We autonomate instead of automate


BLOG 6: Who is in control?

Autonomation or Jidoka [1] is still being suggested by Toyota as the second pillar of its production system focusing on highlighting problems and creating built-in-quality. Every situation that creates a deviation from an expected outcome should be stopped and examined in the line. Since Toyota holds their “no defects philosophy” high, visualizing deviations becomes a huge focus. This implies that several sensors (of all sorts using color, temperature, sound, …) are used to inform the operator of abnormalities. Jidoka also automates those tasks that are dirty, dangerous or awkward and or cause fatigue in the course of the shift or cause repetitive stress injuries over months and years. The latter view may cause some confusion as if the real intention is to eliminate the work from operators. On the other hand, it appears that Jidoka is about error alone. It isn’t. Blog 7 A car navigation automates the need for a map reader. Nowadays my car navigation suggests where I have to drive to and most of the time I just follow what it says, because I’m lacking information. At this point I feel that the machine is more in charge than I am. If my car could provide me with information like real time road blockages, change the route to a more enjoyable environment. If I could get real time information about places to see and visit while driving, I might not just follow the navigation but alter the course based on this new informat ion. Although the navigation tries to find the shortest of quickest route and is able to make many decisions on this using current traffic information, I can always overrule the route by pulling my ‘virtual’ Andon cord.  Automation makes the operator useless BUT autonomation informs the operator so better decisions can be made. It should be the same in a production environment. Autonomation is about providing the operator information to make other decisions than the machine is capable of. If the machine is making an error or if the number of parts are being exceeded, the operator wants to be able to shut it down before more (bad) parts are created because the machine is not capable of all those insights and extra knowledge the operator has.  Autonomation makes things visible. It is designing for control, so the machine is almost forcing the operator to follow, but the other way around. Autonomation in this sense is definitely the next hype in production design. Arkite believes that autonomation is the key driver in any production sector where people are involved, creating the need to intervene and interact with the environment and also navigate the operator to get the job done. Don’t automate, autonomate. People matter. #lean #jidoka #autonomation #augmented reality [1] (http://www.toyotaglobal.com/company/vision_philosophy/toyota_production_system/jidoka.html)

BLOG 5: Next generation work instructions

Increasing the user friendliness for unexperienced operators is one of the most thrilling new features we have been doing at Arkite. A couple of years ago I trained production companies in ‘how to make instructions’. One of the standard examples I used was how to mount a Lego car. What I have typically seen is that instructions were made with lots of text, arrows, pictures and often too many steps combined. It resulted often in a too complex instruction. Everybody recognizes the beep from a microwave, a traffic light, an ambulance siren, a blinking smoke detector, even an Apple starting up (sound used for Wally in the Dblog-5-foto-1isney film), the barcode scanner used at the register in a supermarket, beeping when a truck drives backwards, Nokia’s default ringtone, chalk on a blackboard, thunder and lightning, applause, from blinking to static indication of the level of the elevator, the red light indicator on any electric device that indicates on/off, the green/red indication light for availability of a parking lot, the beeping sound and use of color markers increasing in frequency of a car sensor gives you an idea how far you are from an object. Each sensor and its intended function gives meaning to the user. What do you think of the difference between above daily social instruction as compared to the production instructions below? (picture from source) What do you think the symbols below mean? blog-5-foto-2 For sure you know what these symbols mean so why not completely rely on these? Work instructions can be replaced with dynamic symbols in the production as soon as they are needed, at the exact position, in the right order, leading to clear undoubtable action. Arkite makes people more experienced and creative as they reach more experienced level. By systematically let HIM (Human Interface Mate) question the way of working in a gamified and production-safe way, a better solution to work can be found and thus utilizing the creative power of people. Don’t automate, autonomate. People matter. #lean #instruction #autonomation

BLOG 4: How creativity can excel in production!

blog-4-foto-1In todays standardized production we tend to distinguish a master, experienced and inexperienced operator. In many cases the master helped to define the standardized way of working. But where is the creativity? I have noticed that as people tend to work a long time for a company, creativity decreases. New inexperienced people, who have many interesting ideas, are typically trying to fit in and thus killing their creativity. The interesting ideas don’t get the chance they are supposed to receive. I rather believe that the inability of an inexperienced operator contains an unknown resource for quality and creativity. In the knowledge that an experienced operator blindly proceeds in performing the actions, it seems that the standards have been thought through enough. Before you can reach a standard work process a lot of variant ways to produce have been tried out. All this knowledge is lost in defining the standardized work. This procedure almost automatically kills the creativity the operators have. On the other hand, there are always operators that tend to do something else than the standard prescribes. How many people I have met stating that operators don't want to follow the standard! Even the operators that tend to do something else might have found a better way of working! That people tend to be creative and produce in a different manner should blog-4-foto-2not be wrong! I would embrace this creativity to find new ways of improved standardization. I expect that this is also an ongoing source of renewable ideas. So bottom line, why are we all not using the opportunity for each inexperienced operator to rethink production processes? And why are we not stimulating the experienced operators applying more creativity? At Arkite we believe in supporting creativity once people tend to become more experienced. It is by systematically letting the Arkite HIM (Human Interface Mate) question the way of working in a gamified and production-safe way, a better process can be found. The creative power of people is then used in it most optimal way. People matter. Don’t automate, autonomate.

BLOG 3: How to help others not to fall

People make mistakes, it is as simple as that. And we should be proud of making mistakes because nature took 3.5 billion years to perfect it! In a productblog-3-fotoion environment mistakes are not corrected themselves, we use poka yoke’s. A poka yoke is any mechanism in a process that helps an operator avoid (yokeru) mistakes (poka) without costing time to use. Notice that the performance of the action together with the begin conditions can result in a mistake (not desired possible states of the system or failure mode as common used). Let me give a practical example. If a person is running fast in a corridor, but the tiles are wet, the person might fall. So running (performance of action) and slippery floor (the begin condition) results in an error, in this case falling. The begin condition of the wet floor is not always true. Suppose in this case it is wet due to cleaning. Each begin condition can be tracked back to an action performed earlier. Something must have caused the wet floor. The known poka yoke for this would be to put up a sign saying slippery floor. In fact the possible state of falling of a person could be avoided by adding information in the past eg. putting up a sign after cleaning. The cleaning water could also be altered to better show the presence eg. with color. So adding extra information during a (former) process can avoid the bad possible states later. A better designed poka yoke should make sure that the end situation (clean floor) is without any retribution or issues afterwards (still wet and clean floor). So the poka yoke should report instantly if something is wrong and not wait too long for the mistake to appear. What if you could manage to add real time uniform information step by step making sure that all steps are done? Arkite’s Human Interface Mate assists the operator real time as if the operator was in a simulation helping the operator to navigate to the next step. As if a virtual angel would look over the shoulder and when the operator makes a mistake, the operator is informed about it so action can be taken to not let the mistake flow to the next step! Don’ automate, autonomate. People matter! #poka_yoke #lean #autonomation #jidoka

BLOG 2: Navigating through the field of errors

In the nineties I remember when I was travelling, navigation didn’t existed. We used a book with maps. It was not always easy when several directions where presentedblog-2-foto to know which road you actually had to take because a major city was sometimes announced and then some smaller ones. A decision had to be taken quickly, unless you knew the map by head. In a production environment a navigation system could be called a poka yoke. A poka yoke is any mechanism in a process that helps an operator avoid (yokeru) mistakes (poka) witho ut wasting time. But as with the paper version of navigation and trying to know things by head, in production it happens that we choose the wrong way ending in the wrong destination: an error. An error is something like all possible states (or failure modes) of a system that can be reached by action. A good poka yoke like your car’s navigation system will guide you towards good products. What would happen if your navigation system stopped? Would you still be able to find your destination? If a system takes over all knowledge you also will be lost! Good that I didn’t throw away all my paper maps 🙂 ! So a navigation system should at the one hand guide you towards the destination but also try to help you remember how to get there in a gamified way. To navigate through the process Arkite has come up with a projected navigation system for each step of the process, minimizing the bad possible states but mindful for stimulating the alertness, competence and creativity! Making operators more master chefs than industrialized robots. Don’ automate, autonomate. People matter!

BLOG 1: Eliminating error striving towards 100%

Eliminating error and safety is the main function of a poka yoke. However, for those who have tried to make a poka yoke, always something else seems possible that doesn't make the poka yoke 100% perfect or mistake proof. Basically, when no system is present to correct or detect and inform the error, the so called error is produced through the actions of an operator
  • by mistakenly altering the sequence or
  • while mounting using a "wrong" direction of an object,
  • wrong position of a tool,
  • by not respecting the process parameters of a tool or machine,
  • pushing buttons in the one exception the programmer didn't consider yet,
  • standing in the path of a moving object resulting in an incident or accident.
Hence it is possible to perform an non-compliant action resulting in a mistake or error.blog-1-foto A (technical) poka yoke is any system that results in the correction of the action while performing it OR a system that detects an error in the making and gives a sign to the person or machine performing the action.  The best poka yoke doesn’t add any extra time for the process to complete but are parallel operating.  We should be careful when applying the concept of poka yoke eg. sorting, filtering out bad parts doesn’t prevent bad parts so it can’t be called poka yoke. If the machine produces bad parts the device that solves it, is not a poka yoke. A human interaction is actually needed to talk about poka yoke. If a device checks the part after it has been made, I wouldn’t call this a poka yoke. What we learn from this is that an error can only be made through human action and the right conditions. If the system isn't used or in operation no real issue is at hand. If you can eliminate the action no error can be produced. The action requires a change of "energy". Changing the type of energy needed to create the error sometimes helps to prevent it. In a later blog I will come back to this. So if we can’t eliminate the action, parallel checking and informing or stopping is the only alternative. Most poka yoke’s focus on the end of operation or by checking initial conditions are met, but also the action itself can be checked, reaching closer to a 100% poka yoke. Arkite developed a poka yoke to check the process. One of the key features is without adding extra operator handlings (confirmations) and thus time, Arkite’s product – Human Interface Mate- informs the operator in real time of any potential human process deviations.  What’s new is that it can actually check if an operation is performed like stirring a spoon, making a torque action, cleaning a surface. In an example an operator has to drill holes but sometimes chooses one of the wrong holes. Using the Human Interface Mate the process will be navigated and when any potential deviation would occur the operator would visually be notified to prevent any possible error. Typical alternative solutions will check after the drilling but at this point you’ll have a scrap part. Arkite brings through HIM the logical poka yoke. People matter! Don’t automate, autonomate.