Vehicle manufacturers and their tier1 suppliers are now faced with new challenges in the development of nextgeneration smart vehicles. To support the future of networked mobility, they need reliable and redundant connectivity solutions for high-speed networks, with seamless end-to-end integration encompassing hardware, software and services.

Automotive engineering is facing a paradigm shift that affects not only vehicles themselves, but also how we relate to vehicles and how vehicles interact with their surroundings. As in other sectors, the impacts of digitalization and new business models are noticeable in the automotive industry. According to the study “Car of the Future v4.0” by Citi, we can expect the automotive industry to be characterized by four vertical markets up to 2030. 

These in turn correspond to four vehicle types: robotaxis (mobility on demand), autonomous rideshare vehicles (driverless vehicles), a combination of robotaxis and autonomous rideshare vehicles, and traditional ownership. Connected Mobility encompasses a variety of applications and technologies that represent the future of our smart networked urban transport infrastructure, including advanced driver assistance systems (ADAS), networked infrastructure and vehicle-to-everything (V2X) networking technologies, which make autonomous driving (AD) vehicles possible. However, these technologies are based on very high-performance, reliable and timesensitive high-speed networks, which are currently not yet used.

SPEED IS ESSENTIAL

Within a single generation, electronics and computer systems have changed nearly all aspects of our lives, including vehicle engineering. The scope of electronics, software and processing power in vehicles has changed not only how our vehicles operate, but also how we perceive them – more like mobile computers than machines. The number of control units (ECUs) in vehicles has grown from a handful to a hundred, or even more in top-end vehicles. They control all functions, including engine and energy management and traction control, as well as more recent developments such as active blind spot warning, lane departure warning and adaptive cruise control. At the same time, there are more and more interior and exterior sensors, sophisticated telematics and information systems on board, along with numerous electronic control units and various automotive networks (such as CAN, LIN, FlexRay and Ethernet) for data transport and control.
This trend is accelerated exponentially by the constantly increasing networking and automation of vehicles. Enormous computing power, bandwidth and speed are required to process this massive amount of data in real time. This data volume is generated by numerous sensors including camera, radar, lidar and ultrasonic, and must lead to safe decisions in just a few milliseconds. This requires a universal, standardized and reliable end-to-end network architecture, because any delay or latency in the transmission of critical data in autonomous vehicles can potentially have fatal consequences for vehicle occupants, other road users or pedestrians.

The estimates vary, but in any case highly automated vehicles can generate up to 40 terabytes of data per day. 5G technology is regarded as a key function for highly automated driving. It has the potential to connect vehicles and other equipment in our surroundings, such as IoT devices, to a highperformance data network that provides the speed, response time and range necessary to support the functions of autonomous vehicles. Semiconductor companies are currently developing new microprocessors that enable data transmission with high bandwidths up to 10 Gbps or more with multizone vehicle architectures (Figure 1). This will change vehicles from mobile computers into data centers on wheels and enable autonomous vehicles to make complex decisions in real time.

AUTOMOTIVE ETHERNET

In the wake of constantly changing demands on existing automotive networks, Ethernet has developed into one of the most important protocols for high-speed data transmission. The demands on data rates are constantly rising, leading to the issues of functional safety and data security in networked environments and the desire for standardized intersectoral solutions. Ethernet is a mature, stable and proven technology that offers high speeds and security. However, Ethernet architectures must work under harsher conditions in vehicles than in most conventional Ethernet implementations, such as in data centers. Consequently, such networks must be much more robust because they are exposed to variable temperatures, humidity, water, vibrations and shocks and must withstand electromagnetic interference (EMI).

MULTI-ZONE VEHICLE ARCHITECTURE

Distribution of all the electronic and electric devices into several zones allows existing and new communication protocols, such as CAN, LIN, LVDS and Ethernet, to be integrated into the Ethernet-based vehicle architecture. Special gateways form the interface between these networks and the Ethernet architecture. They aggregate the sensor data and forward it to one or more ADAS/AD processors, where decisions are made and the resulting vehicle commands are transmitted over the high-speed network to the gateways of the individual zones. The gateway forwards the commands (signals) to the control units.

AN EXAMPLE OF THE DISTRIBUTION OF VARIOUS NETWORK TECHNOLOGIES

A wide variety of sensors, such as radar, lidar, ultrasonic sensors and high-resolution cameras, continuously send data about the surroundings and are connected over LVDS and Ethernet (for example) to the gateway. For example, the rear-facing sensor strip (with various sensor technologies) continuously sends image data from cameras and other sensors to the gateway of the rear vehicle zone. The gateway aggregates all incoming data into a large data stream and sends it in real time to the ADAS/AD server over 10 gigabit Ethernet. There the software makes a decision about the signals from the sensor strip. As an object is blocking further travel, it issues a brake command that is sent to the brake control unit, which in turn actuates the brakes to stop the vehicle. All this takes place in real time. Forwarding of the sensor signals is prioritized so that this critical data stream cannot be impaired by other network traffic. The automotive industry relies on Time-Sensitive Networking (TSN) for this sort of real-time capable communication links. TSN involves a set of standards that extend Ethernet with functions for real-time data transmission. The objectives are assured end-to-end latencies, precise time references, minimal packet loss rates, and high availability of network connections. In this way TSN ensures the desired prioritization of network traffic so that safety-critical functions.

cannot be delayed by other traffic, such as a data stream from the infotainment center. 

The required level of safety is achieved through a multi-zone gateway architecture with redundancy to enable a fail-safe functional network. The required safety is additionally achieved by a multi-layer safety approach with device certification.

Plug-and-socket connectors and cables that must ensure signal integrity for the entire life of the vehicle are one of the decisive factors for the reliability of high-speed network connection. Reliability must not be impaired by electromagnetic interference (EMI) or by thermal or mechanical processes. The comprehensive 10 Gbps network solution from Molex fulfils this demanding task by combining sensor, control and infotainment data with integrated multi-zone redundancy and Time-Sensitive Networking (TSN) capabilities for maximum reliability. 

A multilayer safety approach with device certification ensures compliance with the strict safety requirements of vehicle manufacturers. The full potential of Ethernet is by no means exhausted by the new 10 Gbps network system. There is growing demand for even higher speeds, and future network systems will be oriented to these increasing speeds. For the Connected Mobility network system of the future, Molex is cooperating with Accenture, Allgo, Amazon Web Services (AWS), Aquantia, BlackBerry (QNX and Certicom), Broadcom, Cypress, Excelfore, Molex CVS, Microchip, Texas Instruments and Rosenberger and is a member of the NAV Alliance.