holonomic vs nonholonomic motion

Holonomic and nonholonomic describe a deep difference in how freely a robot can move: whether it can instantly travel in every direction it can point in, or whether some directions are blocked at any given moment even though it could still reach them by maneuvering. A holonomic robot can move any way it likes right now — forward, sideways, diagonally, or spin in place — with no direction off-limits. A nonholonomic robot cannot: at this instant some motions are simply impossible, so to go those ways it must first shuffle and turn.

The everyday picture is a shopping cart versus a car. A cart with casters on all four corners is holonomic — you can slide it straight sideways into a tight gap without turning it at all. A car is nonholonomic: it cannot slide sideways into a parking spot, even though that spot is clearly reachable, because its wheels only roll forward and back and steer. That is exactly why parallel parking takes a back-and-forth wiggle: the car has to convert a series of legal forward and reverse arcs into the sideways shift it cannot do directly. The restriction is on which directions are available at each instant, not on which places are ultimately reachable.

This distinction shapes how planners must work. For a holonomic robot, a planner can often draw a route as if the machine were a free-floating dot, since it can follow any squiggle. For a nonholonomic robot, the planner must respect the movement rule at every step — only producing paths the machine can actually trace, full of the curves and reversing maneuvers its wheels allow — which makes the planning genuinely harder. The deeper idea behind the labels is about counting: when the number of directions a robot can command at once equals its total number of independent ways to be positioned (its degrees of freedom), it is holonomic; when it can command fewer directions than that, it is nonholonomic.