Robotics involves designing and programming machines capable of functioning in complex, real-world environments. In many ways, it represents one of the most demanding applications of artificial intelligence, as it draws on nearly every branch of the field.

For instance:

• Computer vision and speech recognition help robots perceive and interpret their surroundings.

• Natural language processing, information retrieval, and reasoning under uncertainty allow robots to understand human commands and anticipate the outcomes of their actions.

• Cognitive modeling and affective computing enable more natural interactions between robots and people by recognizing and responding to human emotions or simulating them.

Because many of these challenges are best addressed using machine learning techniques, machine learning plays a critical role in advancing AI for robotics.

What is a robot?

At its core, a robot is a programmable machine equipped with sensors to perceive its surroundings and actuators to interact with the environment. These components allow it to carry out specific tasks or sequences of actions. While popular media often portrays robots as humanoid figures with stiff movements and robotic voices, real-world robots are usually built with a focus on functionality rather than human resemblance.

In practice, most robots look nothing like humans. Their design depends heavily on their intended use. For example, a dishwasher is not shaped like a person but is still a robotic system that cleans dishes using water jets, sensors, and mechanical components.

Interestingly, robots aren’t limited to stationary machines—many autonomous or semi-autonomous vehicles equipped with sensors and actuators also fall under robotics. However, purely software-based tools, like customer service chatbots (sometimes called "software robots"), do not qualify as actual robots in the traditional sense of robotics.

Source: https://builtin.com/robotics