ADAS (advanced driver assistance systems) and AD (autonomous driving) are now some of the main directions for the automotive industry to improve safety and comfort for drivers and road users.
These systems depend on sensors to gather information processed by sophisticated computers. Currently—and likely for quite a while—cameras are the main sensors used, as they can give a reasonably precise view of the environment at an affordable cost. Other sensors found on cars include radars, lidars, and sonars. Radars are often used for functions such as AEB (automatic emergency braking) and BSD (blind spot detection). Lidars are just emerging for front detection, and sonars are mainly used for parking assistance. These sensors are chosen for specific functions or to improve the reliability of the information given by cameras, but none has the versatility of cameras, which are thus indispensable for ADAS and autonomous driving. There is also a cost factor, and cameras are competitive in that sense; low-range cars, because of their build cost constraints, tend to use only cameras to drive some very useful functions for safety as AEB. Medium-range cars’ systems still use cameras, but bolster performance and accuracy with radars and rangers that are usually simplified lidar with one beam. And high-range cars are beginning to use lidar, though at this point it remains very expensive. The development of autonomous cars at levels 4 and 5 will likely see the use of cameras, radars, and lidars jointly in the same car for redundancy, which is obligatory to achieve adequate safety performance and dependability.
The global market for cameras is currently dominated by the smartphone market with roughly 90% of applications. The market for automotive cameras was around USD $2bn and 10% of the global camera market in 2016, but the automotive camera market is now growing much more rapidly at a rate of 20% per year in volume. Present projections have the market reaching 800 million cameras in 2030. This market is pulled by end users’ strong interest in the real safety benefits of ADAS and AD, and it is pushed by the assessments and regulations already applied or under preparation in different countries.
Front cameras are the most complex ones in automotive service, and Yole estimate their market value at $10bn by 2030. Other cameras, particularly surround cameras, are forecast to make another $10bn market in 2030.
Technical progress for cameras is under way on all their main components. The optics will shrink, perhaps with plastic lenses if thermal issues can be solved, or even with no more lenses over the longer term. The sensors will come to have higher definition, much above the current best 8-megapixel level. We’ll see more and more sophisticated solutions integrating several sensors working together for enhanced possibilities, for instance with 3D, distance measurement, very high sensitivity towards quantum imaging counting each photon, and multispectral systems for better recognition of objects. We’re also likely to see image processing and, more generally, sensor fusion systems and artificial intelligence becoming one of the most differentiating factors of competitivity, given that many actors such as the leader Intel/Mobileye are investing hugely in R&D.