With many North American car companies struggling to compete in the global auto industry, General Motors is hoping its new focus on software and system centered vehicles will help it drive innovation in the sector.
Historically, car companies have been focused on design and mechanical innovation and, according to GM, the role of software engineering has been underutilized. GM’s new strategy, however, is toward what it calls “a new automotive DNA,” which involves taking software and electronics to change the concept of car designs. This means you’ll see a lot more GM researchers with backgrounds in software engineering.
“When we talk about the new DNA, we talk about moving from internal combustion engines to electric propulsion,” said Robert Baillargeon, thrust area leader for electronics, controls, software process, methods, and tools at GM Research and Development. “But to take it further, when we talk about mechanical systems, we’re talking about electronic systems replacing them for both steering and motion of wheels. This means that everything from how we built them, how we service them, and how we sell them is going to have to change.”
In a roundtable keynote speech, as part of the IBM Centre for Advanced Studies Conference (CASCON) in Toronto, Baillargeon outlined today’s major challenges in auto industry that will hopefully be addressed with GM’s change in design principles. Some of the challenges include: the need for low cost, renewable energy, the environment, driver safety, road congestion and affordability.
“GM wants to extend the concept of a standalone system of vehicles that drive down the road, to a vehicle that is interconnected both with other vehicles around it as well as the roads and terrain,” Baillargeon said.
And for him, this is where the development of new innovations through software and electronic design will come into play.
“For the longest time, whenever we talked about a platform, we talked about the physical chassis of the car, now we’re talking about a software platform that extends well beyond that,” Baillargeon said. “It enables a lot of diversity in terms of applications and the locations where we put them within the vehicle.”
Ideally, through emerging technologies, he said, GM can teach vehicles how to feel and sense what’s happening on the road. But, Baillargeon said, to be truly effective, systems needs to be developed that work in collaboration with other cars on the road. In a computer simulated video, Baillargeon showed the audience at CASCON a series of cars using sensors to communicate with each other about situations and conditions on the road.
“Wouldn’t it be nice if other cars driving ahead of you can communicate back to you, automatically through the use of sensors, that there’s ice up ahead,” Baillargeon said. “So, you’d get a heads-up display communicating that to you well in advance and allows you to be prepared ahead of time. The same thing could be used if there were a collision up ahead or lots of congestion.”
Typical systems in the future, he said, will no longer have a single feature in the vehicle performing a function, but rather four or five electronic control units working together on a network to perform a single function.
Another example of potential innovations to a car’s navigation system, according to Baillargeon, could be a system that tells the driver how much money their trip on the 407 Express Toll Route, an electronic toll highway in the Greater Toronto Area, might be expected to cost.
But the ultimate objective, he said, is to get vehicles to drive themselves. With a combination of GPS, digital maps, 360 degree sensing, and vehicle-to-vehicle communication, GM hopes this can become possible. And with further software and electronic breakthroughs, Baillargeon said even more could be possible.
“The next step of ‘take me to where I want to go,’ is in developing algorithms and electronic controls and actuators to get where your car can essentially become a chauffeur,” Baillargeon said.
One of the software development standards Baillargeon cited was the Unified Modeling Language (UML), a general-purpose language that includes a graphical notation to create abstract models of a system.