From Barcodes to AI: Machine-to-Machine Solutions

Machine to machine solutions have been a game changer when it comes to modernizing commerce. Let’s explore what that looks like.

A History

Once upon a time, in lands near and far, small business owners and warehouse workers used to record inventory on thin slices of trees manually. There were perks to this: one could practice cursive handwriting and it was an excuse to carry a fancy pen. But, it was a time-consuming and tedious process that often resulted in human errors. 

This system worked but a constant need for industries to cut costs, a need for increased security for military and railroad applications, as well as businesses, need to beat the competition led to innovations in automation.

UPC Barcodes

In the 1980’s gained widespread adoption as a way to manage inventory for grocery stores. This simple paved the way for automated inventory adjustments, increased productivity, and laid the groundwork for online shopping. Unique product identifiers (UPC) barcodes tell the story, what is this thing, and what information is tied to it. 

That information could be pricing, warehouse location, country of origin, and more. It’s a product line’s fingerprint. It’s unique to every product of the same style and type.

Using a barcode, a scanner, and a computer information can be transmitted from machine to machine reducing the possibility of human error. It could still be possible that a checker at the store failed to scan an item or that the wrong label was placed on the wrong item.


Originally invented around the same time but for a very different purpose was RFID. This takes the technology even further, using radio frequencies as the unique identifier instead of an alphanumeric code.  The beauty of RFID is that the scanner doesn’t need to physically see the code to scan it, just be within range. 

This technology was implemented for military and access control applications. It wasn’t until the electronic product code was invented and the technology became cheaper to implement that it started to be used for commercial applications. In early 2000, ‘passive’ RFID tags were invented and the technology was more affordable for factory, logistics, and supply chain automation.

Suddenly, a store could be alerted if someone was trying to leave without paying for the mascara in their cart. Today it can be used for managing receiving docs, picking and shipping, for logistics, and it even has applications in gathering sports data and updating cell phone apps.

Machine Vision Object Identification

Advancing the technology a bit further, what happens when it’s not practical to add a UPC barcode or electronic product code to every object? What if you need an accurate count on produce coming in the store and you need to identify whether a customer is purchasing a persimmon or a lemon? 

This is where artificial intelligence with deep learning comes in. Machine vision detects objects that it has been trained to look for, processes and learns from the information, and creates an action, such as updating a database, issuing a speeding ticket, opening a gate, signaling the start of an SMS campaign, or anything else you can think of. Using this technology you can have an AI powered warehouse, city, train, airport, or even city.

Object identification is the machine-to-machine solution being heavily utilized today in the parking and public safety industries. For example, if you are entering a secure parking lot and your vehicle is registered the gate will open and you’ll be allowed access. The same technology can be used for identifying inventory being received even before it can be scanned with RFID. If it’s known that a truck is carrying a certain order and it arrives at the gate, the receiving individuals (or robots) can be instructed where the products will go without any physical scanning happening. 

Challenges and Future Trends 

Limited labor forces and a reduction in consumers’ desires to interact with humans have increased the need for automation across industries. Some barriers to entry with this are hardware-driven. If the solution is RFID, a business must purchase handheld and fixed scanners as well as tags. If the desired solution is machine vision, the hardware required is cameras.

The future is trending toward fully automated warehouses using machine vision for inventory, quality control, and more.  For retail, unattended markets and self-service are becoming increasingly popular. Not having to wait in line because you paid on an app or ‘just walked out’ of an Amazon store is the designated future application of machine vision in retail. Big challenges to overcome with stores that have implemented this technology are consumer education and having the right variety of sensors in the right places to capture every purchase. 

Smart cities and communities are another current, trend that is ramping up to make big changes in the future. So many things can be automated: toll roads, site access, security systems, appointment settings, and even traffic tickets. The future is trending toward integrating these systems for one seamless experience. Consumers can get where they are going faster and safer while the industries involved become better at the job of serving consumers with more accurate, real-time, data.

In conclusion, as with every time in history there are still changes to come, but with deep learning neural networks the engine is learning how to perform better every day and with each new application it continues to improve.