Large volumes of data represent a true revolution for robotics: machine learning multiplies the capacity to respond to the environment, making robots more autonomous, intelligent, and agile.

Compared to other fields, robotics has particular characteristics because its goal is to enable a physical agent to interact with the real world.

Traditionally, the field of robotics has been dominated by the desire to find solid models based on the principle of response: optimizing the relationship between the input of a stimulus and the output of action or the prediction of a response to a given stimulus.

This worked well in a factory world with well-defined inputs and outputs. However, for years, it has been challenging for robots to move out of perfectly structured factory or laboratory environments into the “real world,” characterized by its wide range of tasks, stimuli, and possible situations.

Machines are rapidly advancing in the field of stimuli and perceptions: they are beginning to see, hear, read, and touch in ways that were previously impossible. This has led to a true revolution in robotics: the ability to learn and deal with direct inputs from the real world greatly enriches the capabilities of robots.

And here, large volumes of information are key. This is because the possibilities of machine learning multiply. The availability of Big Data and new associated learning techniques have the potential to allow robots to understand and operate in significantly more complex environments. Naturally, this should lead to a qualitative leap in the performance and implementation capacity of robotics in a wide range of practical applications and real-world settings.

Arthur Dubrawski, director of Auton Lab—the robotics institute at Carnegie Mellon University’s School of Computer Science—explains that robots have always relied on data because the operational definition of a robot involves performing the following sequence in a loop: sensing, planning, and acting.

The difference now is that Artificial Intelligence and Machine Learning will increasingly play a leading role in this process.

Looking to the future, robotics and manufacturing will increasingly be defined by big data and the ability of AI systems to analyze and act on production information.

Robots can now acquire skills to perform tasks through advanced machine learning, enabled by the development of integrated sensors and cloud-based analytics. This trend will continue with the help of faster 5G wireless networks.

The factories of the future will use Big Data collection and analysis to enable robots to make quick decisions in the manufacturing process, even when presented with unfamiliar equipment and objects. The goal will be to create robots that are more autonomous, intelligent, responsive, and agile.

 Julio Cesar Blanco – November 14, 2022