This manipulator was developed for the RoboSoft 2016 Grand Challenge (Manipulation scenario). In particular, Manipulation Task 2 required the robot to reach and touch a can located behind multiple vertical obstacles, where avoiding any contact with the obstacles yielded the highest score. With conventional manipulators, reaching such a target while keeping the entire arm collision-free is challenging because it requires continuous whole-body control of the arm’s global shape and joint configuration.

We addressed this by adopting a snail eye-stalk–inspired, tip-growing extension strategy. Instead of pushing an already-deployed body forward, the manipulator “grows” at the tip while the base remains fixed. As a result, obstacle avoidance becomes a tip-navigation problem: once the tip threads between the obstacles, the previously deployed body remains in the cleared path and does not require repeated global reconfiguration. Based on this concept, we built a tip-growing extendable manipulator and competed in the RoboSoft 2016 Grand Challenge.

By conceptually splitting the cross-section of a snail eye-stalk mechanism, we obtain a long ∩-shaped structure where one end is anchored while the other side is pushed from within, producing motion equivalent to tip growth. Inspired by this, we implemented a simple tip-growing mechanism by bending a tape measure into a ∩ shape, fixing one side, and pushing the other so that the tip extends while the base remains stationary.

For steering, we introduced auxiliary steering units that locally bend the body and tilt the growth direction. Once a bend is formed, subsequent extension proceeds along the bent direction, enabling directional growth. To meet the manipulation scenario requirements, we integrated two steering units, each delivered by its own tip-growing extendable mechanism—resulting in a three-layer extendable system in the final manipulator.

We participated in the RoboSoft 2016 Grand Challenge; however, a hardware failure prevented successful task completion.