Bus Systems From a Distributed to a Zonal E/E Architecture

Modern vehicles are packed with electronics, control units and sensors and cameras. They all need to communicate with each other. Data traffic on the buses is increasing.

Three major trends are currently revolutionizing vehicle design: automated, electric, and connected driving. To enable these huge advances, a redesign of the E/E architecture is needed. Today's vehicles and their common E/E architectures consist of up to 150 separate ECUs and complex, weight-intensive wiring. Future technologies such as automated driving (ADAS) require more functions and therefore more sensor data and ECUs to implement each function.

Many electronic control units (ECUs) currently process the majority of sensor information locally. To cope with the growing complexity of vehicle functions, increasing data volumes, and increasing computing power, automotive manufacturers and suppliers are striving to transition from a distributed to a zonal E/E architecture. This new concept results in new network requirements.

Zonal modules are optimally placed in the vehicle and take over communication into and out of a zone. This requires a backbone system that distributes and collects data in the vehicle system. The sensors in the vehicle generate hundreds of megabytes of data per second that must be exchanged between the electronic control units with predictably low latencies (in the range of microseconds to milliseconds).

Automotive Ethernet Time Sensitive Networking

Traditional in-vehicle communication networks allow low-latency communication, but do not meet the high-bandwidth requirements. Due to this limitation, a high-bandwidth, low-latency network protocol is needed to support backbone communications in modern vehicles. Ethernet has been used for quite some time for diagnostic purposes, for feeding data from image sensors and for vehicle infotainment systems within vehicles.

By adding deterministic timing functions, the range of applications for Ethernet can be expanded considerably. The IEEE Time Sensitive Networking (TSN) working group has developed a set of standards to meet the requirements for in-vehicle communications with high bandwidth and low latency.

Advances in TSN standards include support for clock synchronization, resource reservation for different types of traffic, various traffic shapers, support for scheduled traffic, frame preemption, and network management mechanisms. In the future, the network technology Automotive Ethernet and the protocol extension TSN (Time Sensitive Networking) could further bring about significant changes if the vehicle network is divided into subnetworks of spatial zones in which special control units, so-called gateways, realize the connection of local control units.

Traditional automotive bus systems transition

The classic CAN variants have proven their worth for over three decades in various applications, especially in vehicles. With CAN XL, a new variant of the data bus is in the starting blocks, which is intended to close the gap in transmission rates between signal-based communication and automotive Ethernet technology. Not only because of the hardware costs and high transmission reliability, CAN XL is interesting for all users who want to continue their developments based on existing architectures.

In 2018, initial work on the development of the third generation of CAN-based data link layer protocols was started within the framework of CiA with the active participation of Fraunhofer IPMS. CAN XL offers a maximum payload of 2,048 bytes per frame. By means of an "acceptance field" it is now possible to separate the priority function (11-bit priority field) from the address & content function (32-bit acceptance field), which is significant for use in automotive applications, for example.

With a targeted data rate of up to 20 Mbit/s and a maximum payload of 2048 bytes, CAN XL allows the tunneling of TCP/IP packets, thus achieving significantly better interoperability, and opening up for new types of applications. Nevertheless, users can still benefit from the advantages of the classic CAN protocol when using CAN XL:

• Arbitration prevents collisions (i.e. lossless prioritization),

• Robust error handling,

• Low costs for transceivers,

• Very low costs for cabling and

• Almost any topology possible for cabling.

Fraunhofer IPMS IP Core Solutions for the Automotive Industry

Different components in vehicles require different bandwidths and latency, and the complexity of networking continues to increase as the number of ECUs used in vehicles grows. Different approaches in the automotive industry with regard to cost efficiency and flexibility require different network technologies in the vehicle in order to optimally adapt to the increasing requirements of the coming years, both technically and economically.

With a CAN and LIN controller IP cores, an IP core family for Time Sensitive Networking (TSN), Low Latency Ethernet MAC core and a 32-bit RISC-V IP processor core, Fraunhofer IPMS offers a comprehensive range of IP cores to meet the challenges of the new network architectures.

These automotive IP designs are offered platform independent and are suitable for integration into FPGAs as well as ASICs. Fraunhofer IPMS provides comprehensive manufacturer- and technology-independent consultancy on all development issues and adapts and continuously devel­ops the IP cores according to the customers' needs. The multidisciplinary IP design team is available as a competent development partner in every development phase. A special focus is on safety-critical applications.

The use of ASIL-D

The ISO 26262 standard specifies the requirements for safety-critical electronic systems in the automotive sector so that hardware failures can be prevented and brought under control. The use of ASIL-D ready pre-certified IP cores from Fraunhofer IPMS significantly simplifies the approval process of the overall system, as relevant documents such as the FMEDA (Failure Mode and Effects Analysis) and the safety manual are already available and necessary safety functions are fully implemented and tested.

This guarantees the minimization of costs and time required for system development. SGS TÜV Saar is a certification partner for Fraunhofer IPMS IP Cores in the automotive sector.

Demonstration Multiprotocol Automotive Subsystem

The advantages of CAN bus systems in combination with other protocols offer a wide range of applications. As an example, an automotive networking solution is outlined below — a Fraunhofer IPMS Automotive Networking Subsystem. The Fraunhofer IPMS EMSA5-FS uses among others the CAN-CTRL IP Core as peripheral device. The EMSA5-FS is a RISC-V processor core for functional safety and is designed as a 32-bit processor with a five-stage pipeline according to the open RISC-V instruction set architecture (ISA). The EMSA5-FS IP Core is available as a stand-alone processor or as a pre-configured subsystem.

Automotive bus systems require guaranteed low latencies and sometimes high bandwidths for the transmission of user data. New communication protocols such as CAN-XL or Ethernet-TSN are used in the context of automotive Ethernet in parallel to classic protocols such as CAN-FD, CAN2.0 or LIN. Bridging and Interbus switching is also integrated so that all the advantages of various bus systems can be used in automotive application scenarios.

Data packet between CAN XL, Ethernet and TSN

Gateways allow, for example, the bridging of data packets between CAN XL and Ethernet and vice versa, as well as the integration of real-time standards such as TSN. This allows real-time data streams from powertrain, chassis, cabin and infotainment, each of which has different specific requirements, to be combined in efficient E/E architectures.

The targeted combination of diverse protocols thus enables a tailored, application-oriented solution. Figure 1 demonstrates this type of customer-specific automotive networking subsystem, which combines not only CAN but also LIN, TSN and the EMSA5-FS controller for management tasks in one system.

* Monika Beck is Research and Development at Fraunhofer IPMS, Dresden.

(ID:48689998)