KUKA Robotics

Automation is nothing new - it’s been used in manufacturing for decades. But with every new technology on the factory floor we are reminded of how many different tasks can easily be done by people, but it takes creative programming and other efforts (also done by humans) – for robots or machines to achieve the same tasks, even when they look simple at first sight.

Don’t get me wrong – the automation of tasks is needed. Not only to compensate for worker shortage but also to protect workers from injuries due to dangerous or unergonomic tasks and to enable manufacturers to leverage human intelligence for more critical operations than doing simple and repetitive operations.

Looking at the integration of a mobile robot fleet in a production environment, mobile robots are often used to automate transportation tasks in which workers push carts around all day or drive forklift trucks or load/unload different various types of machines.  So what are some of the steps to automate these kinds of tasks?

First, all manual steps that are done today need to be identified and converted to a step-by-step process. How does the worker know that it’s time to refill a station with certain parts and which parts are needed? He (or his co-workers) might just use their eyes to look at the stations to see what’s needed. How does one know where to find the parts? Maybe they’re just always at the same place and if not, under normal conditions workers can easily see that this is the wrong part and look for it at another place. All these different steps need to be identified and as with every automation, the eyes, ears, voices and brains of us humans that are used for the process needs to be translated into steps.

Every step (and the tasks necessary) for the operation has to be predictable, every part needs to have a known place and everything has to be identifiable so that a machine or robot know exactly what it is, where it is and it has to be predefined what has to happen in case of unexpected events. It’s also necessary to coordinate the mobile robot task with the rest of the manufacturing process and connect all parts via software and controls. Taking the time to carefully analyze the task does more than setting things up for automation - it will also help identify pain points; thus improving and optimizing the process itself.

“Taking the time to carefully analyze the task does more than setting things up for automation - it will also help identify pain points; thus improving and optimizing the process itself.”

Another important consideration is safety. Again, the question to begin with: How does the worker currently handle the parts?  Do they simply use sight to find parts and coordinate the motion of his body, arm, hand and fingers to execute a task?  What role does experience, common sense, work instructions and consideration of his own and other people’s safety play when handling the part?

Robots and other automation are used for handling parts for a long time, so this step is very well known. Using the weight, shape and size of the part are key pieces of information needed to choose the appropriate mobile robot for a specific application. The same information together with other parameters like stiffness or fragility of the part all go into choosing possible solutions for picking and placing the parts on and off the mobile robot. There are plenty of different solutions possible, e.g., worker, forklifts, cranes, gripper attached to a robot or a conveyor system handling parts onto a conveyor mounted on the mobile robot.

In discussing a mobile task, the driving aspect has to be looked into very closely as well, beginning with analyzing the starting point: how does the worker find his way from one location to another?  How does he decide where to go? What if the location he went to is occupied? What if there are areas with bumps? All these things must be looked at “through the mobile's eyes”. Mobile robots come with different types of navigation, but there is a clear trend toward feature-based autonomous navigation based on lidar data which makes it more flexible to integrate. These mobile robots receive their tasks through a fleet management software and react to obstacles that hinder their way and stop immediately if someone comes too close. The software is also calculating the routes and coordinating the whole mobile robot fleet, deciding what robot is best to do a certain task and where every single mobile robot has to go and when. And regarding the floor – depending on the specific mobile robot type some fixes might be needed or areas being blocked for the robot to drive into.

The automation of human tasks can range from very simple to very complex. Keeping in mind that humans have capabilities that are easy to overlook because we are humans ourselves helps with identifying all necessary steps of a process. The better the analysis is done in the very beginning of a project, the smoother the deployment of the mobile robots will become at the end. Doing it right from the beginning is faster and easier than fixing issues that come up later.