Mobile Robotics

Definitional Boundaries & Taxonomy

Mobile robotics in the industrial and intralogistics context primarily bifurcates into two distinct taxonomical branches: Automated Guided Vehicles (AGVs) and Autonomous Mobile Robots (AMRs). The definitional boundary lies strictly in their navigational autonomy, computational payload, infrastructure dependence, and dynamic decision-making capabilities. AGVs are fundamentally deterministic, infrastructure-bound robotic platforms. They rely on physical or virtual guides: such as magnetic tape, inductive wires embedded in concrete, passive RFID tags arranged in Cartesian grids, or reflective laser targets: to follow predefined, rigid paths. Their control loop is fundamentally closed around following a static trajectory; if an obstacle is encountered, the deterministic logic dictates a complete halt until the path is cleared.

In stark contrast, AMRs are probabilistic, highly cognitive, and infrastructure-free systems. They utilize Natural Feature Navigation (NFN) paired with sophisticated SLAM (Simultaneous Localization and Mapping) to dynamically understand their topological environment, predict the vectors of dynamic obstacles, and autonomously recalculate trajectories to achieve a predefined goal state. An AMR is a true autonomous intelligent agent operating within a stochastic, unstructured environment, computing potential fields and dynamic windows on the fly. Conversely, an AGV is an automated execution unit operating within a highly structured environment where the "intelligence" resides in the centralized traffic controller rather than the edge device. The taxonomy extends further into specific form factors: towing vehicles (tuggers), unit load carriers (deck-based platforms), pallet trucks, automated forklift AGVs (reach trucks, VNA trucks), and specialized heavy-payload omnidirectional platforms for aerospace and automotive chassis transport.