Speedy material handling with minimum footprint

If you are in intra-logistics you should know about our HKM robot, watch our webinar before you continue reading: https://youtu.be/5jZYhngO_-I

If you are working in any other area, you should also listen in if you think robotic handling is too slow, costly or bulky. Also if you are simply interested in speedy long-reach pick-and-place from an engineering point of view, you will benefit from the webinar by considering the fundamentals of motions for moving materials. Let me explain...

The Hybrid-Kinematic Manipulator (HKM) principle/IP is a result from some of us at Cognibotics having 'second thoughts' after earlier experiences from working at ABB Robotics in Sweden (a pioneering company with brilliant engineers that are limited by a global organization). For pick-and-place within easy reach of a human arm, there are many robots to use, including Delta and Scara arms when speed is key. When longer reach is needed while avoiding to take the step into mobility, you need a robot with longer reach.

But what about speed and foot-print (floor and carbon)? Think about the following (when watching the HKM in the webinar):

1) First ignoring carbon footprint (energy consumption) and speed (over entire workspace), some fundamental aspects are already explained in "Big Robot Reach, Small Robot Footprint" meaning that rotary joints are to be used, with four axes as a basis. But, instead of the Scara on rails we aim for another type of arm that scales better for longer reach, such that we normally can do without rails, thereby reducing the floor footprint, and making it possible to stack several arms close to each other.

2) Reducing carbon footprint of motion means that almost all energy should be used for moving the payload, not burning energy on moving an arm with big-mass links, or holding that heavy arm up against gravity. Hence, all links should be made from light-weight carbon fiber, and all main-axes (1-3) motors and gears should be located at the base to minimize inertia, just like the HKM is designed.

3) Handling in general often requires some special end-effector that might be heavier than the grasped objects. On a normal articulated arm (be it a standard 6DOF or a 4DOF palletizer) this means additional static load has to be taken care of by motor torques. A perfect robot arm for handling, like the HKM, has some location along the arm (the elbow in the HKM case) where balancing of static load (including the outer arm in the HKM case) can be arranged mechanically. The inner arm is statically arranged such that no gravity torques are needed (just like for Scara arms, which however cannot have a balanced vertical motion).

Thereby, for the total vertical component of HKM motions, the carbon footprint is reduced even further, while also cost/weight can be reduced for that drive-train of vertical motion at the robot base.

4) Whereas the location of motors and transmissions near the base enables speedy motion from a mass and inertia point of view, it also means (since all arm elements are located after the speed reducer) that high forces and torques (both bending and torsional) are to be transmitted along the carbon-fiber arm structure. Even if carbon fiber is very stiff, in particular compared to weight, the long robot reach means long links, which implies a low stiffness and thereby quite resonant physical dynamics.

The effective motion dynamics for the user is, however, nice and stable due to active damping via feedback from an accelerometer at the wrist, accomplished by the algorithms in software. Thus, special algorithms enable the high-performance control of the light-weight structure, which has been mechanically designed such that cross-resonances are avoided (other robot arms will [by laws of physics, no matter the control] experience inaccurate motions if the arm is made too light and slim).

5) Finally, the configuration of HKM arm segments is such that the best combination of reach and speed is achieved. Considering the previous items this is an optimization that is done based on the differential-algebraic equations, which Cognibotics has worked out in detail such that a Digital Twin (also for future upcoming models) will have emulated controlled dynamics that are very close to the physical system, and hence machine learning can be applied both offline and online. The control software also utilizes the real-time version of dynamics for tool-changing on the fly as needed to provide the ultimate speedy material handling.

These five steps describe our 'second thoughts' that we have combined with plenty of practical know-how ranging from safety and cabling in the robot to programming and interfaces on user level. Apparently, the HKM is the perfect robot for mobile material handling too, but that is a later story. First, enjoy the webinar.

Klas & Team Cognibotics