Designing a human sized industrial robot
John Hayes posted on September 05, 2013 |

Manufacturing robots have historically been fast, dedicated machines for high-volume manufacturing. For lower-volume repetitive tasks, North American manufacturers have turned to outsourcing overseas.

Rethink Robotics was launched in 2008 to create a robot that would meet the needs of North American custom manufacturers. The design challenge was to make a robot that was:

  • Human size and human pace – not ultra high speed,
  • Less expensive – Target around US$22,000,
  • Safe to work around,
  • Easy to program.

I spoke with Michael Lewis, senior mechanical engineer about their latest robot, Baxter. 

In particular, he mentioned how the robot could mimic the wide range of motion of human joints. Lewis pointed out the very long arms with seven degrees of freedom from shoulders to elbows to wrist.

For safety, Lewis’ team gave Baxter's movements some "give" through a spring system on every gear output that senses when Baxter encounters an object, like a person. 

To design that spring mechanism, there is a sensor that measures the load on the spring through a combination of a magnet and a Hall effect sensor.  When a traditional industrial robot hits an object, it keeps going. Baxter retracts.  

To develop a commercial robot, Lewis pointed out their humble start in 2009 when they worked with gearboxes that they scavenged from other assemblies.  

Now Baxter is made from 550 unique parts. The full assembly combines a wide range of manufacturing processes including sheet metal, plastic injection molding, die casting, welding, fastening and gluing to name a few. 

Managing the complexity of these sorts of designs and processes is not trivial, but through extensive use of contract manufacturing and handing off fully featured CAD models, the design team has developed a successful product.

As of now, Baxter is in use in lightweight repetitive tasks, such as packing boxes, putting together plastic assembly kits and providing insight into grasping mechanism research.   

Recommended For You

Recommended For You