The automotive industry is undergoing a period of intense innovation, with a supply chain invigorated by the possibilities of increased automation in cars. Self-driving capabilities have the potential to redefine the car and its usage in society, with people’s use of their commute time being freed up for other activities and their idea of car ownership being re-examined.
There are barriers to get to this autonomous future though: Cost is a major factor; the expensive compute and sensor solutions currently being tested in prototypes around the world need to be made more affordable.
Compute capabilities unique to the constraints of automotive applications are required to move from prototype systems currently being tested in limited numbers to mass deployment. A large amount and different types of processing power is required to collect data from the sensors, perceive the environment of the car, decide on and perform the appropriate action. Energy efficiency is a key consideration so the compute power in the vehicle does not reduce range or create too much heat for the car to radiate away.
Safety of the autonomous systems is of prime importance, with redundancy and functional safety capabilities a must. Security technologies must be employed as well as safety, to protect the complex systems from attack and manipulation.
Autonomous driving is a complex algorithmic task, which means that software size and complexity can be daunting. The elements of the autonomous software stack needs careful consideration.
Arm sits at the centre of an ecosystem of partners who have been supplying the automotive industry for over two decades. This ecosystem can provide the processing solutions, software technologies and meet the safety requirements to deliver the autonomous future at scale.
