Humanlike Articulated Robotic Headform for Respirator Fit Testing

These prototypes can replace humans in testing of respiratory protection devices.

The testing of individual respiratory protection (IRP) devices is now accomplished with panels of human wearers. Historical at tempts to simulate the human face and head have been unsuccessful for a variety of reasons that include imprecision in reproduction of facial dimensions and unrepresentative textures of the surfaces applied to headforms.

Figure 1. Static headform with Anatomically Accurate Facial “Skin” thicknesses. The skin was rigorously measured with a lance to validate the accuracy of the tissue thicknesses. The completed headform was mounted on a portable acrylic base and prepared for use in the laboratory.
Two new headforms have been designed and developed: one static, and one articulated and actuated to accurately reproduce the facial gestures and movements of a human whose anthropometric dimensions fall within the window identified by the National Institute for Occupational Safety and Health (NIOSH) as a short/wide headform, and just outside the window for a medium headform.

The static headform was built first as a prototype to work out many of the details of material composition and configuration. Skin thickness was specified to reproduce a set of ultrasound measurements reported for a panel of young adult male Caucasians. This was accomplished by an inverse forensic reconstruction technique in which a mold of the NIOSH medium headform was cast and clay was applied to the (negative) surface of the mold in a layer whose local thicknesses matched the measured skin thickness. Pins inserted into the negative surface guided the clay thickness. The skull form was then cast to the clay negative surface, and used subsequently as a complementary mold surface to cast the polydimethylsiloxane (PDMS) skin (Frubber™) with anatomically accurate thickness (see Figure 1). For this, the molds were oriented with locating pins, and for assembly, the skull form and skin included locating dimples in areas noncritical to respirator fitting.

A simplified fit test under static conditions at NIOSH showed this head to achieve fit factors (FFs) slightly better than average values measured for the complete protocol, which includes movement. The result was considered completely successful for the first stage, and a process of design refinement and construction of an articulated, robotic headform in the same dimensions was undertaken.

Feedback that the skin thickness near the top of the head was a bit too thin led to a jig that has since been used to verify dimensions during assembly. The temples and back of the head also were too small (see Figure 2), which resulted from deformation of the surface during the assembly process. Jigs were devised to control the surface form during assembly, and to verify dimensions prior to shipment.

Figure 2. Feedback from NIOSH about the Static Headform indicated that the temples and back of the head were too small (violet and dark blue areas), which resulted from deformation of the surface during the assembly process. Jigs were devised to control the surface form during assembly, and to verify dimensions prior to shipment.
Incorporation of anchors used to connect the skin through cables to individual servos caused moderate (5%) swell ing of the PDMS after removal from the molds. A series of washing and baking steps restored the original dimension, but the texture of the skin was unsatisfactory, which triggered several cycles of reformulation and casting that led to a Frubber™ composition that exhibits both stable dimensions and appropriate mechanical properties. Routing of the breathing tube down through the throat was precluded by the array of components in the lower jaw used to drive lower facial expressions and movements, a problem aggravated by an increase in its inside diameter from ¾" to 1". The solution selected was to route the air up the front of the face, over the top and down the back of the head, and then down the back of the neck.

The stiffest challenge was achieving an airtight seal around the electronic and mechanical components enclosed in the skull form from aerosolized water and salts in both the external environment and the breathing tube, through which the same aerosol was drawn through a test respirator and eventually delivered to sampling equipment. After several false starts, this was accomplished by inverting the physiology of the oral cavity in an elastomeric casting that was sealed with an adhesive to the lips of the Frubber™ skin, and that engaged the breathing tube on its upper face.

Initial fit testing results suggest that the static headform accurately reproduces the dimensions and textures of the NIOSH medium head. The first application targeted is upgrading N95 respirator certification under 42 CFR 84 to a basis of protection factor (PF) rather than particle exclusion by the medium; however, markets to support respirator design, leak testing, use of hazardous challenges, and PF measurements during extreme exercise/stress are anticipated.

This work was done by Joseph Wander of the Air Force Research Laboratory, and David Hanson and Richard Margolin of Hanson Robotics. AFRL-0224