Abhay Samant of NI tells Jonathan Newell about the need for a highly flexible approach to prototyping and testing sensor systems for the cars of the future.
Think of the new cars that were available four years ago. Not much different to those that are gracing today’s showrooms? Except for maybe the absence of a few “nice to have” bells and whistles, the option of a hybrid version and maybe autonomous braking.
If you think the same will be true for the next four years…. think again. A battle is being waged between the top global automotive suppliers to deliver advanced connectivity, driver assistance technology, autonomy and intelligent mobility. Before the end of the decade, these are set to become the defining factors in the purchasing decisions of discerning private buyers and fleets alike.
RF and sensor proliferation
To deliver this technology, carmakers are finding themselves up against technical hurdles that just seem to get higher as the enabling technologies emerge and then become redefined to fine tune them.
The key elements are the connectivity environment and the sensors that are used throughout the car to monitor its environment. Both are proliferating at an extraordinary rate, with HORIBA MIRA having recently described the cars of today and the immediate future as “the largest and most complex piece of consumer electronics that anyone is likely to purchase.”
Infotainment consoles, Advanced Driver Assist Systems (ADAS), automatic emergency call technology and infrastructure communication systems use an array of information and communication technologies – from the familiar CANbus to cellular 4G, Bluetooth, Dedicated Short Range Communications for vehicle-to-vehicle and vehicle-to-infrastructure (V2x) communications and 802.11 wireless. If that weren’t complex enough, the RF environment is constantly changing, with emerging technologies such as 5G and the multi-gigabit per second 802.11ad WiGig standard.
National Instruments has long been involved in the development of test and prototyping platforms and counts the automotive industry as one of its major sources of customers. In support of the industry, the company recently attended the Automotive Testing Expo in Stuttgart, where Dr Abhay Samant, the Section Manager of RF Product Marketing at National Instruments demonstrated how NI’s platform based approach can be used on the connected car.
In what turned out to be a coincidental case in point for demonstrating the usefulness of vehicle connectivity, I spoke to Samant using Voice over IP, as he was being chauffeured through the German countryside to his next event.
The now commonplace Engine Control Unit (ECU) isn’t the only on-board computing system in a modern car, with most of them having up to 10 such control units for telematics, brake control, active safety and other in-vehicle functions, all coupled to each other and the outside world with an array of communications protocols. When added to the proliferation of sensors in driverless and semi-autonomous vehicles, the testing and prototyping become critical. In this respect, a platform based approach provides the industry with a means to support the entire value chain from early design through prototyping to final production testing.
“Testing is complex and critical with vehicle sensors and other systems having both software and RF. A platform-based approach provides the best means of achieving the test requirements throughout the value chain,” said Samant.
HIL testing enables specific elements of a complex system to be “slotted” into the test bed and have the full range of test cases applied to them. As products are developed, simulated hardware can be replaced with real components as they’re created.
I asked Samant how NI handles the complexities of the various industries it works in and in this case the rapidly evolving automotive industry. “Domain expertise is a crucial element since it’s a highly evolutionary industry and standards are constantly emerging and changing. NI experts define and support all the elements of the platform to make sure that all the test cases can be and are put together correctly,” he said.
Connectivity and sensors
The test cases for sensors can become complicated, because, apart from having to emulate different targets, as ADAS technology evolves new targets are identified. Pedestrians, cyclists, animals and roadside furniture all present different challenges, especially if moving.
The ability to cover all scenarios is something that insurance underwriters are particularly interested in, to ensure that factors influencing liability in the use of autonomous vehicles have been taken into account. This means that the test platform has to be flexible enough to be able to take adjustments for new test cases.
There is also the complication of sensor range variation, which is dependent on environment and connectivity. Commonly, LiDAR sensors operate over close ranges, cameras operate up to 100m and radar up to around 300m in rural environments. These ranges decrease in urban settings.
Communication technology has the ability to expand these ranges too and 5G has been tipped as the means of extending sensor range through community sensing, exploiting the cloud infrastructure and increased bandwidth.
However, according to Samant, although changing from dedicated short range communications to 5G has the ability to increase scope, there are disadvantages such as security which will require different implementations of the technology.
Currently, the radar bandwidth is between 800 and 1200MHz, but an increase to 2GHz has been predicted to be possible within a few years. This will offer finer resolution in complex urban environments.
5G and autonomy
National Instruments has the advantage that it has been involved in the early stages of the development of 5G technology and works closely with global telecommunications companies. It is also involved in the early development of autonomous vehicles and so has insights that span the boundaries between the industries.
With the right implementations and the availability of secure bandwidth at low latency, 5G has the potential to be the unifying communication technology for driverless or highly autonomous transport.
Many companies are already hanging their hats on the 5G peg for safety-critical vehicular control functions. Goods vehicle manufacturer, Scania and telecoms company, Ericsson have recently entered into a partnership to evaluate the use of 5G for vehicle platooning, which involves heavy goods vehicles being operated in close formation to reduce drag, improve fuel efficiency and reduce emissions.
Since the vehicles are driven on trunk routes closer to each other than would allow the driver to be able to react quickly enough, platooning relies on extremely low latency vehicle-to-vehicle communications so that all the individual vehicle control system can act as a single unit.
This is one example of how semi-autonomous control, sensor technology and communications technology are combined in a safety-critical application that would benefit significantly from a platform based approach to prototyping and testing.
The combination of 5G and autonomy is being tipped as being the driver of the next generation of test beds. “Those who operate a platform based approach will really help their companies to navigate through the complex test scenarios of RF and software performance”
Read more on Scania’s approach to 5G networks for V2x communications