A perfect design: The multifunctional muscle

How Multifunctioning Joints Produce Highly Agile Limbs in Animals with Lessons for Robotics

This paper reviews how multifunctioning joints produce highly agile limbs in animals with lessons for robotics. One of the key reasons why animals are so fast and agile is that they have multifunctioning joints in their limbs. The multifunctioning joints lead to a high degree of compactness which then leads to a host of benefits such as low mass, low moment of inertia and low drag. This paper presents three case studies of multifunctioning joints-the human wrist joint, knee joint and foot joints-in order to identify how multifunctioning is achieved and what lessons can be learned for robotics. It also reviews the multifunctioning nature of muscle which plays an important role in joint actuation. A key finding is that multifunctioning is achieved through various means: multiple degrees of freedom, multifunctioning parts, over-actuation and reconfiguration. In addition, multifunctioning is achieved through highly sophisticated layouts with high levels of integration and fine-tuning. Muscle also makes an important contribution to animal agility by performing multiple functions including providing shape, protection and heat. The paper reviews progress in achieving multifunctioning in robot joints particularly for the wrist, knee and foot. Whilst there has been some progress in creating multifunctioning robotic joints, there is still a large gap between the performance of animal and robotic joints. There is an opportunity to improve the agility of robots by using multifunctioning to reduce the size and mass of robotic joints.

Universal optimal design in the vertebrate limb pattern and lessons for bioinspired design

This paper broadly summarizes the variation of design features found in vertebrate limbs and analyses the resultant versatility and multifunctionality in order to make recommendations for bioinspired robotics. The vertebrate limb pattern (e.g. shoulder, elbow, wrist and digits) has been proven to be very successful in many different applications in the animal kingdom. However, the actual level of optimality of the limb for each animal application is not clear because for some cases (e.g. whale flippers and bird wings), the basic skeletal layout is assumed to be highly constrained by evolutionary ancestry. This paper addresses this important and fundamental question of optimality by analysing six limbs with contrasting functions: human arm, whale flipper, bird wing, human leg, feline hindlimb and frog hindlimb. A central finding of this study is that the vertebrate limb pattern is highly versatile and optimal not just for arms and legs but also for flippers and wings. One key design feature of the vertebrate limb pattern is that of networks of segmented bones that enable smooth morphing of shapes as well as multifunctioning structures. Another key design feature is that of linkage mechanisms that fine-tune motions and mechanical advantage. A total of 52 biomechanical design features of the vertebrate limb are identified and tabulated for these applications. These tables can be a helpful reference for designers of bioinspired robotic and prosthetic limbs. The vertebrate limb has significant potential for the bioinspired design of robotic and prosthetic limbs, especially because of progress in the development of soft actuators.

On the role of recurrent inhibitory feedback in motor control

This article reviews presumed roles of recurrent inhibition in motor control, that have been proposed over the past five decades. The discussion is structured in an order of increasing complexity. It starts out with the simplest and earliest circuit, that is recurrent self-inhibition of skeleto-motoneurons, and related functions. It soon becomes clear that in order to understand recurrent inhibition, we must look beyond the simple self-inhibitory CNS circuit. First, recurrent inhibition must be seen in the context of other neural circuits. Second, some quantitative features appear to be correlated with features of the neuromusculo-skeletal periphery. Third, the aspect of lateral inhibition between different members of a motoneuron pool as well as between different motoneuron pools points to the essential multiple input-multiple output structure of recurrent inhibition that again can be understood only by correlating it with features of the neuromusculo-skeletal periphery. Another extension results from the discovery that recurrent inhibition affects not only skeleto-motoneurons, but also gamma-motoneurons, Ia inhibitory interneurons mediating reciprocal inhibition between antagonist motoneurons, other Renshaw cells and cells of origin of the ventral spinocerebellar tract (VSCT). Then the view broadens again, investigating the potential role that recurrent inhibition plays in two far-ranging theories of motor control, the inverse-dynamics approach and the equilibrium-point hypothesis. Finally, the present author tries to formulate, in broad strokes, a personal functional interpretation of recurrent inhibition. All the functional considerations, right or wrong, should yield ideas for new experiments, and this then is the last objective of this review.

Interaction of recurrent inhibitory and muscle spindle afferent feedback during muscle fatigue

Mammalian skeletal motor units have differing properties including their different susceptibility to fatigue. The question discussed in this paper is whether and to what extent proprioceptive feedback via muscle spindles can contribute to shape the firing patterns of motor units so as to minimize their loss of force during fatiguing contraction. The firing of a skeleto-motoneuron dispatches signals which are fed back to the same and homonymous as well as synergistic motoneuron. Two feedback pathways are of concern here: one via the related muscle unit and muscle spindle afferents (proprioceptive path), and one via recurrent motor axon collaterals and Renshaw cells (recurrent inhibitory path). It is suggested that the contraction of a motor unit or a small group of adjacent ones is signalled to the homonymous alpha-motoneurons via proprioceptive afferents, the signal being filtered and enhanced by spinal recurrent inhibition. This is effected by timed correlation of the signals which are propagated through the two feedback loops. The effects of the correlation can be strengthened by (i) topographical order of the feedback connections, (ii) heterosynaptic interaction, and (iii) tendencies towards synchronous discharge between motoneurons. These mechanisms render the possibility more likely that information about the unfused contractions of a muscle unit (or a small group of them), mediated via proprioceptive afferents, play a role in shaping the precise discharge pattern of the innervating motoneuron(s). These mechanisms could be used to optimize the force output of large fatiguing motor units during long activation, during which their activation rates normally decrease (adapt) over time. Our results show that during adapting motoneuron firing Renshaw cells and muscle spindle afferents may show discharge patterns which at least in part are in keeping with such an hypothesis.