In the hit 1970s television series, The Six Million Dollar Man, technology of the future would remake Steve Austin into a “stronger, faster, better” human being. But that was science fiction. Today, humanoid robots promise to do human work as well or better than people can — without the need to rest or eat, the risk of injury, or the fear of potential hazards. Also, for a humanoid to be successful, it needs to cost far less than The Six Million Dollar Man.
The engineering challenges in designing a truly capable humanoid robot at an affordable cost are considerable. Not even 10 years ago, viral videos of humanoids falling down at the DARPA Robotics Challenge provided the general public with an amusing look at the difficulties involved. Just to stay on their feet — let alone perform complex tasks — humanoids require sensing, processing and motion systems far beyond the capabilities of most industrial robots. But these challenges are rapidly being solved thanks to exponential progress in artificial intelligence, CPU speeds, integrated circuit miniaturization and other technologies.
Every technology used in a robot, however, is in the service of motion. The humanoid is useful only to the degree that it can move with the balance, precision and dexterity of a human — or even “stronger, faster, better” than a human.
The biggest opportunity for humanoid robots is in performing tasks on a human scale in environments built for humans. That could mean anything from moving boxes in a warehouse to replacing human labor in dangerous environments, helping with elder care, or even doing routine household chores. A growing labor shortage is incentivizing the development of these human-scale robots, with Goldman Sachs Research forecasting a $6+ billion market achievable within just 10 to 15 years.
To realize the opportunity, engineering teams need to address the two most important motion requirements for a successful humanoid robot:
1. Minimize power consumption for prolonged battery life.
2. Maximize torque density so that robots can carry their own weight while also manipulating significant external dynamic loads.
These two requirements are closely related and must be solved simultaneously in the same motor. That’s why Kollmorgen believes motors and actuators should be designed to be robot-ready, rather than borrowing from the motion requirements of drones or other non-robotic applications.
Understanding the Challenges of Torque, Speed, Weight and Efficiency
The maximum recommended weight limit for human workers under the National Institute for Occupational Safety and Health’s Lifting Equation is 51 lb (23 kg). UPS requires labeling and special procedures when shipping packages over 70 lb (31.5 kg). Implementation of the Manual Handling Directive in the European Union varies by member state, but several countries specify a 25 kg limit for men and 15 kg limit for women.
These weight limits provide a general guideline for the dynamic loads that most humanoid robots should be expected to carry — but robots also need to support and move their own weight, which, at a truly human scale, could be two to three times the weight of the load.
Fixed, collaborative robots working in manufacturing environments have been achieving these and greater load-handling capabilities for years. However, consider that even the most sophisticated and flexible industrial robots typically provide 6–7 degrees of freedom. Humanoid robots, in contrast, might have 30–40 or even more axes of motion. These axes give the humanoid the freedom to move about, manipulate its environment and perform sophisticated, human-like tasks.
Each of these axes adds weight and bulk to the robot while consuming energy. Advancements in artificial intelligence, visual systems, kinesthetic sensors and data processing speeds are all important. But nothing could be more crucial to engineering a new generation of successful humanoid robots than maximizing torque density, minimizing size and weight, and reducing the power consumption in each robotic joint.
Understanding Servo Motor Requirements
Motor size, weight and torque are key specifications for humanoid robotic joints. Calculating the optimum load-torque-speed is relatively straightforward for an industrial collaborative robot, but humanoids present a different set of challenges.
Humanoid joints do not operate within the relatively narrow speed range of a cobot. Navigating and performing work within an unpredictable environment, each humanoid joint must be capable of very quick bi-directional accelerations — from zero to high speeds and back again — in a continuous, dynamic dance of balance, precision and power.
Given these requirements, conventional measures of motor performance — such as continuous torque and speed ratings — are of limited use when selecting motors. Instead, relevant benchmarks should be based on the motor constant, or Km, calculated by dividing the torque constant (Kt) by the square root of the line-to-line resistance of the motor windings (Km = Kt / sqrt Rm). Km is essentially a measure of motor efficiency when comparing motors of similar size.
A highly efficient motor can operate with minimal thermal rise, helping ensure reliable performance for the motor and heat-sensitive components such as lubricants and electronics within the tight confines of a robotic joint. Calculating Km per gram of motor weight also provides useful data for choosing the lightest motors that will deliver the performance you need.
An additional way to optimize torque in a lightweight, compact robotic joint is to use motors that take advantage of the D2L rule, which essentially states that doubling the moment arm length results in a fourfold increase in torque without affecting the motor’s axial length. In other words, D2L allows you to build a more powerful joint with a simple increase in motor diameter while keeping the all-important joint width at a minimum.
Choosing the Right, Robot-Ready Motors
Servo motors designed specifically for the size, weight and performance requirements of robots can help engineering teams design and build a more capable, more marketable humanoid robot. The prime example of a robot-ready motor is the TBM2G series of frameless servo motors from Kollmorgen.
TBM2G motors are available in seven frame sizes, each of which can be optimized around three different stack lengths. This is a huge advantage over competing motors that are typically available in only three to five frame sizes. With so many options, TBM2G motors can be specified to achieve the ideal balance of size/weight and torque for each joint in each robot, depending on the robot’s intended use.
When adding up the many robotic joints that go into a humanoid, the weight and size savings with each joint can make for a substantially lighter robot that requires far less energy to support and move its own weight. Compact, lightweight, right-sized TBM2G motors are ideal for achieving this engineering goal.
These innovative frameless motors also incorporate advanced materials and multiple winding options to help engineers achieve optimized mechatronic solutions across the speed and torque demands of humanoid robotic joints. TBM2G frameless servo motors offer:
- Rapid acceleration with consistent torque across the range of high-performance robotic joint requirements.
- Greater continuous torque in smaller package sizes, with form factors that take advantage of the D2L rule.
- Reliable response and precision for the highly dynamic motion required in humanoid arms and legs.
- Superior energy efficiency across the operating range in mobile applications powered at 48 VDC and below.
- Low thermal rise to extend the life of lubricants, electronics and other robotic joint components.
- A large inner-diameter thru-bore to accommodate encoders, cables, shafts, tools and more.
The TBM2G series joins a complete selection of frameless motor solutions from Kollmorgen. Learn more about Kollmorgen’s expertise for robotics and humanoids then contact us to consult with a Kollmorgen robotics specialist. Together, we can help bring the humanoid robots of the future to life.
