Novel Method for the Optimal Design of a Robotic Neck Mechanism to Mimic Human Neck Motion: a Data-Driven Approach

In recent decades, the development of robotic head and neck mechanisms able to mimic human actions has been gaining attraction in several fields of application, resulting in a fertile research area. Most past works focused on studying human-like interactive features crucial in the context of healthcare robots without concentrating on reproducing the human neck motion naturally. Nowadays, in a world where human safety is increasingly gaining importance in daily life and working environments, robotic neck mechanisms that accurately mimic human neck behaviour represent a game-changer in performing tests to develop personal protective equipment and estimate head and neck injury risks. Up to the present day, the design of serial robotic neck mechanisms developed exploiting the biomimetic approach is restricted to defining the limits of the movements and velocity of the neck by using a statistically unrepresentative sample of anatomical measurements. This work proposes a novel data-driven method to find the optimal design for a four Degrees of Freedom serial robotic neck mechanism. First, an experimental campaign is conducted to acquire neck motion data on six human subjects in three height male percentile groups. Then, an optimisation process computes the optimal geometrical parameters for each group. Finally, a prototype of the robotic neck mechanism is developed and tested. This work is the first step in a broader study that aims to design a robotic neck mechanism with an adjustable geometry that mimics the head motion of any subject given the percentile group and some anatomical measurements.