Marrying AI and manual labor for more effective assembly lines
Digitization, automation, and Industry 4.0 are driving far-reaching changes in the manufacturing industry. But most of the mass production in industries such as automotive is still done by manual labor.
When it comes to manufacturing, there is quite naturally much talk about Industry 4.0. As you know, the industrial internet of things (IIoT), along with automation, artificial intelligence (AI), machine learning, augmented and virtual reality (AR and VR), are frequent buzzwords.
And with good reason. These are potent tools that are already transforming manufacturing, and this development shows no signs whatsoever of slowing down. As one of the persons at Axis in charge of driving Industry 4.0 solutions, this area is obviously close to my heart. And believe me, many exciting innovations are going on.
Quite recently, however, I was reminded how easy it is to start taking things for granted based on your work and frame of reference. So I was pretty surprised to learn that some 72 percent of assembly in mass production is still carried out by manual labor.
I came across this number when talking to a new partner. They are called Drishti and aim to improve productivity, quality, and training in various industries where manual assembly is still frequent and where lean production is generally the norm. They have a solid position within the automotive industry but also in medical devices and electronics and other long-tail discrete manufacturing industries. Drishti is also a valued member of the Axis Application Development Partner (ADP) Program.
Connecting manual labor and AI
When talking about Industry 4.0, we mainly focus on connecting smart components, letting the machines communicate to reap the benefits. In this perspective, Drishti offers a kind of hybrid solution. They use videos to connect the manual labor at the assembly line with AI. This way, they gather masses of data in the process, which is analyzed and form the foundation for future improvements.
It is a huge step forward since most mass production units still use the traditional way of measuring productivity by using a stopwatch and then writing down the results. However, this is a very limiting model because you miss out on valuable data that can be used to improve the operation.
On the other hand, Drishti places cameras above each workstation, which quickly and continuously measures the process. The analytics on top of the cameras and AI translate the video streams into actionable data. You will, for example, get information on how long each moment takes and if there are any anomalies, bottlenecks, or repetitive work moments that add time or even put quality or safety at risk.
You can measure the time spent at each workstation and if all the actions in a process step have been performed. If a step is missed, the system will raise an alarm.
Easier to find the root cause
Basically, the Drishti setup is all about data. Simple, accessible data. If you should detect a faulty product or get a customer claim, it is easier to get to the root cause. For example, I watched a video where a worker tightened the same screw twice – that is, one time too many. Maybe that doesn’t sound so serious, but ultimately, the torque used to secure this screw might have doubled. It may cause errors later in the process or at the usage stage.
If all the steps were performed correctly, but then you have something out of the ordinary, such as the problem mentioned above, the analytics software will classify it as a production error. And you can go back and check what might have caused the error.
In the end, it all boils down to providing business value for the customers. Drishti specifies four pillars of value, claiming that their setup offers deep visibility into production, reduces the number of mistakes, helps associates add more value, and speeds up production improvements. And, as always, the proof is in the pudding. For example, one customer improved throughput by 21 percent, another reduced their product line labor cost by 34 percent, and a third reduced their defect rate by an impressive 48 percent.
The operators become ambassadors
I’m not about to pitch the Axis portfolio to you here and now, but let’s say that our cameras with edge capabilities are perfect sensors in Drishti’s applications, and as such we are Drishti’s exclusive camera provider. We will talk more about specific solutions in an upcoming article, backing it up with more customer experiences as well as application and business insights.
But surely, the operators can’t like all this, feeling watched over and controlled?
No, as a matter of fact, it turns out that this is not the case. On the contrary, according to Drishti’s surveys, the operators often become ambassadors for the model, which is probably not surprising when you think about it. It brings an element of objectivity and is a way for them to prove their worth in a world that is getting increasingly digital and automated. The collected data can also form a solid foundation for training and future competence upskilling, benefitting both the operators and the company.
But of course, using Drishti’s services, it is crucial to guarantee GDPR compliance and operator privacy in the video streams. So, it is a primary customer concern. Here, Drishti got help from a partner – yes, you guessed it – where the installed cameras carry privacy-masking software. This solution operates in real-time and without affecting the quality or functionality of the video footage.
Massive amount of data
As you can imagine, Drishti’s services amass vast amounts of bytes. All these hundreds of thousands of hours of video and data need to be streamed, analyzed, and stored. In the upcoming article, we will discuss how Axis compression technology keeps these enormous data volumes manageable.
I’m looking forward to describing in more detail how these solutions work and how Axis solutions help make Drishti’s services improve productivity and quality in manual assembly production. So please keep your eyes open because I’m sure that it will be interesting. Especially if you are working with lean processes in the manufacturing industry and manual assembly is a vital part of your operation.