There is probably no human endeavor more complex, involving more people and companies, or more expensive than an automotive vehicle program. Billions are invested in creating a new body and powertrain. Hundreds of millions are invested in even minor upgrades. The bulk of that is in the manufacturing plant and machinery—stampings, tooling, robots, fixtures. That is a lot of financial risk. With so much at stake, the culture among automotive manufacturing engineers is necessarily conservative. New pressures from customers and new opportunities in technology could change that.

A panel discussion at the SAE World Congress conducted in April explored a number of exciting new ways cars could be built to satisfy a world demanding choice even while driving costs down. The panel was hosted by Omron Automation (Hoffman Estates, IL), organized and moderated by Monika A. Minarcin, automotive marketing manager for the company.

Problem: Complexity on the Rise

The consensus of the panel is that automakers are facing growing manufacturing complexity. The vast array of worldwide emissions regulations require automakers to deliver different engines and powertrains depending on the country or region. Buyers are demanding choices in the number of doors, seats, and interiors. Electronics and software further complicate the picture. Not only are there entertainment systems and electrified powertrains, but there’s also the growth of automated driving assistance systems, or ADAS. A major luxury carmaker with a plant in the United States has a single vehicle model with over 400 trillion buildable combinations, according to Adrian Jennings, vice president of RTLS technology for Ubisense (Denver; London), a speaker on the panel.

Another complexity is product timing. The basic vehicle platform can easily last 15 years or more—a car with 200,000 miles on it is no longer a rarity. Yet consumers familiar with other complex products, like cell phones and laptops, are used to ever-shorter product lifecycles and upgrades.

New companies in the field are also changing perceptions. “It is common to hear today of Tesla owners saying they are getting a software upgrade overnight,” said Curtis Wilson, vice president of engineering and research for Omron Delta Tau Data systems, another panel member. In the near future it won’t just be software. “There may come an expectation in the very near future where someone may purchase a car with two rows [of seats] and then demand a third row when they have a child, rather than purchase a new car,” he said.

Physically changing the configuration of an existing car will be no simple task. It will be impossible if the industry sticks with its current manufacturing model, according to the consensus of the panel. What is needed is “a factory of the future.” What could that be?

From Fixed to Configurable Production Lines

“The factory of the future is any factory that breaks the paradigm of the fixed workspace,” Jennings said. The almost universal model of the fixed-workstation, moving production line is over a hundred years old, pioneered by Henry Ford, he said. “These fixed lines exist within a fixed space marked as ‘The Factory.’ With production moving at a set pace in a straight line, there is only so much you can do that is different [to improve it],” he said. Accommodating an expanded palate of choices will require new, possibly, radical thinking.

Conveyance is the key, according to Jennings.

“You have to break the paradigm of the single line of conveyance, where a car being built enters on one end and does not get off until you get to the other end,” he said.

How? While there are any number of ways that could happen, one example is with a production line built around automated guided vehicles, or AGVs.

AGVs, programmed to run autonomously, are the opposite of a single line processing at a fixed rate. Outfitted with some of the same ADAS technology that enables self-driving cars, they need not be confined to a fixed space or fixed routes.

“They can be programmed in real time,” he said. With some innate on-board intelligence, including what version of the car is being built—three rows or two, hatchback or sedan—the self-driving AGV could guide itself to the correct station to build the right car.

“AGVs can also go in straight lines at a constant speed—which means they can be integrated into current assembly processes [at low risk],” said Jennings. This kind of line, reconfigurable on the fly and able to accommodate the familiar past as well as the future, could also accommodate product upgrades. “A car needing an upgrade to a third row seat would simply skip the other stations and go directly to the one that it needs,” Jennings said.

“We should also think about changing the structure of robots themselves,” said Ana Djuric, a professor at Wayne State University (Detroit) and a member of the panel. “The one-armed, six-degrees-of-freedom robot is very rigid. We can easily advance into [a factory] with distributed communications, with modular, reconfigurable machines, AGVs, and even drones,” she said.

Information, Culture and Tribes

While self-guiding, autonomous technology is exciting to talk about, it will require a higher level of integration. That is where the existing culture could be a barrier, according to Jennings.

“Within a plant there are these very strict stovepipe organizations, such as between controls and IT,” he said. The controls engineers work in the world of machine controllers, PLCs, and Ethernet IP dedicated to creating products. Information technology, or IT, deals with product lifecycle management tools, and networking with corporate-wide systems.

“There are knowledge gaps in these groups that create barriers in communication,” Djuric said. These different worlds—different tribes—have grown up in semi-isolation and getting them to talk to each other will be important.

“To me, the hardest thing is going to be the [information] glue that is going to put all of these things together,” said Wilson. As he noted, the notion of AGVs is a great idea, but he said the system might break down when an AGV has to interact with other machinery in the plant. “How would an AGV talk to a hoist that lifts up the part [that it is carrying]?” he asked rhetorically. There are communication difficulties outside the plant as well, between design engineers who use CAD and manufacturing engineers who use CAM. “Engineers create these beautiful sculpted surfaces, but I find that not one in 100 know enough about the dynamics of the machine tool cutting path traversing a curve, taking into account velocities or accelerations,” he said.

The consensus seemed that to advance to a future factory, current organizational thinking needs to advance as well. 

Moving beyond the standard six-degrees-of-freedom one-armed robot requires a global kinematic model an n-DOF reconfigurable machinery structure, according to Professor Ana Djuric of Wayne State.

Demise of Lean Thinking

Minarcin of Omron Automation, the panel moderator, also challenged one of the fundamental tenets of modern manufacturing—lean thinking. Most of the panel agreed with her premise. The industry must move beyond lean.

“The constructs of lean that developed our current system has been around since before we had personal computing,” she said. It was lean thinking that developed the linear production lines of today, with production time and efficiency per operation ranked as more important than anything else. “The problem I have with lean, Six Sigma and just-in-time today is that it is an incremental and often localized technique that does not allow for the innovation process required to address optimizing product complexity or consumer choice,” she said. “We need to look at the construct of the factory system.”

Optimizing in a linear fashion limits flexibility and consumer choice while not addressing the return on investment which is possible. It is at odds with the changing automotive market.

“Lean has to die,” she said.

The demise of lean also has to do with how best to use humans in a global marketplace faced with encroaching automation, according to Jeffrey Liaw, engineering manager for Martinrea (Vaughan, ON), a supplier of automotive body parts. Liaw is in a unique position where his body parts plant located in Ontario also has a factory making the same parts in Mexico. “If I lean out my plant and save, say, $5 million dollars annually, the plant in Mexico would do the same thing,” he said. “We in the first world cannot compete with low-cost country wages trying to achieve the lowest cost through traditional efficient lean techniques. It provides us no advantage.”

Operating from the premise that flexible manufacturing is a plus, the question he posed to himself is how to best employ creative humanity to their best potential—not just how best to lean out any particular process. “The questions we should ask ourselves are: what can we do to offer people more salary? How do we make people who work for us more creative and therefore more valuable?” he said.

Liaw also agrees that culture may be the biggest roadblock. “Traditional lean assumes one thinker and many followers,” he said.

“But a single perspective becomes a barrier, it becomes impossible to understand all of the problems in a plant or opportunities to improve. If you push solving problems down to the local workcell, they become smaller, and easier solve,” he said. What he did was essentially raise the expected skill level of each of his workers, with positive results.

A reconfigurable plant where complexity and choice are the most important features, where locally empowered workers are free to solve their own problems, will challenge many fundamentals of today’s thinking.

“Manufacturing constraints are the culprit,” he said. “The real issue is, are we giving enough leeway to the plant to get solutions.”