The ADAS are on their way to becoming an integral part of the vehicle—with interfaces to many different clusters of electric/electronic systems in the car. Comparable to the human body, numerous functions have to be implemented and networked: sensors such as cameras, radar and ultrasonic, as well as processing units and actuators. Taking into account the complexity of the use cases and the different vehicle areas that have to exchange information, it is obvious that an adequate network infrastructure is of essential importance for the efficiency of the system. From a functional point of view, driver-assistance systems have started to enlarge the range of classical infotainment systems.
As shown in Fig. 1 “Evolution of E/E Architecture” the driver assistance domain is becoming an integral part of the E/E ecosystem. ADAS and infotainment systems are expected to grow together in the future.
Fig. 1: Evolution of E/E Architecture
Typical emerging driver-assistance applications include:
- Parking Assistance
- Collision Warning
- Traffic-Sign Monitor
- Lane-Departure Warning
- Advanced Lane Guidance
- Pedestrian Warning
- Night Vision
- Adaptive Cruise Control
- Pre-Crash Warning
Driver assistance systems lead to special requirements at the network level.
These are characterized by:
- Transfer of control, video, packet and IP data
- Highest quality of service
- Hard real-time determinism and low latency
- Flexible topologies, such as star, daisy chain and ring
- High bandwidth
- Remote-Control features
- Safety aspects
- Robustness and maturity
Beyond this list, cost efficiency is a key requirement.
The next section demonstrates that a multi-channel network approach with an underlying TDMA mechanism is an excellent choice.
Typically, driver-assistance systems have to deal with a variety of sensor data. In order to cope with the complexity, you often find a hierarchical approach with different abstraction levels and timing constraints. On the lower level, there is a high amount of raw data where high bandwidth as well as coherent and fast transmission is required. On the medium level, objects and attributes need to be transported. Finally,