Universal Robotics develops flexible automation software solutions to react and adapt to their surroundings, and perform tasks that are costly, dangerous or difficult for humans to undertake. The U.S. Department of Labor reports that transportation and warehousing has the highest non-fatal occupational injury rate of all sectors. Overexertion injuries are the biggest cause of this injury incident rate. Universal targets automating logistics tasks for robots such as mixed depalletization, random bag picking, and random bin picking.
Inbound logistics specialists purchase, coordinate, and move materials, parts, and finished inventory from suppliers to plants, warehouses, distribution centers and stores. These workers unload trucks often piled high with boxes of various sizes, and depalletize partial or full pallets with mixed cases. Most retail businesses unload 8 to 12 boxes or cases per minute. Maximum weight of such boxes is typically 60 lbs (27 kg). Thus, an automated robotic solution must identify and move up to 12 random cases per minute weighing up to 60 pounds each.
However, doing this requires robotic machine vision guidance that senses 3D variability. True 3D has six degrees of freedom (DOF) – position (X,Y,Z) and rotational pose (Rx, Ry, Rz). For instance, identical boxes, cases or stock-keeping-units (SKUs) on different layers of a pallet have completely different X,Y,Z locations (width, height, depth). This is not the case with parts in an automated assembly line – a mechanical fixture holds them in place, or they can be found in the same place every time.
True 3D sensing measures variations in pose (Rx, Ry, Rz) as well. In contrast, a flat part on a conveyor only requires 2D vision. It can only rotate around one axis perpendicular to the belt. However, in a distribution center, a palletized group of random boxes may vary in orientation along one axis to three axes. A neat stack of boxes may only rotate perpendicular to the floor, but tilted boxes require 3D vision to describe orientation in all three axes. As an example, random truck unloading involves processing boxes/packages whose orientation shifts during transit. This variability in box position requires a system that can handle all three axes. Similarly, random bin picking has objects arbitrarily mixed in a container, varying in all three axes.
SPATIAL VISION – 3D Made Easy™
The Spatial Vision Robotics (SVR) product delivers true 3D object recognition and machine vision guidance. Universal engineers are experts at constructing solutions that combine Universal’s patented software with sensors (vision, laser, microwave, infrared, etc.) to generate accurate and reliable 3D position and pose (in six degrees of freedom).
Random objects and locations require a large field of view to cover a 3D region of interest within the workcell. Typical 2D or 2½ D machine vision is limited to a smaller area associated with a fixture. Corresponding to this, machine vision often requires micron levels of accuracy.
In contrast, the 3D accuracy is typically 2 – 5mm for moving cases and boxes. Pliable vacuum cups on the robot’s end of arm tool (EOAT) are designed for this level of flexibility while picking up objects varying from ounces to hundreds of pounds.
To meet this range of accuracy needs, SVR provides three levels of sensor suites with increasing levels of accuracy: Standard (± 5.0mm accuracy), Enhanced (± 2.5mm accuracy), and Premium (± 0.5mm accuracy).
AUTONOMY – Automated Robot Programming
3D sensing is only part of the solution. Moving random objects to and from various locations (i.e. pallet to conveyor) typically requires a dexterous robot. The scope of motion control changes from fixed point-to-point locations to regions of interest with dynamic target locations. These movements can exercise the full range of a robot’s degrees of freedom. If the robot is retrofitted into a workcell, and the task requires a wide range of random movement, a 6-axis robot ensures it can manipulate objects in tight quarters. If the robot arm can swing free, or the range of motion is more limited, a 5-axis robot is sufficient. If pick and place tends to occur from the top of the box and case, then a 4-axis robot works.
A robot conducting a high-speed task that requires a high degree of dexterity results in custom motion control, and additional programming. Universal Robotics Autonomy is automated robot programming software that significantly reduces the programming time required. It automates the process of creating robot jobs. Autonomy generates inverse kinematics and collision avoidance between robot arm(s), tools and other objects. It provides high-speed response up to 500 Hz with single or dual arm robots up to 17 axes.
NEOCORTEX – Software with an IQ™
Logistics tasks encounter random, unexpected situations. A pallet full of cases may arrive on the dock damaged, wet, leaning, or loose. A worker picks up a box, and the bottom falls out. Increased randomness requires intelligence that can both adapt to changes as well as react to unanticipated events.
Universal uses patented machine-learning technology, called Neocortex™, which utilizes memory and sensory data to control robot behavior. It recognizes complex objects and can identify them in random locations, increasing flexibility. Neocortex is able to move boxes it has never seen before, learning on the fly, in dynamic and changing environments.
Inbound logistics is in a state of flux: ramping for holidays, lowering inventory, overcoming shortages, and meeting on-time delivery commitments. Neocortex learning identifies changes in process or product over time. By applying intelligence to volumes of data, meaningful patterns emerge. Automated tasks can be completed even when trends shift.
Automating inbound logistics solutions requires 3D sensing, control, and intelligence. Universal Robotics and their team of partners provide flexible automation systems to move never-before-seen objects at industrial speeds.