports real-time guarantees by reducing the amount of data transmitted over the network and increasing the speed of response to events.

A major challenge arises from the fact that vehicles are in the field for a long period of time, so future systems should be prepared for changing crypto computing power and key length requirements by considering or balancing newer crypto techniques such as post-quantum cryptography etc. For example, the ongoing development of quantum computers poses a particular challenge by challenging the security of traditional asymmetric key exchange methods [10].

In summary, ensuring cybersecurity and real-time capability in autonomous driving is a challenging task that requires a combination of different technologies and concepts. The integration of edge computing, parallel key renewal and authentication, as well as the adaptation to future crypto requirements are key elements to ensure the security and performance of autonomous vehicles in the connected world.

### **4. Real-time management**

Our adaptive-real-time-manager (ARM) is an innovative concept that aims continuously assessing and optimizing the real-time capability of autonomous driving systems. This section discusses the basic design of the ARM and its advantages compared to existing solutions.

Factors such as vehicle environment, traffic conditions, visibility, and network connection quality influence the real-time capability of autonomous driving systems. The ARM constantly evaluates these factors and adjusts driving speed and strategy accordingly (**Figure 4**).

A crucial aspect of the ARM concept is the Round Trip Time (RTT) of the closed information chain from the vehicle's sensors and actuators to the cloud and back. The RTT varies depending on the preferred cybersecurity mechanisms, which can be selectively integrated at different security levels.

The ARM assesses the real-time capability of the respective closed information chain by considering the RTT and, if necessary, other system parameters. This enables optimal adjustment of driving speed and strategy to the respective conditions.

#### **Figure 4.**

*The ARM might suggest a speed of 50 km/h when the connection quality is good and the vehicle environment is favorable, but only 30 km/h when the connection quality is poor or the vehicle environment is more complex.*

The ARM can reduce the impact of traffic control systems on the real-time capability of autonomous driving systems. This is achieved by continuously adapting driving strategies and speeds to the current conditions and, if necessary, to the information provided by traffic control systems.

A real-world scenario illustrates this benefit of ARM: An autonomous vehicle stops before a green light at an intersection. One possible explanation for this behavior is that the intelligent traffic light has informed the autonomous driving system of the time remaining in the green phase. However, the ARM has suggested a driving speed that is not sufficient to cross the intersection without a collision, so in this case the vehicle waits for the next full green phase.

Compared to existing solutions, the ARM offers a more dynamic approach to real-time assessment and optimization of autonomous driving systems. The continuous analysis of influencing factors and the adaptation of driving speed and strategy increase the safety, efficiency and flexibility of these systems.

Another advantage of the ARM is the ability to selectively incorporate cybersecurity mechanisms at different security levels. This ensures data security and system integrity without unnecessarily compromising the real-time capability of the autonomous driving system.

In summary, the automITe Adaptive Real-time Manager offers a promising approach for addressing the challenges related to autonomous driving and cybersecurity. Continuous real-time assessment and optimization, selective incorporation of cybersecurity mechanisms, and enhanced interaction with infrastructure make the ARM a unique and forward-looking solution in this field. It remains to be seen how the ARM will prove itself in real application scenarios and what further developments and optimizations are possible in the future (**Figure 5**).

Together with the ARM there are two essential building blocks that can accelerate the integration into an ADS system:

• Cloud-Broker-Concept: This ensures independence from the cloud provider and a uniform interface on the ADS side to the cloud. An essential step here is the

*Autonomous Driving and Cybersecurity by Design DOI: http://dx.doi.org/10.5772/intechopen.112042*

#### **Figure 6.**

*Vehicle architecture with middleware and adaptive-real-time-manager in place.*

integration and the interface management of the cloud broker into the system of the ARM (see **Figure 6**).

• Simulation of the Adaptive Real-time Manager and the Cloud Broker: For this purpose, we are currently designing a driving simulator that can be used and extended to simulate the driving of a vehicle in a city, with all driving information obtained from the cloud. By using the simulator, the effort required for testing in the field can be reduced, as many shortcomings are already revealed by simulations.
