Robotics
Robotics is the field of designing and controlling machines that act in the physical world. It brings together mechanisms, sensing, planning and control, with or without machine learning.
A warehouse robot needs to estimate where it is, plan a route to a shelf and command its motors. A camera or distance sensor supplies observations; planning selects a path; control turns that plan into movement. These parts must work together even when a sensor reading is imperfect.
A language model might turn “take this box to aisle three” into a task request. It does not remove the need for localization, motion planning or control of the actual machine. A software-only artificial intelligence (AI) agent can choose actions too, but it does not necessarily have a physical body.
The distinction matters because physical actions have constraints a text response does not. The robot must account for obstacles, actuator limits and the timing of its measurements. Testing a plan in simulation is useful, but the real machine still needs tests that expose differences in sensing and motion.
Sources
- Dellaert and Hutchinson: Models for robotics — Connects world state, robot state, action models, sensors and low-level controllers.
Go deeper
- Dellaert and Hutchinson: Introduction to Robotics and Perception course
Follow worked robot examples from trash sorting to mobile robots and autonomous vehicles.