Autonomous driving: safety first

By Gina Roos, editor-in-chief

The state of the autonomous vehicle market is in flux after two fatalities with self-driving and automated driving cars in 2018. This has pushed automotive OEMs and their component partners to take a step back and do a better job of balancing safety with high-tech advanced features. A big problem is that drivers still need to watch the road, but how do these safety systems ensure that drivers keep their eyes on the windshield and not on other distractions?

The market will be huge. Allied Market Research forecasts that the autonomous vehicle market will reach $557 billion in 2026, growing at a compound annual growth rate of 39.5% from 2019 to 2026. Some of the biggest factors impacting growth — positively or negatively — include the need for improved safety and a reduction in traffic congestion, government regulations, connected infrastructure growth, high manufacturing costs, higher demand for connected cars, and the development of smart cities.

Junko Yoshida’s discussions with several industry analysts make it clear that automotive OEMs and tier-one suppliers still face big challenges in their goal to attain autonomy, causing some to scale back their expectations for autonomous vehicles (AVs) and put more focus on advanced driver-assistance systems (ADAS). Uber’s fatality in March 2018 really set back the AV market, putting technologies, safety, and development practices into question, according to most analysts.

One result is perhaps a much sharper focus on ADAS by the automotive OEMs and their suppliers. There continues to be a growing number of automated and semi-automated safety systems that are packed with electronics that use radar, light detection and ranging (LiDAR), cameras, and a host of sensors along with software to decipher the data. And there are all kinds of advanced driver-assistance safety systems that range from blind-spot warning and lane-departure warning to driver awareness detection and collision avoidance.

These new safety systems are using infrared cameras and pupil tracking to monitor driver awareness and vehicle-to-vehicle communications to share all kinds of information like speed and braking with nearby vehicles to trigger braking or collision avoidance features. All of these new systems will eventually lead to autonomous driving, but we’re not there yet.

The Department of Transportation uses SAE International’s “Levels of Driving Automation” standard that defines the six levels of driving automation. Many of the above-mentioned driver-assistance features such as blind-spot detection or lane-departure warning are no longer surprising, although there is still a high percentage of vehicles on the road that have no assisted driving features, as Semicast Research’s Colin Barnden points out. He believes that the sensors, GPUs, and sensor fusion software needed for autonomous driving at Levels 4 and 5 are cost- and power-prohibitive, but OEMs are quickly adopting lower-cost camera- and radar-based Level 2 ADAS for the mass market. So expect to see more autonomous emergency braking, lane keep assistance, blind spot information systems, and driver-monitoring systems moving into mass-market vehicles.

What’s making this happen is innovation in the component market. For example, LiDAR sensors are moving from niche applications into the automotive market with 3D data and depth resolution for ADAS and autonomous vehicles. LiDAR is now being used with cameras in automotive applications to identify a host of things like pedestrians, signs, and traffic lights. One of the biggest trends in the market is the move to solid-state technology, along with the shrinking size and cost of LiDAR sensors.

LiDAR will play a big role in the automotive market over the next several years. Market research firm Yole Développement forecasts the automotive LiDAR market to reach $5.2 billion in 2023, up from $325 million in 2017.

Sensors have been playing a big role in automotive applications, and it keeps on growing. The automotive sensor market has come a long way from the days when airbag sensors and tire pressure monitoring sensors were the big things. Today, vehicles are packed with sensors that measure things like pressure, temperature, position, speed, humidity, and occupation, and their usage will only grow as more ADAS and camera systems are installed in mainstream vehicles. Some of the newest designs include global shutter image sensors for driver monitoring systems, MEMS motion sensors (IMUs), and LiDAR sensors for ADAS and autonomous vehicles.

A part of the AV world is connectivity. The connected car ecosystem is growing, and microprocessors are at the heart of protecting these communications systems from cyber-threats. What will it take to protect these vehicle-to-everything (V2X) communications? Majeed Ahmad tells us that it will include securing V2X processors, hardware-centric security, and software to bolster security features .

At the same time, connected cars are being driven on multiple fronts — vehicle-to-vehicle, vehicle-to-infrastructure, vehicle-to-network, and vehicle-to-pedestrian. The ultimate goal is to connect vehicles to each other, the network infrastructure, and roadside traffic lights, radar, and other functions for added safety and security. Alix Paultre says that the convergence of these solutions and services will integrate the functionality at the device level more and more as core technologies continue to mature.

Now add cloud connectivity for infotainment and navigation systems, and this connectivity is going to drive the need for intelligent antennas, sensors, secure processors, ADAS, and advanced communications as well as new communications standards. What do you have now? The building blocks of a smart traffic environment, said Paultre.

The development of autonomous and ADAS technologies, which are driving big changes at the component level, is also pushing more OEMs and system manufacturers to work with non-traditional partners like cloud providers or software vendors to develop autonomous vehicle platforms. We also see more and more auto OEMs collaborating to advance AV technologies.

But there are still big challenges in prepping for production-ready AVs. Renesas’ Tim Grai explains that, when it comes to production-ready AVs, the vehicle must function continuously and safely at all times, which means that they need to be fail-safe and fail-operational. “And that is the great engineering challenge.” Grai shares five questions facing automotive engineers as they navigate the path to production-ready autonomous cars.

Cover image: Shutterstock