From bone to plausible bipedal locomotion using inverse kinematics

A Guide to Inverse Kinematic Marker-Guided Rotoscoping using IK Solvers

X-ray Reconstruction of Moving Morphology (XROMM) permits researchers to see beneath the skin, usually to see musculoskeletal movements. These movements can be tracked and later used to provide information regarding the mechanics of movement. Here, we discuss “IK marker-guided rotoscoping”—a method that combines inverse kinematic solvers with that of traditional scientific rotoscoping methods to quickly and efficiently overlay 3D bone geometries with the X-ray shadows from XROMM data. We use a case study of three Nile crocodiles’ (Crocodylus niloticus) forelimbs and hindlimbs to evaluate this method. Within these limbs, different marker configurations were used: some configurations had six markers, others had five markers, and all forelimb data only had three markers. To evaluate IK marker-guided rotoscoping, we systematically remove markers in the six-marker configuration and then test the magnitudes of deviation in translations and rotations of the rigged setup with fewer markers versus those of the six-marker configuration. We establish that IK marker-guided rotoscoping is a suitable method for “salvaging” data that may have too few markers.

An approach to drastically reduce the required legs DOFs for bipedal robots and lower-limb exoskeletons

We present a method to drastically reduce the required number of degrees-of-freedom (DOFs) needed for walking for each leg of bipedal robots and lower-limb exoskeletons. This approach releases more legs DOFs in the null space to do other tasks, instead of unnecessarily constraining them. It uses relative reference frames to control relative motion between the two feet, instead of the usual method of controlling foot movement with respect to fixed reference frames. In its basic form, it controls the bipedal walking holistically using two controllers: (1) world space control using relative feet motion and (2) null-space control of the legs posture.

Integrating morphology and in vivo skeletal mobility with digital models to infer function in brittle star arms

Brittle stars (Phylum Echinodermata, Class Ophiuroidea) have evolved rapid locomotion employing muscle and skeletal elements within their (usually) five arms to apply forces in a manner analogous to that of vertebrates. Inferring the inner workings of the arm has been difficult as the skeleton is internal and many of the ossicles are sub-millimeter in size. Advances in 3D visualization and technology have made the study of movement in ophiuroids possible. We developed six virtual 3D skeletal models to demonstrate the potential range of motion of the main arm ossicles, known as vertebrae, and six virtual 3D skeletal models of non-vertebral ossicles. These models revealed the joint center and relative position of the arm ossicles during near-maximal range of motion. The models also provide a platform for the comparative evaluation of functional capabilities between disparate ophiuroid arm morphologies. We made observations on specimens of Ophioderma brevispina and Ophiothrix angulata. As these two taxa exemplify two major morphological categories of ophiuroid vertebrae, they provide a basis for an initial assessment of the functional consequences of these disparate vertebral morphologies. These models suggest potential differences in the structure of the intervertebral articulations in these two species, implying disparities in arm flexion mechanics. We also evaluated the differences in the range of motion between segments in the proximal and distal halves of the arm length in a specimen of O. brevispina, and found that the morphology of vertebrae in the distal portion of the arm allows for higher mobility than in the proximal portion. Our models of non-vertebral ossicles show that they rotate further in the direction of movement than the vertebrae themselves in order to accommodate arm flexion. These findings raise doubts over previous hypotheses regarding the functional consequences of ophiuroid arm disparity. Our study demonstrates the value of integrating experimental data and visualization of articulated structures when making functional interpretations instead of relying on observations of vertebral or segmental morphology alone. This methodological framework can be applied to other ophiuroid taxa to enable comparative functional analyses. It will also facilitate biomechanical analyses of other invertebrate groups to illuminate how appendage or locomotor function evolved.

Learning Kinematic Structure Correspondences Using Multi-Order Similarities

In this paper, we present a novel framework for finding the kinematic structure correspondences between two articulated objects in videos via hypergraph matching. In contrast to appearance and graph alignment based matching methods, which have been applied among two similar static images, the proposed method finds correspondences between two dynamic kinematic structures of heterogeneous objects in videos. Thus our method allows matching the structure of objects which have similar topologies or motions, or a combination of the two. Our main contributions can be summarised as follows: (i) casting the kinematic structure correspondence problem into a hypergraph matching problem by incorporating multi-order similarities with normalising weights, (ii) introducing a structural topology similarity measure by aggregating topology constrained subgraph isomorphisms, (iii) measuring kinematic correlations between pairwise nodes, and (iv) proposing a combinatorial local motion similarity measure using geodesic distance on the Riemannian manifold. We demonstrate the robustness and accuracy of our method through a number of experiments on synthetic and real data, outperforming various other state of the art methods. Our method is not limited to a specific application nor sensor, and can be used as building block in applications such as action recognition, human motion retargeting to robots, and articulated object manipulation amongst others.

Robust identification of three-dimensional thumb and index finger kinematics with a minimal set of markers

This study presents a methodology to determine thumb and index finger kinematics while utilizing a minimal set of markers. The motion capture of skin-surface markers presents inherent challenges for the accurate and comprehensive measurement of digit kinematics. As such, it is desirable to utilize robust methods for assessing digit kinematics with fewer markers. The approach presented in this study involved coordinate system alignment, locating joint centers of rotation, and a solution model to estimate three-dimensional (3-D) digit kinematics. The solution model for each digit was based on assumptions of rigid-body interactions, specific degrees of freedom (DOFs) at each located joint, and the aligned coordinate system definitions. Techniques of inverse kinematics and optimization were applied to calculate the 3-D position and orientation of digit segments during pinching between the thumb and index finger. The 3-D joint center locations were reliably fitted with mean coefficients of variation below 5%. A parameterized form of the solution model yielded feasible solutions that met specified tolerance and convergence criteria for over 85% of the test points. The solution results were intuitive to the pinching function. The thumb was measured to be rotated about the CMC joint to bring it into opposition to the index finger and larger rotational excursions (>10 deg) were observed in flexion/extension compared to abduction/adduction and axial rotation for all joints. While the solution model produced results similar to those computed from a full marker set, the model facilitated the usage of fewer markers, which inherently lessened the effects of passive motion error and reduced the post-experimental effort required for marker processing.