The idea that the physical design of a robotic arm like Festo's Bionic Handling Assistant could reduce internal fragmentation in robotic vision systems is a shortcut that ignores the real complexity.
To reach full performance expression, much more than an articulated structure is needed; the integration of sensors and vision is primarily a colossal software and algorithmic task.
Take a warehouse robot: if the cameras can't compensate for lighting changes or poorly positioned objects, it's not the arm's fault but the algorithm that needs improvement.
We must seek optimization without compromise at every level, not a miracle solution from a single component.