The warehouse robots are coming.

The "dummy" artificial intelligence can make impressive things. But they need human assistance in every unexpected situation. 

The problem with artificial intelligence is that if artificial intelligence needs flexibility it requires machine-learning ability. But still, most artificial intelligence systems are operating in the closed areas, and they can be flexible without machine learning ability. 

The term dummy artificial intelligence means that the system cannot independently learn new things. There are standard models of how they must react. Those models are fixed. And they cannot be automatically changed. 

And in the cases of surprises, the system asks the human operator to decide for it. These kinds of systems are used in warehouses. The moving robots are like remote-control cars, and the control system is locating in the fixed systems that are in the computer centers. Those systems might use coded radios because that makes them safer. And the possibility for the error signals is minimized. That denies the misfunctions and dangerous situations. 

In that case, the system doesn't need the ability to learn new things. If some route where the robot is moving the merchandise is blocked the system supervisor can just deny using that route. In that case. The cargo robots need only the fixed models of how to move. If one or two possible routes are blocked. In this kind of system, the robots are "dummies". And the fixed computers that are in computer centers are controlling the movements of robots. 

In that case, the robots might only have the radio control system, and the camera that is delivering the data to control computer and system supervisors. If there are people in the hall, the system just stops the movement. And in the case that there is some block in the route of the robot. The system can just wait for the assistance of the human operator. The operator looks at the camera image and decides on how the robot should continue. 

 That kind of self-learning algorithm is hard to make. And for the cases where artificial intelligence is operating in closed and controlled areas. They don't need the ability to learn new things. 

All artificial intelligence-controlled systems are not learning new things. Their programming bases two or three tables where is the route to the solution. And the system selects the route by following certain algorithms. 

That kind of system can use in the robots that are working in limited areas like warehouses. Those robot carriers might have certain routes for how they are carrying the cargo. And if some route is blocked, the robot changes some other route.  The system might observe areas by using cameras and triangular meters. That is locating the RFID system of the carriers for locating them. In that kind of system, the merchandise that is going the same way can be put on the shelves connected to the fixed cranes. 

That is putting those shelves on the carriers. That thing keeps the moving robots as simple as possible. In this kind of system, the robots that are moving shelves have a minimum number of parts. All other parts are in the fixed positions. And that thing means that there is a minimum number of possible errors in the moving systems. The thing is that the automatic forklifts need more parts like the photovoltaic cell-based positioning system, which is positioning the lifter in the right position. 

And the sensor what is measuring the pressure when the system puts the lifter under the shelf. The system looks for the right merchandise by using RFID systems. The sorting of that merchandise can happen on the conveyor belt. When packages are moving on the conveyor belt. The system selects them by checking the destiny of those packages. And how the system knows what shelf or box it would select? 

The system knows how many packages are going to a certain location and the size of those packages. Then the system will just load the box on the carrier, which is driven under the point where the system pushes the packages out from the conveyor belt. But those boxes can be on the fixed cranes and the cargo robot will drive them to the right gate. 

These kinds of systems are complicated, but they are not learning anything new. When robots are collecting things from the warehouse, they must only know the recognition number and then find the gate where they will drive those merchandise. 

The only thing that changes are the aiming numbers. Those kinds of systems are suitable in the closed areas, where is always similar order. There are no big surprises and the system might be made safe by denying people go to the working area of robots. The only surprise that might come to the front of the robots is that some route is blocked. 

Image.()https://blog.igus.eu/should-your-warehouse-invest-in-robotics-in-2021/