kinematic redundancy

A robot is kinematically redundant when it has more joints to move than the job strictly requires — extra wiggle room left over after the task is satisfied. The clearest example is your own arm. To touch a fixed spot on a table, your hand only needs to land at that point, yet your shoulder, elbow, and wrist together give you far more freedom than that one task demands. The proof: keep your fingertip pinned on the table and you can still swing your elbow up and down. That leftover motion, which changes the arm's shape without moving the goal, is exactly what redundancy buys you.

We measure it as a simple subtraction: the number of joint freedoms minus the number of freedoms the task actually needs. A common arm has six joints, and placing a tool at a point with a chosen orientation also takes six freedoms, so it is just barely sufficient with nothing to spare. Add a seventh joint and you have one degree of redundancy. That spare degree shows up as self-motion: the robot can rearrange its own elbow, shoulder, or middle links while the working tip stays perfectly still on target — like reshuffling the slack in a loose rope without moving either end.

This spare freedom is enormously useful, which is why advanced arms and humanoids are deliberately built redundant. The extra motion lets a robot keep working while dodging an obstacle in its path, steering clear of awkward locked-up poses, staying away from its joint limits, or simply holding a more comfortable and energy-saving stance. The cost is that there is no longer a single right answer for how to bend the joints — infinitely many joint settings reach the same goal — so the control software must add a tie-breaking rule (be smooth, be efficient, avoid that wall) to choose gracefully among them.