A biologically inspired controller for hexapod walking: simple solutions by exploiting physical properties

The Role of Muscle Spindle Feedback in the Guidance of Hindlimb Movement by the Ipsilateral Forelimb during Locomotion in Mice

Safe and efficient locomotion relies on placing the foot on a reliable surface at the end of each leg swing movement. Visual information has been shown to be important for determining the location of foot placement in humans during walking when precision is required. Yet in quadrupedal animals where the hindlimbs are outside of the visual field, such as in mice, the mechanisms by which precise foot placement is achieved remain unclear. Here we show that the placement of the hindlimb paw is determined by the position of the forelimb paw during normal locomotion and in the presence of perturbations. When a perturbation elicits a stumbling corrective reaction, we found that the forelimb paw shifts posteriorly relative to body at the end of stance, and this spatial shift is echoed in hindlimb paw placement at the end of the swing movement. Using a mutant mouse line in which muscle spindle feedback is selectively removed, we show that this posterior shift of paw placement is dependent on muscle spindle feedback in the hindlimb but not in the forelimb. These findings uncover a neuronal mechanism that is independent of vision to ensure safe locomotion during perturbation. This mechanism adds to our general knowledge of how the nervous system controls targeted limb movements and could inform the development of autonomous walking machines.

Climbing favours the tripod gait over alternative faster insect gaits

To escape danger or catch prey, running vertebrates rely on dynamic gaits with minimal ground contact. By contrast, most insects use a tripod gait that maintains at least three legs on the ground at any given time. One prevailing hypothesis for this difference in fast locomotor strategies is that tripod locomotion allows insects to rapidly navigate three-dimensional terrain. To test this, we computationally discovered fast locomotor gaits for a model based on Drosophila melanogaster. Indeed, the tripod gait emerges to the exclusion of many other possible gaits when optimizing fast upward climbing with leg adhesion. By contrast, novel two-legged bipod gaits are fastest on flat terrain without adhesion in the model and in a hexapod robot. Intriguingly, when adhesive leg structures in real Drosophila are covered, animals exhibit atypical bipod-like leg coordination. We propose that the requirement to climb vertical terrain may drive the prevalence of the tripod gait over faster alternative gaits with minimal ground contact.

Numerous selective forces shape animal locomotion patterns and as a result, different animals evolved to use different gaits. Here, Ramdya et al. use live and in silico Drosophila, as well as an insect-model robot, to gain insights into the conditions that promote the ubiquitous tripod gait observed in most insects.

Rigidity and Flexibility: The Central Basis of Inter-Leg Coordination in the Locust

Many motor behaviors, and specifically locomotion, are the product of an intricate interplay between neuronal oscillators known as central pattern generators (CPGs), descending central commands, and sensory feedback loops. The relative contribution of each of these components to the final behavior determines the trade-off between fixed movements and those that are carefully adapted to the environment. Here we sought to decipher the endogenous, default, motor output of the CPG network controlling the locust legs, in the absence of any sensory or descending influences. We induced rhythmic activity in the leg CPGs in isolated nervous system preparations, using different application procedures of the muscarinic agonist pilocarpine. We found that the three thoracic ganglia, each controlling a pair of legs, have different inherent bilateral coupling. Furthermore, we found that the pharmacological activation of one ganglion is sufficient to induce activity in the other, untreated, ganglia. Each ganglion was thus capable to impart its own bilateral inherent pattern onto the other ganglia via a tight synchrony among the ipsilateral CPGs. By cutting a connective and severing the lateral-longitudinal connections, we were able to uncouple the oscillators’ activity. While the bilateral connections demonstrated a high modularity, the ipsilateral CPGs maintained a strict synchronized activity. These findings suggest that the central infrastructure behind locust walking features both rigid elements, which presumably support the generation of stereotypic orchestrated leg movements, and flexible elements, which might provide the central basis for adaptations to the environment and to higher motor commands.

Fast damage recovery in robotics with the T-resilience algorithm

Damage recovery is critical for autonomous robots that need to operate for a long time without assistance. Most current methods are complex and costly because they require anticipating potential damage in order to have a contingency plan ready. As an alternative, we introduce the T-resilience algorithm, a new algorithm that allows robots to quickly and autonomously discover compensatory behavior in unanticipated situations. This algorithm equips the robot with a self-model and discovers new behavior by learning to avoid those that perform differently in the self-model and in reality. Our algorithm thus does not identify the damaged parts but it implicitly searches for efficient behavior that does not use them. We evaluate the T-resilience algorithm on a hexapod robot that needs to adapt to leg removal, broken legs and motor failures; we compare it to stochastic local search, policy gradient and the self-modeling algorithm proposed by Bongard et al. The behavior of the robot is assessed on-board thanks to an RGB-D sensor and a SLAM algorithm. Using only 25 tests on the robot and an overall running time of 20 min, T-resilience consistently leads to substantially better results than the other approaches.

EMULATION OF HUMAN FEELINGS AND BEHAVIORS IN AN ANIMATED ARTWORK

The behavior of an animated artwork, survivor — a classroom chair which walks, with a dynamics which some viewers find haunting — reflects an attempt to emulate (and suggest to viewers) some feelings and behaviors that are typical of survivors of landmine blasts, learning to use crutches. The artwork itself is intended for sensitizing viewers to the horror experienced by those who survive, and those who do not. The behavior of such a survivor is affected by several factors: some are due to the objective difficulty of using prosthetic legs, and some are due to emotional factors, e.g., fear, "shame" of being in such situation, and pain. The mechanical structure, strongly conditioned by artistic requirements, was combined with a control system that exhibits appropriate behaviors. Behavioral control, a technique developed for the control of mobile robots, was used in survivor, and implemented over a modified version of the traditional Brooks' subsumption architecture. This technique makes it possible to emulate normal locomotion behaviors such as the need of avoiding obstacles and typical animal feelings such as curiosity, hunger, fatigue and fear. We describe the mechanics and viewers' response, and formalize aesthetic response. We briefly survey computer modelling of emotions, robotic art, and biomimetic locomotion in robotics.

Thomas Graham Brown (1882--1965), Anders Lundberg (1920-), and the neural control of stepping

Thomas Graham Brown (1882--1965) undertook experiments on the neural control of stepping in the University of Liverpool laboratory of Charles Sherrington (1857--1952) in 1910--13 and his own laboratory in 1913--15 at the University of Manchester. His results revealed the intrinsic capability of the spinal cord in the guinea pig and cat to generate a stepping output pattern whose timing did not depend upon descending or sensory inputs. This idea was then revolutionary because the prevailing viewpoint was that the stepping rhythm was generated by spinal reflexes. Sadly, Graham Brown's GBR peers gave little credence to this seminal accomplishment, except perhaps Sherrington, who waxed but largely waned on the potential significance of the work. It remained for the Swedish neuroscientist, Anders Lundberg (1920-), to rescue Graham Brown's concepts from obscurity: in seminars presented in several countries between 1957 and 1980, and in widely read articles and reviews (1965--1981). Graham Brown had proposed mutually inhibitory connections between a pair of intrinsically active flexor and extensor "half-centers" on each side of the spinal cord, with the rhythmic output modulated by sensory proprioceptive input. Lundberg, Elzbieta Jankowska (1930-), and their colleagues provided seminal, compelling evidence for spinal half-center interneuronal circuitry implicated in the control of stepping and Lundberg and Ingemar Engberg (1935--2005) made behavioral EMG observations on unrestrained cats that supported a central generation of the rhythm. Subsequently, models of the spinal pattern generators for mammalian locomotion have become progressively more complex but they mostly still include a half-center component.

A dynamic model of thoracic differentiation for the control of turning in the stick insect

Leg movements of stick insects (Carausius morosus) making turns towards visual targets are examined in detail, and a dynamic model of this behaviour is proposed. Initial results suggest that front legs shape most of the body trajectory, while the middle and hind legs just follow external forces (Rosano H, Webb B, in The control of turning in real and simulated stick insects, vol. 4095, pp 145-156, 2006). However, some limitations of this explanation and dissimilarities in the turning behaviour of the insect and the model were found. A second set of behavioural experiments was made by blocking front tarsi to further investigate the active role of the other legs for the control of turning. The results indicate that it is necessary to have different roles for each pair of legs to replicate insect behaviour. We demonstrate that the rear legs actively rotate the body while the middle legs move sideways tangentially to the hind inner leg. Furthermore, we show that on average the middle inner and hind outer leg contribute to turning while the middle outer leg and hind inner leg oppose body rotation. These behavioural results are incorporated into a 3D dynamic robot simulation. We show that the simulation can now replicate more precisely the turns made by the stick insect.

Insect walking is based on a decentralized architecture revealing a simple and robust controller

Control of walking in rugged terrain requires one to incorporate different issues, such as the mechanical properties of legs and muscles, the neuronal control structures for the single leg, the mechanics and neuronal control structures for the coordination between legs, as well as central decisions that are based on external information and on internal states. Walking in predictable environments and fast running, to a large degree, rely on muscle mechanics. Conversely, slow walking in unpredictable terrain, e.g. climbing in rugged structures, has to rely on neuronal systems that monitor and intelligently react to specific properties of the environment. An arthropod model system that shows the latter abilities is the stick insect, based on which this review will be focused. An insect, when moving its six legs, has to control 18 joints, three per leg, and therefore has to control 18 degrees of freedom (d.f.). As the body position in space is determined by 6 d.f. only, there are 12 d.f. open to be selected. Therefore, a fundamental problem is as to how these extra d.f. are controlled. Based mainly on behavioural experiments and simulation studies, but also including neurophysiological results, the following control structures have been revealed. Legs act as basically independent systems. The quasi-rhythmic movement of the individual leg can be described to result from a structure that exploits mechanical coupling of the legs via the ground and the body. Furthermore, neuronally mediated influences act locally between neighbouring legs, leading to the emergence of insect-type gaits. The underlying controller can be described as a free gait controller. Cooperation of the legs being in stance mode is assumed to be based on mechanical coupling plus local positive feedback controllers. These controllers, acting on individual leg joints, transform a passive displacement of a joint into an active movement, generating synergistic assistance reflexes in all mechanically coupled joints. This architecture is summarized in the form of the artificial neural network, Walknet, that is heavily dependent on sensory feedback at the proprioceptive level. Exteroceptive feedback is exploited for global decisions, such as the walking direction and velocity.

Assessing sensory function in locomotor systems using neuro-mechanical simulations

Computer simulations are being used increasingly to gain an understanding of the complex interactions between the neuronal, sensory, muscular and mechanical components of locomotor systems. Recent neuro-mechanical simulations of walking in humans, cats and insects, and of swimming in lampreys, have provided new information on the functional role of specific groups of sensory receptors in regulating locomotion. As we discuss in this review, these studies also make it clear that a full understanding of the neural and mechanical mechanisms that underlie locomotion can be achieved only by using simulations in parallel with physiological investigations. The widespread implementation of this approach would be enhanced by the development of freely available and easy-to-use software tools.

Perturbation of leg protraction causes context-dependent modulation of inter-leg coordination, but not of avoidance reflexes

All animals capable of legged locomotion execute fast, adaptive compensatory movements in response to perturbation of a step cycle. In terms of motor control, such adaptive behaviour typically involves changes in the kinematics of the perturbed limb as well as changes in coordination between legs. Moreover, the unpredictable variety of real life situations implies that compensatory responses should be sensitive to the behavioural context of the animal. We have investigated the extent to which the compensatory response of a walking stick insect (Carausius morosus) adapts in parallel to strong context-dependent adaptation of step kinematics and inter-leg coordination. The behavioural contexts we chose were straight walking and visually induced curve walking, for both of which the steady state limb kinematics and inter-leg coupling strengths were known. In case of curve walking, we further distinguished contexts according to whether the inner or the outer leg was perturbed. The three contexts differed strongly with respect to the set of joint actions before perturbation. Upon mechanical perturbation of front leg protraction, we studied context-dependent differences in a local avoidance reflex of the perturbed leg, as well as in coordination mechanisms that couple the step cycles of the perturbed leg to its unperturbed neighbours. In all three walking contexts, obstacle contact caused an avoidance movement of the front leg that deviated from the unperturbed swing trajectory. Swing duration was increased while step distance was decreased; however, both effects vanished in the subsequent unperturbed step. The prevailing immediate reaction of the three leg joints were retraction of the coxa (>76%), levation of the femur (>80%), and flexion of the tibia (>80%), regardless of the behavioural context and, therefore, joint action prior to perturbation. Moreover, activation of each one of these joint actions was shown to be independent of the other two. Thus, local avoidance reflexes are not modulated by the descending visual information that causes transition from straight to curve walking, but are composed of context-independent joint actions. Perturbation of the front leg also caused significant shifts of the touch-down position of the perturbed leg and of its unperturbed neighbours. If the inner front leg was perturbed, this shift could persist until the subsequent step. Perturbation affected both the spatial location and the timing of touch-down and lift-off transitions in unperturbed neighbouring legs. These effects on inter-leg coordination were context-dependent. For example, time delay to lift-off of the contralateral neighbour was shortened in inner and straight walking legs, but not in outer legs. Finally, a targeting mechanism that determines foot placement in stick insects was shown to be affected by perturbation in a context-dependent manner. We conclude that the immediate compensatory response of the perturbed leg is not adapted to the behavioural context in spite of strongly differing step kinematics, whereas the compensatory effect on inter-limb coupling is context-dependent.

Behaviour-based modelling of hexapod locomotion: linking biology and technical application

Walking in insects and most six-legged robots requires simultaneous control of up to 18 joints. Moreover, the number of joints that are mechanically coupled via body and ground varies from one moment to the next, and external conditions such as friction, compliance and slope of the substrate are often unpredictable. Thus, walking behaviour requires adaptive, context-dependent control of many degrees of freedom. As a consequence, modelling legged locomotion addresses many aspects of any motor behaviour in general. Based on results from behavioural experiments on arthropods, we describe a kinematic model of hexapod walking: the distributed artificial neural network controller walknet. Conceptually, the model addresses three basic problems in legged locomotion. (I) First, coordination of several legs requires coupling between the step cycles of adjacent legs, optimising synergistic propulsion, but ensuring stability through flexible adjustment to external disturbances. A set of behaviourally derived leg coordination rules can account for decentralised generation of different gaits, and allows stable walking of the insect model as well as of a number of legged robots. (II) Second, a wide range of different leg movements must be possible, e.g. to search for foothold, grasp for objects or groom the body surface. We present a simple neural network controller that can simulate targeted swing trajectories, obstacle avoidance reflexes and cyclic searching-movements. (III) Third, control of mechanically coupled joints of the legs in stance is achieved by exploiting the physical interactions between body, legs and substrate. A local positive displacement feedback, acting on individual leg joints, transforms passive displacement of a joint into active movement, generating synergistic assistance reflexes in all mechanically coupled joints.

Signals From Load Sensors Underlie Interjoint Coordination During Stepping Movements of the Stick Insect Leg

During stance and swing phase of a walking stick insect, the retractor coxae (RetCx) and protractor coxae (ProCx) motoneurons and muscles supplying the thorax-coxa (TC)-joint generate backward and forward movements of the leg. Their activity is tightly coupled to the movement of the more distal leg segments, i.e., femur, tibia, and tarsus. We used the single middle leg preparation to study how this coupling is generated. With only the distal leg segments of the middle leg being free to move, motoneuronal activity of the de-afferented and -efferented TC-joint is similarly coupled to leg stepping. RetCx motoneurons are active during stance and ProCx motoneurons during swing. We studied whether sensory signals are involved in this coordination of TC-joint motoneuronal activity. Ablation of the load measuring campaniform sensilla (CS) revealed that they substantially contribute to the coupling of TC-joint motoneuronal activity to leg stepping. Individually ablating trochanteral and femoral CS revealed the trochanteral CS to be necessary for establishing the coupling between leg stepping and coxal motoneuron activity. When the locomotor system was active and generated alternating bursts of activity in ProCx and RetCx motoneurons, stimulation of the CS by rearward bending of the femur in otherwise de-afferented mesothoracic ganglion terminated ongoing ProCx motoneuronal activity and initiated RetCx motoneuronal activity. We show that cuticular strain signals from the trochanteral CS play a major role in shaping TC-joint motoneuronal activity during walking and contribute to their coordination with the stepping pattern of the distal leg joints. We present a model for the sensory control of timing of motoneuronal activity in walking movements of the single middle leg.

Insect walking and robotics

With the advent of significant collaborations between researchers who study insect walking and robotics engineers interested in constructing adaptive legged robots, insect walking is once again poised to make a more significant scientific contribution than the numbers of participants in the field might suggest. This review outlines current knowledge of the physiological basis of insect walking with an emphasis on recent new developments in biomechanics and genetic dissection of behavior, and the impact this knowledge is having on robotics. Engineers have begun to team with neurobiologists to build walking robots whose physical design and functional control are based on insect biology. Such an approach may have benefits for engineering, by leading to the construction of better-performing robots, and for biology, by allowing real-time and real-world tests of critical hypotheses about how locomotor control is effected. It is argued that in order for the new field of biorobotics to have significant influence it must adopt criteria for performance and an experimental approach to the development of walking robots.

Statocysts in crabs: short-term control of locomotion and long-term monitoring of hydrostatic pressure

Crabs show well-coordinated locomotion. They have proprioceptors similar to those of lobsters, but they differ in terms of their balancing systems and their condensed nervous system, which allows rapid interganglionic conduction. Typically they exhibit dynamically stable locomotion with a highly developed semicircular canal system that codes angular acceleration in each of three orthogonal planes (horizontal and vertical at 45 degrees and 135 degrees to the pitching plane). Left and right interneurons each code one direction of angular acceleration, carrying information between the brain and the thoracic ganglia. Cell A codes head-up vertical plane angular accelerations. Cell B codes rotations in the horizontal plane. Interneurons C and D code headdown vertical plane information, carrying it ipsilaterally and contralaterally respectively. These interneurons have a central role in locomotion. They are activated and have their responsiveness to angular acceleration enhanced before and during locomotion. Such simple activation pathways point to how an angular-acceleration-controlled robot (CRABOT) could be constructed. Hydrostatic pressure information carried by the thread hairs, which also sense angular acceleration, is filtered out from direct pathways onto the interneurons, but spectral analysis shows that it still has an influence via central pathways. Long-term recordings from equilibrium interneurons in free-walking crabs taken from the wild into constant conditions show tidally changing frequencies