Robots as a tool to help humans be more productive was a common theme of the 2017 ROBOBusiness Conference held in Santa Clara. In the above interview, Steve Crowe, Managing Editor for Robotic Trends, discusses the panel he moderated, which featured leaders from DHL, GE, Intel and NASA, and how machine and deep learning are helping robots to act autonomously by seeing and reacting to their respective environments. Reinforcing Crowe’s comment was one of the speakers from one of the breakout sessions who suggested that the addition of natural language interfaces, combined with machine learning, allow quality inspection engineers to “train” robots by talking to it to help identify good and bad parts.

These co-robots, or cobots, simply become another tool for a craftsperson to do her job more efficiently. As echoed by speakers in the breakout sessions, automation isn’t about jobs, as much as it is about tasks. The industrial use-case for robots eliminates drudgery and helps workers perform higher-level tasks, while the robots perform repetitive tasks. One speaker compared the evolution in automation to that of going from a regular to an electric screwdriver. As an example, Crowe mentions a warehouse use-case where productivity increased by 40% with the same number of people.

Still there are cultural and human-machine interaction issues that manufacturers and service providers have been dealing with and the results may inform other forms of automation, such as autonomous vehicles in public spaces. At first, there can be trepidation by the human workforce to having autonomous bots whizzing about, but after a while the incumbent workforce becomes comfortable with their new co-workers to the point of even naming them. Sometimes the human workers become too comfortable, as an example was described whereby impromptu floor meetings would block the aisles, causing the autonomous robots to pause, waiting for the end of the meeting.

One of the big advantages to the autonomous mobile robot (AMR) approach over traditional methods that use guided tracks (e.g. via tape or wires, autonomous guided vehicles, AGV) is it gives the manufacturing plant or warehouse flexibility in the routes it sets up for its robotic workers. That is, with an autonomous mobile robot, a new route map can be loaded into its brains, as needs change with no need to change a wired infrastructure.

Similarly, one speaker, suggested that the market will evolve to encompass three classes of autonomous mobile robots, which he described as

• “motorcycle like” for short, narrow paths
• “milk-run”, which would be medium distances
• “long-haul”, which would across the factory floor

The above classes could parallel what might be seen as autonomy moves to the public realm and into transportation, as their could be right-size solutions addressing specific needs such as neighborhood travel, crossing the city and regional journeys. In a sense then, the factories and warehouses being served by the robots are the testing grounds for what will be a proliferation of autonomous robots in the public space in the coming decades.

Author’s note:

After seeing Stanford’s Jackrabbot outfitted with its nicely pressed tie and its ability to sense and interact with its surrounding environment combined with artificial intelligence, one has to wonder how long before it could be doing the interviews, as depicted in this edited photo.