Automakers are increasingly using image sensors and graphics chips to build advanced driver-assisted systems (ADAS), like automatic braking and blind-spot warnings. In its latest attempt to bolster its ADAS processors, Freescale recently agreed to buy Cognivue, a company that designs image cognition chips.
The deal was the final step in an ongoing collaboration. Freescale has worked with Cognivue to write algorithms that can identify objects in photographs and video. Freescale said in a statement that it would incorporate similar technology into new “automotive-grade” chips.
The terms of the deal were not disclosed. It is also not clear how NXP Semiconductors’ acquisition of Freescale will affect the deal.
Freescale had also used Cognivue’s image processors in its S32V automated driving technology, which is currently in pre-production. The company's APEX2 processors can identify images from four cameras located around the vehicle and extract features used to classify objects. The processors are linked to a graphics chip, which constructs a three-dimensional model of the vehicle's surroundings.
Patrick Morgan, Freescale’s vice president and general manager of ADAS Radar and Safety, has said that these models can be combined with machine learning programs, allowing vehicles to prevent collisions by taking control from drivers.
The deal follows the growing market for sensor technology in automobiles. According to IHS Automotive, around 83 million sensors are expected to be plugged into vehicles in 2020, more than five times the number of sensors used by the automobile industry in 2012. In addition, the market for advanced driver-assisted systems is expected to earn revenues of around $15 billion in 2016, according to technology research firm Specialty Analytics.
Freescale says that it has shipped more than 20 million sensors, microcontrollers, and microprocessors for advanced driver-assisted systems. The company expects to increase that output significantly with the Cognivue acquisition.