EMC: Making an impact

Anthony Martin and Alastair Ruddle from Horiba Mira consider how to ensure resilience among increasingly complex vehicle systems within their intended electromagnetic environment.
In 1998, the Volkswagen Golf Mark IV had 17 ECUs, two Can networks and just 434 Can signals. In 2010, the Golf Mark VI had 49 ECUs, five Can networks and 6516 Can signals. Today, typical vehicle electrical systems have more than 60 primary networked (Can or Flexray) ECUs and as many as 20,000 Can signals.
This growth in complexity over the past 20 years has been driven by the consumer’s desire for comfort and convenience, and the industry’s drive for market share through value, efficiency and safety. However, this will quickly be surpassed by the rapid integration of new technologies into future vehicles.
Autonomy
It is widely acknowledged that autonomous vehicles offer the application that artificial intelligence and machine learning have been waiting for, and that this introduction will be sooner than we think. There have been significant strides in the development of basic algorithms used in machine learning and an increase in the amount of quality data available.
Infra-red sensors, lidar systems, 360˚ vision systems, wireless connectivity and more combine to provide machine learning algorithms with a wealth of rich information.
Wireless technologies and their associated benefits are now an indispensable part of modern society. With demand increasing and implementation costs reducing, they are becoming available across most vehicles on the market.
Last year saw a significant increase in sales of internet connected vehicles including the Chevrolet Cruze and Spark (OnStar), Audi A3 and Q5 (Audi Connect) and the Mercedes-Benz GLC-Class (Mbrace). It is envisaged that this type of connectivity, along with intelligent transport systems, will deliver the much awaited connected car and the connectivity backbone for autonomous vehicles.
Fully autonomous vehicles, such as Google’s Waymo, aren’t strictly reliant on information from vehicle-vehicle or infrastructure (V2X) communications, but instead rely on existing GPS satellites and on-board sensors coupled with detailed digital maps to interpret their location and surroundings.
When implemented, V2X communications will enrich the data sources available to support autonomous driving, and enhance the sophistication of interactions between them. The availability of V2X could help overcome potential technical limitations of automated driving. For example, there are issues with GPS signal availability in underground locations such as road tunnels, and even in some locations that are above ground. Supplementary location information could be provided from roadside beacons, or from other vehicles in locations that can access GPS.
EMC
For electromagnetic compatibility (EMC) engineers, who must get to the nuts and bolts of what is required to validate these systems, a thorough understanding of the technology underpinning vehicle features and all associated risks will be essential. This may require new levels of understanding, test methods and risk mitigation processes.
From its introduction in the 1970s due to radio reception issues caused by ignition system interference, automotive EMC has developed into a key part of the development lifecycle of today’s vehicles. It is now concerned not only with quality issues that could impact a customer’s enjoyment of the vehicle but the safety of the vehicle occupants and pedestrians.
Changes in automotive EMC testing over the past 40 years have generally been related to new technologies for improved measurement efficiency. For example, Fast Fourier Transform (FFT) allowed measurement times in excess of one hour to be reduced to a matter of seconds. Additionally, the efforts of national and international committees mean that EMC testing is now more accurate and controlled providing new levels of confidence in test methods and the products they cover.
Today’s cars are more reliable, safer, durable, better performing and more efficient than they have been in the past 130 years. However, this safety and performance has largely been based on a relatively low level of complexity in vehicle control architectures, allowing for almost full EMC verification test coverage.
Current verification strategies involve one or more of the following risk mitigation methods, and more likely a complex mixture of all:
- 1. Whole vehicle verification
- 2. System verification
- 3. Component verification
- 4. Simulation
- 5. Technical review
- 6. Best practice design
Recently, the emergence of adas and the associated increase in complexity of vehicle control systems has created several challenges for EMC verification. Adas has a high level of risk due to the potentially catastrophic nature of the failure modes, making vehicle-level testing vital.
Current capabilities in EMC testing were not developed to verify autonomous and assisted driving functions, and they will require a step change in the product integrity verification process and significant changes to the facilities, equipment and methods used for whole vehicle EMC testing.
In the future, advanced verification will make greater use of early risk assessment, as well as analysis and simulation at the design stage coupled with component characterisation to reduce the requirements for vehicle verification. These will be combined to provide an argument for safety assurance (due diligence) and provide confidence in product integrity much earlier in the development cycle.
However, there must be a closer link between EMC and safety, with safety assurance being a key element in the verification of autonomous and assistance features to deliver system integrity and credibility when accidents occur. In fact, safety assurance will be the only arguable mechanism available to OEMs and tier ones in the event of an accident involving self-driving vehicles.
Standards will form a key role in the engineering process, with ISO 26262 for functional safety representing the reasons for achieving high levels of system confidence. EMC will need to be considered alongside functional safety in a more unified approach.
Effective verification and validation are required to evaluate whether the level of resilience is adequate and effective. The diversity of the automotive supply chain makes it is essential to perform this verification for individual hardware and software components, complete embedded systems and at vehicle level, to ensure that all elements are properly integrated.
It is clear that there are still challenges on the horizon yet to be fully addressed, but with a coordinated approach to safety, security and functionality, the industry will be able to manage the potential risks for connected and autonomous vehicles.
Anthony Martin is EMC chief engineer and Alastair Ruddle is a computational electromagnetics specialist, both for Horiba Mira
