STMicroelectronics discusses the different levels of vehicle automation, and how they are enabling the industry on its journey towards autonomous driving.
There continues to be a strong acceleration in transitioning to serve new mobility trends, including increasing vehicle electrification, connectivity, advanced driver assistance systems (ADAS), networking, and autonomous driving, to name a few.
According to a recent presentation by STMicroelectronics, the total addressable market for automotive electronics is worth $35.3 billion in 2019. Today, around 65% of the total electronics in vehicles comprise traditional automotive core electronics, while digitalization and electrification systems account for the remaining 35%.
“Ten to 20 years ago, it would probably take 15, 20 years, or even more, to change this partitioning,” says Davide Bruno, Head of Marketing and Application for MEMS Analog, MEMS and Sensors Group (AMS), Asia Pacific, at STMicroelectronics. “With the current speed of new projects, new developments and new technologies, we will move to more than 60% of digitalization and electrification, and less than 40% will be for the traditional automotive core electronics. This will happen in just three to five years from now.”
STMicroelectronics is already engaged in all of these applications—such as ADAS vision and radar, networking, positioning and connectivity, and many other aspects of vehicle electrification. “The key point here is that the speed at which we are going to change and shape the automotive market today will be very fast—even faster than before,” says Bruno.
COVID-19: Challenges and Opportunities
The COVID-19 pandemic this year, however, has significantly impacted the automotive market, with global car production expected to decline by 20% this year from figures forecasted in 2019. But somehow, the recovery is very promising, according to Bruno.
What’s even more important is that the pandemic has somehow accelerated the transition to electric vehicles (EVs). In 2020, EVs account for only 16% of the total vehicles produced. By 2022, this figure is expected to increase to 22%; and by 2024, around 37% of all vehicles manufactured are EVs. This, in turn, is expected to result in significant increase in semiconductor content—especially sensing.
“Two new trends are coming in strongly,” says Bruno. “One is the shared mobility. By 2024, more than 10% of private vehicles will be shared mobility, which will provide a lot of opportunities to companies like ST. This means the traditional automotive model is changing, so many OEM and Tier 1 companies are rethinking the way their new cars will serve the consumers and the people that will buy the car.” Another trend is car connectivity which, together with shared mobility, are becoming a social need. This is opening up a lot of opportunities for ST in terms of products, especially sensors.
Automotive Sensor Opportunity
Vehicles nowadays are increasingly being integrated with cameras, LiDAR, long- and short-range radar, and ultrasound systems for applications such as cruise control, collision avoidance, 360° surround view, parking, navigation, safety, and vehicle dynamics—the mega systems, in Bruno’s words.
“Which sensors are we seeing in these mega subsystems? Clearly, motion sensors have been used 40 years ago for embedded safety. Environmental sensors are something new. Microphone is even newer,” says Bruno.
Such image is a very complex system, with a lot of subsystems that need to talk to each other. “Therefore, we are talking about fusion, about how different systems need to communicate with each other,” says Bruno. “This is a very important topic—the fusion of all these subsystems and sensors.”
Figure 2: The many opportunities for automotive sensors.
ST has identified four mega growth drivers: shared mobility and access control; road noise cancellation, which aims to give the passenger a better experience inside the car by reducing the noise on the road; and then the Level 4 and Level 5—5G and driving assistance—the key enablers that will allow cars to be more autonomous.
Each growth driver has key technologies that will accelerate the adoption of these new applications. For instance, in shared mobility and access control, the key is to have low-power: such as low-power accelerometers for passive key entry.
“The road noise cancellation system is a bit more complex: we have the accelerometer for detecting the wheel vibration; the microphone to sense and capture the acoustic noise inside the cabin—and we are not talking about only one microphone, but many OEM solutions that consist of four, six and eight microphones, in addition to the four accelerometers for each wheel of the car,” says Bruno. “This is a very complex and interesting trend, because you can integrate all of these different subsystems, where each has a specific sensor.”
Next are the connected cars. For vehicle-to-everything else (V2X), telematics box, car alarm, and many others, ST has been a pioneer on these technologies and, with its automotive TESEO processors, offers a full set of kits, especially for the MEMS and the inertial module unit. ST offers a 6-degree of freedom device (6DOF), in which there is a 3-axis accelerometer and a 3-axis gyroscope fully integrated into the navigation system.
But what is really changing the landscape is the driving assistance, according to Bruno. “It is no longer a dream, but a reality,” he says. “Everybody is working on this. What we have here is the fusion of systems that include radar, LiDAR and camera.” ST also has sensors targeted at these applications. In fact, according to Bruno, ST not only sells the sensors that come into the cars, but also the sensors used in when the car is being manufactured in the production line.
“The inclination and how the radar is installed in the car are very important. That is why we are selling specific devices for the inclination or stabilization for the radar during car production. In any LiDAR or radar system, the proper inclination of the whole system in regards to the car plane is instrumental for the right measurement. So you need the inclination sensor to detect any misalignment both during the installation of the LiDAR/radar itself, and during the use in the normal operation of the car, while driving,” explains Bruno.
The Future of Autonomous Driving
According to Bruno, there are five levels of vehicle automation (see Figure 3). “We used to have cars at Level 0, in which the driver fully controls the vehicle,” he says.
ST believes that one third of the cars that will be produced in 2021 will be at Level 2 and Level 3. But to move to Level 4 and Level 5, what the industry needs is a clear system roadmap. “We need more fusion between different subsystems including radar, LiDAR, and so on. Clearly, at Level 5, the drivers no longer drive—they just take the car to move from point A to point B,” Bruno says.
Figure 3: The five levels of vehicle automation.
At present, the cars are en route to Level 3 (around year 2022), while Level 5—full automation, no driver required—is expected to happen around 2040. “To reach Level 5, we need more than 30 sensors to work, in an interoperative way and altogether,” says Bruno.
Three Waves of Sensor Adoption
There have been three waves of sensor adoption in cars, according to Bruno. The first wave started with the airbag in 1974, which has now become active safety. This wave includes pressure sensors for the driver-side, dual front, and side impact airbags, sensors for vehicle dynamic control (VDC) and electronic stability control (ESC), and pressure sensors in tire-pressure monitoring systems (TPMS).
Next is the arrival of the infotainment system—this is considered the non-safety application. This involved the use of telematics, navigation, door control unit, and passive key element—for which ST is recognized worldwide—and the period between the adoption of active safety and non-safety applications defined the second wave.
Strategy Analytics noted in its market report last year the increasing adoption of inertial sensors for non-safety applications. For example, eCall, telematics, and V2X systems will see increasing penetration of inertial sensors—from 32% in 2018 to 66% in 2026. Passive key elements, meanwhile, will see sensor penetration of 55% by 2026, up from 29% in 2018. And according to Bruno, ST is well represented in these subsystems through its inertial sensor products—one of which is the ASM330 that features 6DOF.
Another trend ST is looking at in this regard is the microphone, wherein MEMS devices are expected to replace ECMs for all applications. “I think the microphone is a very interesting product for automotive. First of all, there is no automotive standard for microphones. ST is one of the pioneers that want to understand the microphone performance that will fit the automotive application,” explains Bruno. “Today, we see four kinds of applications that are driving the demand for microphones in automotive: the hands-free call, noise cancellation, the emergency call, and the in-car call applications. The MP23DB01HP is the industrial grade product we offer to our partners and customers.”
The third wave, which started with active safety, continues all the way to the eventuality of autonomous driving. “It is this wave that we are now designing with all the Tier 1 and OEM car makers,” says Bruno.
While inertial sensors for airbags and electronic stability are already being used, with the consolidated market mainly enforced by law, the fastest growing segment is coming from ADAS and autonomous driving applications.
But first, according to Bruno, manufacturers should need to understand the sensitivity, accuracy, stability, and linearity of the sensors and parameters required by the application. Because of the fact that there is not so much difference between a sensor used in a smartphone and the sensor installed in a car to achieve Level 5—autonomous driving. What is important and more challenging are the accuracy, linearity, and stability because, at Level 5, there are no margin for errors, so there should be no possibility for mistakes.
To address this issue, most of the OEMs and Tier 1 companies are using redundancy configuration for the inertial modules or sub-systems, Bruno said. “Instead of using only one device, we put together several devices and operate in the way that can give the minimal error. Clearly, this is costly and is very complex because you need to have several devices to operate together—being the same does not mean identical in this case. So, you need to have complex software to synchronize all the signals and verify all the information,” he explains. “The future is to have one single 6-axis or x-axis system, where x can be 4, 5, or 6 DOF, plus the functional safety.”
Another important element in any MEMS sensor for automotive applications is the packaging, according to Bruno. “A ceramic package gives us much more stability and better linearity. This is the key. Ceramic package is the future for automotive when we talk about safety,” he says.
When asked about the possibility of integrating some of the sensors into packages to reduce physical count, Bruno noted that this is the difference between the automotive and the rest of industries.
“The integration of more sensors in one package is a very common approach; well adopted in consumer and industrial markets. For the consumer accelerometer and consumer gyroscope, it is important to reduce the cost dramatically. You will have one 3-axis accelerometer in one single mechanical unit. The same for the gyroscope,” he explains. “But when you move into the automotive application, considering the ASIL D requirements, and your need is to have safe reading of each axis, you have to avoid any possible interfering behavior of one axis on another. So, you might prefer to not have a 3-axis accelerometer in one single mechanical unit for safety, for example.”
He further went to say that the integration in one package or even in one mechanical structure of more axes is not an issue. “This is what we are already doing in the consumer space, and our IMU is already designed in this way, or 3-axis accelerometer and magnetometer in one single package,” Bruno says. “In automotive, the point is that you want to have a reliable system, where redundancy and independency of one axis from another is guaranteed. This is why you might prefer to keep each gyroscope axis separated. It is very difficult to achieve ASIL D from the component point of view. If you have the gyroscope done with only one mechanical structure, you will need 3 dies, one for axis X, one for Y, and one for Z to achieve the ASIL D safety standard.”
Drowsiness, Driver Distraction
The industry is still far away from autonomous driving, so, technology providers are turning towards addressing issues such as drowsiness and driver distraction to help save lives.
Drowsiness is responsible for 20% to 25% of all car crashes in Europe (INVS/AFSA). Meanwhile, more than 50% of drivers are texting while driving. And while we do have the capability to drive while texting, statistics and realities show that this can cause the death of many people.
ST is helping the automotive industry save lives by releasing a new family of image sensors especially engineered to add high resolution vision to machines. “Together with proper processing and systems, we can recognize the degree of attention from our eyes and facial expressions when we drive, so, when we are not able to be in control of ourselves but need to control the car, the system will enable us to pull over, stop the car, and take a rest. It is something very interesting and I think ST will contribute a lot to it,” says Bruno.
One of the key enablers of autonomous driving in the future is the LiDAR, which is compact and cost-effective semiconductor solution for ADAS. ST is cooperating with LeddarTech, to create a LiDAR Evaluation Kit that will demonstrate technical concepts and offer development capabilities in a functional LiDAR for automotive Tier 1-2 suppliers and industrial system integrators to develop a LiDAR solution based on LeddarEngine technology. ST is providing an Emitter Submodule, a system made by MEMS, mirror and optics.