The Difference Between Autonomous Cars and Autonomous Mobile Robots

What you’ll learn:

• What are autonomous mobile robots (AMRs)?
• What are the differences that separate autonomous cars from mobile robots?

At this point in time, there aren’t really any autonomous cars or autonomous mobile robots (AMRs), at least for an extended operating time. They’re automated. To be truly autonomous, they need to be able to act on their own. 

Both cars and robots do have failsafe mechanisms such as a high-level controller or human intervener in the cloud who can take over and deal with sticky situations that exceed the capabilities of local artificial-intelligence (AI) algorithms.

With this clarification in place, I’ve wondered why cars can drive on roads “on their own” while mobile robots in industrial settings seem to be struggling with basic routing challenges. 

If you travel in San Francisco, you almost can’t help but have a Waymo vehicle without a driver pull up alongside you. On the other hand, mobile robots moving inventory throughout a facility can be stymied by something as simple as a spill. And if they’re carrying a load such as ice cream while waiting for the mess to be cleared, you may have a second spill on your hands.

AMR vs. AV Design Complexity

Designing an autonomous robot should be a much simpler task than building an autonomous car. Robots work in highly controlled environments that are mapped, beaconed, and regulated. The main controller manages every mobile entity in the facility. It can even optimize movement by advising one robot that another is using a particular lane and direct the robot to use an alternate route. 

Unpredictable elements like people are restricted to areas outside of robotic movement zones. In addition, if a person does violate a robot’s safety boundary, the robot shuts down to prevent injury. 

Now compare this to the challenge of driving on the road. When computers can barely win a game of chess against a chess master, how can they possibly handle the far more complicated game of driving? 

To a degree, automated cars can be constrained to a “fenced “area. Still, driving on the road with distracted human drivers, adverse weather conditions, and variations such road work, to name just a few of the challenges of driving, makes even automated driving extremely complicated.

So why is autonomous car technology so much more advanced than robotic technology?

Consolidation and Coordination of Autonomous Vehicles

Simply put, there’s more consolidated research on automating cars when compared with guiding robots. Overall, the automotive industry has fewer players and a huge support ecosystem coordinated to leverage knowledge into innovation. Since the major vendors sell millions of cars, there’s simply more money to invest in autonomous technology as a differentiating feature. Thus, the problems that can be solved for vehicles are deeper and more extensive. 

Cars also have the advantage of numbers as well as repetition. Many cars drive past the same areas, using the same roads. Collectively, they can gather a tremendous amount of data in a short time about an individual segment of road. This data accelerates how quickly AI can learn to accommodate circumstantial hazards. It also confirms the AI’s effectiveness to navigate them and capture exception cases that still need to be uncovered and addressed.

Though an increasing number of robots are sold every year, AMRs represent only a portion of that total. Also relevant is that individual robotic companies tend to sell comparably few robots. 

Put another way, a company that only produces 1,000s of robots has limited resources to develop AI for those robots. In addition, third parties developing autonomous technology for robots must make small sales to many companies to survive. This means they also have a higher burden to adapt their IP to integrate with a greater number of proprietary platforms.

In the case of automated mobile robots, individual innovation may be slowing the industry down. Because each robotic vendor is effectively reinventing the same wheel over and over again, the limited research dollars and technical resources available are being used to less than advantage.

Yes, there’s an economic advantage to controlling innovation. However, if the robotics industry consolidated their research, they could progress at substantially faster rates of innovation. To truly capitalize on the value of innovation, a company needs to achieve volume production. 

For a market with many companies creating specialized robots targeted at narrow applications, robotics companies may see greater benefit from sharing information and working together. They won’t have a technological advantage over each other, but that may not matter as each of them will be able to serve a wider market with greater capabilities for a lower investment cost. In the end, a coordinated approach may yield significantly higher gains than traditional business models.