Caninoid Necro-Robots: Geometrically Selected Rearticulation of Canine Mandibles ^†

Abstract

In line with Sustainable Development Goal 9 (sustainable industrialisation and innovation), environmentally responsible engineering designs in modern robotics should consider factors such as renewability, sustainability, and biodegradability. The robotics sector is growing at an exponential rate and, as a consequence, its contribution to e-waste is a growing concern. Our work contributes to the technological development of caninoid necro-robots, robots that are built from the skeletons of deceased dogs. The already formed skeletal structures of deceased dogs (and other animals) are ideal natural material replacements for synthetic robotic architectures such as plastics, metals, and composites. Since dog skeletons are disarticulated, simple but effective methods need to be developed to rearticulate their bodies. The canine head is essentially a large end effector, but its mandible is held together by a fibrocartilaginous joint (symphysis) that degrades at a higher rate than the bone itself. The degradation of the symphysis would ordinarily negate the utility of a canine head as a necro-robotic end effector; however, in this research, we consider simple methods of mandible reinforcement to circumvent this problem. Our research uses 3D scans of a real canine head, which is modelled using the finite element method to ascertain optimal geometrical reinforcements for the mandible. The full head structures and their reinforcements are printed and adhesively connected to determine the most effective reinforcing strategy of the mandible. Here, we elucidate geometrically selected reinforcement designs that are evidenced through mechanical testing, to successfully increase the stiffness of a disarticulated mandible.

1. Introduction

Despite a global push towards net zero by 2050 [1] and the necessity of finding new and innovative solutions to reduce waste and environmental harm [2,3], there has been very little development in the robotics sector. This may be attributed to the nascent nature of the robotics industry, with a limited number of companies offering robots for use at scale. As a result of this, the products available within the robotics sector rely heavily on virgin materials, petroleum-based plastics and energy intensive manufacturing processes. Necro-robotics aims to provide an alternative solution to these growing issues.

A necro-robot is a robotic architecture formed using the body parts of deceased animals. This is different from cyborg animals (such as cyborg insects [4,5,6,7,8]) where the animal subject is still alive and electrical stimuli are used to control the animal body parts. The design and application of necro-robots is a novel field of research with very limited existing literature mainly focusing on spider cadavers [9] and extinct species [10]. The possibility of utilising bone as a key structural material offers enormous scope to re-purpose vast quantities of waste material [11,12]. Through this, both the energy and expense of building a complete robot can be considerably reduced. Necro-robots offer a greatly decreased environmental impact (due to the biodegradability of bone [13]) in their end of life compared to traditional synthetic robots. When combined with the abundance of low-cost materials, they possess the unique quality of being more disposable. This lends them to use in high-risk and sensitive natural environments.

This paper focuses on research into caninoid necro-robots and their potential applications, specifically the head as a load-bearing end effector [14]. In order to achieve this, the key weaknesses within the mandible [15] require reinforcement. Variations in the mandible geometry between the specimens call for the use of adaptive methods such as 3D scanning, computer-aided design (CAD), finite element analysis (FEA), and fused filament fabrication (FFF) prototyping to develop unique reinforcements that can be easily applied. Having procured a historical skull and mandible from the Royal (Dick) Edinburgh Veterinary School (R(D)SVS Veterinary Ethical Review Committee: VERC Ref 209.23) this paper aims to justify the proposed method of designing reinforcements. This paper will then demonstrate that the appropriate application of reinforcements can prevent failure at key sites of weakness within the mandible.

2. Materials and Methods

2.1. 3D Scanning

An EinScan Pro 2X 3D scanner (Shining 3D, Stuttgart, Germany) facilitated the capture of the surface profile of a real dried canine head geometry as a tessellated.stl output (see Figure 1) rendered using Solid Edge Shining 3D software Edition 2022. The surface model file produced by the scanner was post-processed in Mesh-Mixer to smooth the surface geometries and reduce triangle density to maximise ease of processing when using the model in other software. The resulting simplified surface model was converted to a parasolid file using Solidworks and was then imported into various cloud-based modelling software including OnShape for computer-aided designs (CADs) and SimScale for finite element analysis (FEA).

2.2. FEA

A linear elastic isotropic material model was used to simulate the canine mandible under a static structural loading of 50 N. This model provided an understanding of the suitable reinforcement location and design. In addition, the selection of suitable boundary conditions allowed for the design of a fully integrated PLA testing bracket that would provide a realistic response when the FFF archetypes were tested mechanically.

2.3. Reinforcement Design

Three reinforcement iterations were designed to fit the natural curvature of the ventral aspect of the mandible (see Figure 2). The reinforcements were designed with CAD and through a mapping of the mandible 3D scans. The three designs were manufactured using the FFF of polylactic acid (PLA) and then attached to the mandible using a polyepoxide-based adhesive. Each of the three reinforcements was manufactured with the same volume fraction.

2.4. Recreating Natural Weakness

Natural decomposition of the fibrocartilagenous in vivo jointing at the interface intersected by the median plane represents a fundamental weakness caused by maceration and subsequent ageing. To recreate this weakness within the FFF archetypes, the specimens were cut in half at the same location and reconnected with a siloxane-based adhesive.

2.5. Mechanical Testing

A custom designed and built force implementer was integrated with an Instron 3369 mechanical testing unit to apply load to all mandible specimens at both 0 and 37° angles rotated in the dorsoventral axis. For each angle, each specimen was loaded at two positions along the length of the mandible, offering a “normal” and “worst-case” loading scenario. Figure 3 shows the manufactured testing rig mounted in the Instron 3369 Universal Testing Machine with a specimen in position to be tested. The Instron 3369 was linked to the Instron Bluehill 3 Software Version 3.51, which recorded the displacement and load values during testing. The testing rig was held in place through the upper and lower clamp attachments, which were rotatable through 90°, enabling the force applicator to be aligned with the specimen.

To cover different loading techniques, we used the following three parameters: loading position, specimen angle, and reinforcement type. The loading positions were in the middle and at the end (see Figure 4) to replicate an alternative loading scenario. Loading at the end of the mandible replicated a worst-case scenario, with all the force acting through the front teeth; this created the largest bending moment on the mandible. The two specimen angles were 0° and 37° (see Figure 5). Lastly, the specimen type referred to which reinforcement was used, where the reinforcement types were as follows: control, truss, plate, and strut.

3. Results and Discussion

3.1. FEA Results

The FEA results validated our FFF archetype mandible bracket design and aided in predicting the failure location under load. The results for the naturally derived boundary conditions for the mandible geometry can be seen below, highlighting where the stress concentrates (see Figure 6). The results show that stress localises as a maximum on the ventral aspect of the mandible close to the edge of the fixed boundary conditions (see Figure 6A). The second highest stress concentration is found caudal to the last molars (see Figure 6B). This suggests a natural failure point between these two highest stress points.

The reinforcements were then added to the CAD models, and the FEA process was run again for each model. The results were similar and showed near identical locations for maximum stress. It should be noted that the FEA simulations did not include the use of epoxy to bond the reinforcements, and hence they were modelled as one solid object rather than consisting of two components.

3.2. Testing Results

3.2.1. FFF Mandible Archetype Testing

The FFF archetype mandibles and reinforcements were printed and tested. From the data, the values of stiffness were calculated when testing had occurred both in the middle and at the end of the mandible. In all cases the reinforcements outperformed the control at both the 0° and 37° testing angles (see Figure 7 and Figure 8).

3.2.2. Bone Testing

The control test on the bone and the test with the 3D-printed PLA reinforcement returned similar stiffness values. This was unexpected given the improvement shown in the tests on the 3D-printed mandibles. To investigate this, a steel plate reinforcement was attached to the bone sample, which returned a lower stiffness value. It was originally suspected that due to the use of hot-melt adhesive instead of epoxy, there was not a strong enough bond to transfer the bending stress to the reinforcement. Using ImageJ Version 1.54a, an image processing program, the videos were analysed, and it was found that there was no deflection in the bone mandible. This suggests that the displacement observed (Figure 9) was due to the improper fixing of the mandible within the bone testing moulds.

4. Conclusions

All the reinforcements had a positive effect on the strength of the printed models at 0 and 37°. Valuable information on the failure method of the mandible, especially in the symphysis, can be drawn from our testing and will inform the future work in the design of reinforcements for canine necro-robot mandibles. The benefit of reconstructing the mandible is in the improved structural functionality. Whereas in live animals, the mandibular integrity at the symphysis enables the prehension of food, mastication, swallowing, and the carrying of food or objects, a caninoid necro-robot mandibular reconstruction serves the primary purpose of enabling end effector functionality by creating stability, rigidity, and load bearing capacity. Simple mechanisms for articulating the mandible, as researched herein, have broad applicability. This is important as dog skulls come in a variety of sizes and shapes and simple mechanisms are easier to standardise across a wider range of skulls. Further research should be conducted to develop standard methods for connecting and actuating dog skulls and skeletal body parts, as doing so would enable the application of “one size fits all” manufacturing, which would make the use of caninoid necro-robots significantly more feasible in the future.