The modern car is rapidly changing from a chunk of complicated steel artwork to a sophisticated machine with huge amount of electronics. The innovations happening in this piece of engineering marvel are mostly electronics driven and focus on providing features like increased safety, comfort, efficiency and reduced emissions.

This drive of innovation through electronics in cars calls for anywhere between 30 to 100 small computers performing functions like managing fuel efficiency, transmission, steering, anti-locking brakes, vehicle stability etc. These computers inside the car are called as Electronic Control Units (ECU).

An ECU is like an individual computer with electronic hardware such as the micro-controller, and also a host of software including the operating system, network components, device drivers and application software embedded within. Each of these ECUs is in a network with all others bringing in additional software complexity of communication, network management, diagnostics etc.

Most of the modern cars now have several million lines of software code. It is expected that soon over 40% of the car cost would be electronics and a large part of that would be contributed to software. Most of the software is safety critical and requires real-time performance.

The high industry demand for innovative solutions is a result of:

  • Increasing electronic content
  • High quality demands mixed with complexity of applications
  • Constantly reducing cost of the vehicle

The early trends and the motivation for change

A decade ago the ECUs in a vehicle were designed to meet OEM specific and Vehicle specific needs, the software in this case was monolithic and proprietary. This included the device drivers, communication software, other interface software and control application software that the ECUs were supposed to perform for example, engine control. While the software was quickly built and was optimised to deal with resource constraints such as memory and bandwidth it started becoming extremely expensive to maintain and reuse such software.

Any new feature addition or reuse for newer vehicles meant substantial changes, and lot of expense and development time. It also increased the risk of inducing newer defects.

Each OEM started dealing with the challenge by taking a layered approach to the software development (separate the application layer, communication layer and the layer that interfaces with the hardware). Each OEM started dealing with the challenge by taking a layered approach to the software development (separate the application layer, communication layer and the layer that interfaces with the hardware).

Experience and cost advantage

This layered approach to the software development opened a space for software service providers, who had specialised in working with OEMs, Tier 1 Automotive suppliers and Semiconductor companies, to develop ECU software

As the cost pressures and the content of software in a vehicle increased, the OEM car makers worldwide started looking at standardising the software architecture of the ECUs, to facilitate software reuse across vehicles & across subsystems. As a result of this quest an industry forum called Autosar (AUTomotive Open System ARchitecture) was formed by the various industry leaders. This forum has been extremely active and today is set to redefine the trends for future vehicle software. The following is the overview of the Autosar architecture. (For more information, visit ).

The standardisation of the interfaces is expected to pave the way for plug-and-play of various software products from different vendors to work together in the ECU allows for:

  • Major cost amortisation
  • Reduced development time
  • Better quality

With most of the industry players actively driving this in Europe, the desire of the industry to reduce cost and development time seems to be now taking shape. Most of the OEM car makers and Tier 1 suppliers are part of the Autosar consortium and appear keen to see the success of this initiative.

On the flip side like any major initiative of this scale, there is widespread speculation about how successful this one would be. However, for past couple of years if the buzz at the key industry forums is any indication, success seems to be on the way.

For software companies operating in this space the development of Autosar was a great opportunity to leverage change. The companies involved are currently involved in three major tasks:

  • Working with leading Semiconductor manufacturers to provide Autosar based software services.
  • Developing its own Autosar based vehicle network products to be available for world market in co-operation with a leading car maker.
  • Offering services to migrate ECU software to Autosar standards with its vast experience in this area and knowledge of Autosar.

Anup Sable is vice president of automotive and allied embedded and tools at KPIT Cummins