Maximum walking speed and lower limb length in hominids

In 1984, Helene (Am. J. Physics 52:656) and Alexander (Am. Scientist 72:348–354) presented equations which purported to explain how lower limb length limited maximum walking speed in humans. The equations were based on a simplified model of human walking in which the center of mass (CoM) “vaults” over the supporting leg. Increasing walking speed by increasing stride frequency or stride length would increase the upward acceleration of the CoM in the first half of stance phase, to the point that it would be greater than the downward pull of gravity, and the individual would become airborne. This constitutes running by most definitions. While these models ignored various mechanical factors, such as knee flexion during midstance, that reduce the vertical movement of the CoM, the general idea is plausible inasmuch as the CoM of the body does oscillate vertically with each step. One hypothesis tested here is whether it is indeed the interaction between the pull of gravity and the individual's own upward acceleration that determines at what speed (or cadence) he changes from walking to running. Another hypothesis considered is that increased lower limb length (L) was selected for in early hominids, because of the locomotor advantages of longer lower limbs. Results indicate, however, that while L was clearly related to maximum possible walking speed, it was not an important factor in determining maximum “comfortable” walking speed. These and other results from the recent literature suggest that increased lower limb length provided no selective advantage in locomotion, and other explanations should be sought. © 1996 Wiley-Liss, Inc.     

Walking is not like reaching: evidence from periodic mechanical perturbations

The control architecture underlying human reaching has been established, at least in broad outline. However, despite extensive research, the control architecture underlying human locomotion remains unclear. Some studies show evidence of high-level control focused on lower-limb trajectories; others suggest that nonlinear oscillators such as lower-level rhythmic central pattern generators (CPGs) play a significant role. To resolve this ambiguity, we reasoned that if a nonlinear oscillator contributes to locomotor control, human walking should exhibit dynamic entrainment to periodic mechanical perturbation; entrainment is a distinctive behavior of nonlinear oscillators. Here we present the first behavioral evidence that nonlinear neuro-mechanical oscillators contribute to the production of human walking, albeit weakly. As unimpaired human subjects walked at constant speed, we applied periodic torque pulses to the ankle at periods different from their preferred cadence. The gait period of 18 out of 19 subjects entrained to this mechanical perturbation, converging to match that of the perturbation. Significantly, entrainment occurred only if the perturbation period was close to subjects' preferred walking cadence: it exhibited a narrow basin of entrainment. Further, regardless of the phase within the walking cycle at which perturbation was initiated, subjects' gait synchronized or phase-locked with the mechanical perturbation at a phase of gait where it assisted propulsion. These results were affected neither by auditory feedback nor by a distractor task. However, the convergence to phase-locking was slow. These characteristics indicate that nonlinear neuro-mechanical oscillators make at most a modest contribution to human walking. Our results suggest that human locomotor control is not organized as in reaching to meet a predominantly kinematic specification, but is hierarchically organized with a semi-autonomous peripheral oscillator operating under episodic supervisory control.     

Resonant frequencies of arms and legs identify different walking patterns

The present study is aimed at investigating changes in the coordination of arm and leg movements in young healthy subjects. It was hypothesized that with changes in walking velocity there is a change in frequency and phase coupling between the arms and the legs. In addition, it was hypothesized that the preferred frequencies of the different coordination patterns can be predicted on the basis of the resonant frequencies of arms and legs with a simple pendulum model. The kinematics of arms and legs during treadmill walking in seven healthy subjects were recorded with accelerometers in the sagittal plane at a wide range of different velocities (i.e., 0.3-1. 3m/s). Power spectral analyses revealed a statistically significant change in the frequency relation between arms and legs, i.e., within the velocity range 0.3-0.7m/s arm movement frequencies were dominantly synchronized with the step frequency, whereas from 0.8m/s onwards arm frequencies were locked onto stride frequency. Significant effects of walking speed on mean relative phase between leg and arm movements were found. All limb pairs showed a significantly more stable coordination pattern from 0.8 to 1.0m/s onwards. Results from the pendulum modelling demonstrated that for most subjects at low-velocity preferred movement frequencies of the arms are predicted by the resonant frequencies of individual arms (about 0.98Hz), whereas at higher velocities these are predicted on the basis of the resonant frequencies of the individual legs (about 0.85Hz). The results support the above-mentioned hypotheses, and suggest that different patterns of coordination, as shown by changes in frequency coupling and phase relations, can exist within the human walking mode.     

Running perturbations reveal general strategies for step frequency selection

Recent research has suggested that energy minimization in human walking involves both a fast preprogrammed process and a slow optimization process. Here, we studied human running to test whether these two processes represent control mechanisms specific to walking or a more general strategy for minimizing energetic cost in human locomotion. To accomplish this, we used free response experiments to enforce step frequency with a metronome at values above and below preferred step frequency and then determined the response times for the return to preferred steady-state step frequency when the auditory constraint was suddenly removed. In forced response experiments, we applied rapid changes in treadmill speed and examined response times for the processes involved in the consequent adjustments to step frequency. We then compared the dynamics of step frequency adjustments resulting from the two different perturbations to each other and to previous results found in walking. Despite the distinct perturbations applied in the two experiments, both responses were dominated by a fast process with a response time of 1.47 ± 0.05 s with fine-tuning provided by a slow process with a response time of 34.33 ± 0.50 s. The dynamics of the processes underlying step frequency adjustments in running match those found previously in walking, both in magnitude and relative importance. Our results suggest that the underlying mechanisms are fundamental strategies for minimizing energetic cost in human locomotion.     

单侧前庭神经切断后猫运动方式的改变和补偿

本工作首次将一种运动分析系统,即视觉自动电视摄相信息处理系统（ｏｐｔｉｃａｌａｕｔｏｍａｔｉｃＴＶ－ｉｍａｇｅｐｒｏｃｅｓｓｏｒ,Ｅ．Ｌ．Ｉ．Ｔ．Ｅｓｙｓｔｅｍ）用于动物运动方式的定量分析,比较了猫一侧前庭神经切断前后在跨越转动横梁测试中运动方式的变化。结果如下：在横梁静止条件下,猫述后行走的初始身高明显下降,步距变短,步 频降低,呈缓慢的蹒中山步态,继而各参量逐渐增加,分别在术后４５和７０ｄ恢复     

Artificial neural network model for the generation of muscle activation patterns for human locomotion.

Skilled locomotor behaviour requires information from various levels within the central nervous system (CNS). Mathematical models have permitted researchers to simulate various mechanisms in order to understand the organization of the locomotor control system. While it is difficult to adequately characterize the numerous inputs to the locomotor control system, an alternative strategy may be to use a kinematic movement plan to represent the complex inputs to the locomotor control system based on the possibility that the CNS may plan movements at a kinematic level. We propose the use of artificial neural network (ANN) models to represent the transformation of a kinematic plan into the necessary motor patterns. Essentially, kinematic representation of the actual limb movement was used as the input to an ANN model which generated the EMG activity of 8 muscles of the lower limb and trunk. Data from a wide variety of gait conditions was necessary to develop a robust model that could accommodate various environmental conditions encountered during everyday activity. A total of 120 walking strides representing normal walking and ten conditions where the normal gait was modified in terms of cadence, stride length, stance width or required foot clearance. The final network was assessed on its ability to predict the EMG activity on individual walking trials as well as its ability to represent the general activation pattern of a particular gait condition. The predicted EMG patterns closely matched those recorded experimentally, exhibiting the appropriate magnitude and temporal phasing required for each modification. Only 2 of the 96 muscle/gait conditions had RMS errors above 0.10, only 5 muscle/gait conditions exhibited correlations below 0.80 (most were above 0.90) and only 25 muscle/gait conditions deviated outside the normal range of muscle activity for more than 25% of the gait cycle. These results indicate the ability of single network ANNs to represent the transformation between a kinematic movement plan and the necessary muscle activations for normal steady state locomotion but they were also able to generate muscle activation patterns for conditions requiring changes in walking speed, foot placement and foot clearance. The abilities of this type of network have implications towards both the fundamental understanding of the control of locomotion and practical realizations of artificial control systems for use in rehabilitation medicine.     

Sex difference in the pattern of lower limb movement during treadmill walking

To evaluate the characteristics of stereo-typed movement of the lower limb during treadmill walking, the step length and duration of 200 steps were monitored consecutively and calculated by means of a computerized system, consisting of a position sensor, shoes with foot switches and a minicomputer. Eleven male and 10 female subjects walked at various constant speeds ranging from 60-130 m.min-1. Mean, standard deviation (SD) and coefficient of variation (CV) of the time-distance component at each speed were utilized for the assessment of stereotyped movement. When compared with males, females had a tendency to increase their speed by increasing their cadence. The difference of the walking pattern was specifically related to their height. The SD and CV of the time-distance component at a given speed were significantly greater in females than in males. Regression analyses revealed that in the relationship between the walking speeds and the SDs or CVs of the time-distance component, the significant quadratic equations could be fitted. The speed, at which the SD of step length was minimum, was estimated to be about 90 m.min-1 in both males and females. This was regarded as the free walking speed or as the walking speed resulting from a mechanically efficient step length which suited the subject's body size.     

Voluntary bipedal walking of infant chimpanzees

The voluntary bipedal walking of infant chimpanzees was studied by the analysis of foot force and by motion analysis. The infants were trained to locomote on a level platform without any restrictions on the locomotor pattern. The voluntary bipedal walking was compared with the other types of locomotion at the same age and with the trained bipedal walking performed by other chimpanzees, including adult chimpanzees. The characteristics of voluntary bipedal walking in the infant until one year of age were: (1) high-speed walking with short cycle duration; (2) short stance phase duration; (3) small braking component of the preceding leg and large acceleration of the following leg; (4) one downward peak in the vertical component; and (5) a relatively small transverse component. Bipedal walking usually continued for less than one second and ended in quadrupedal locomotion. During walking, the preceding foot touched the floor, heel first, as in the case of older chimpanzees and humans. At this age, bipedal walking was similar to high-speed locomotion. The voluntary bipedal walking of the two-year-old and frour-yearold chimpanzees was characterized as follows: (1) slower speed than during quadrupedal locomotion, (2) relatively long periods and distances; (3) well balanced accelerating and braking components; and (4) a vertical component showing two downward peaks and a trough in between during numerous trials. The last characteristic means that the body center of gravity is higher in the single stance phase, just as in the bipedal walkinbg of the adult chimpanzees and humans. The bipedal walking of infant chimpanzees was discussed in comparison with the walking of humans, including infants.     

Gait‐specific metabolic costs and preferred speeds in ring‐tailed lemurs (Lemur catta), with implications for the scaling of locomotor costs

Metabolic costs of resting and locomotion have been used to gain novel insights into the behavioral ecology and evolution of a wide range of primates; however, most previous studies have not considered gait‐specific effects. Here, metabolic costs of ring‐tailed lemurs (Lemur catta) walking, cantering and galloping are used to test for gait‐specific effects and a potential correspondence between costs and preferred speeds. Metabolic costs, including the net cost of locomotion (COL) and net cost of transport (COT), change as a curvilinear function of walking speed and (at least provisionally) as a linear function of cantering and galloping speeds. The baseline quantity used to calculate net costs had a significant effect on the magnitude of speed‐specific estimates of COL and COT, especially for walking. This is because non‐locomotor metabolism constitutes a substantial fraction (41–61%, on average) of gross metabolic rate at slow speeds. The slope‐based estimate of the COT was 5.26 J kg^?1 m^?1 for all gaits and speeds, while the gait‐specific estimates differed between walking (0.5 m s^?1: 6.69 J kg^?1 m^?1) and cantering/galloping (2.0 m s^?1: 5.61 J kg^?1 m^?1). During laboratory‐based overground locomotion, ring‐tailed lemurs preferred to walk at ～0.5 m s^?1 and canter/gallop at ～2.0 m s^?1, with the preferred walking speed corresponding well to the COT minima. Compared with birds and other mammals, ring‐tailed lemurs are relatively economical in walking, cantering, and galloping. These results support the view that energetic optima are an important movement criterion for locomotion in ring‐tailed lemurs, and other terrestrial animals. Am J Phys Anthropol, 2012. © 2012 Wiley Periodicals, Inc.     

Frequency tuning in animal locomotion

In locomotion that involves repetitive motion of propulsive structures (arms, legs, fins, wings) there are resonant frequencies f(*) at which the energy consumption is a minimum. As animals need to change their speed, they can maintain this energy minimum by tuning their body resonances. We discuss the physical principles of frequency tuning, and how it relates to forces, damping, and oscillation amplitude. The resonant frequency of pendulum-type oscillators (e.g. swinging arms and legs) may be changed by varying the mass moment of inertia, or the vertical acceleration of the pendulum pivot. The frequency of elastic vibrations (e.g. the bell of a jellyfish) can be tuned with a non-linear modulus of elasticity: soft for low deflection amplitudes (low resonant frequency), and stiff for large displacements (high resonant frequency). Tuning of elastic oscillations can also be achieved by changing the effective length or cross-sectional area of the elastic members, or by allowing springs in parallel or in series to become active. We propose that swimming and flying animals generate oscillating propulsive forces from precisely placed shed vortices and that these tuned motions can only occur when vortex shedding and the simple harmonic motion of the elastic elements of the propulsive structures are in resonance.     

Coupling of cardiac and locomotor rhythms

The pressure within exercising skeletal muscle rises and falls rhythmically during normal human locomotion, the peak pressure reaching levels that intermittently impede blood flow to the exercising muscle. Speculating that a reciprocal relationship between the timing of peak intramuscular and pulsatile arterial pressures should optimize blood flow through muscle and minimize cardiac load, we tested the hypothesis that heart rate becomes entrained with walking and running cadence at some locomotion speeds, by means of electrocardiography and an accelerometer to provide signals reflecting heart rate and cadence, respectively. In 18 of 25 subjects, 1:1 coupling of heart and step rates was present at one or more speeds on a motorized treadmill, generally at moderate to high exercise intensities. To determine how exercise specific this phenomenon is, and to refute the competing hypothesis that coupling is due to vertical accelerations of the heart during locomotion, we had 12 other subjects cycle on an electronically braked bicycle ergometer. Coupling was found between heart rate and pedaling frequency in 10 of them. Cardiac-locomotor coupling appears to be a normal physiological phenomenon, and its identification provides a fresh perspective from which to study endurance.     

Biomechanics and allometric scaling in primate locomotion and morphology

Body size has a dominant influence on locomotor performance and the morphology of the locomotor apparatus. In locomotion under the influence of gravity, body mass acts as weight force and is a mechanical variable. Accordingly, the application of biomechanical principles and methods allows a functional understanding of scaling effects in locomotion. This is demonstrated here using leaping primates as an example. With increasing body size, the decreasing ratio of muscle force available for acceleration during takeoff to the body mass that has to be accelerated dictates both the movement pattern and the proportions of the hindlimbs. In an arm-swinging movement, the long, heavy arms of the large-bodied leapers are effectively used to gain additional momentum. A new perspective on decreasing size identifies the absolutely small acceleration distance and time available for propulsion as factors limiting leaping distance and extensively determining locomotor behavior and body proportions. As the mechanical constraints differ according to body size for a given mode of locomotion, a typological approach to morphology in relation to locomotor category is ruled out. Across locomotor categories, dynamic similarity (sensu Alexander) can be expected if the propulsive mechanisms as well as the selective pressures acting upon locomotion are the same.     

A dynamical model for reflex activated head movements in the horizontal plane

 We present a controls systems model of horizontal-plane head movements during perturbations of the trunk, which for the first time interfaces a model of the human head with neural feedback controllers representing the vestibulocollic (VCR) and the cervicocollic (CCR) reflexes. This model is homeomorphic such that model structure and parameters are drawn directly from anthropomorphic, biomechanical and physiological studies. Using control theory we analyzed the system model in the time and frequency domains, simulating neck movement responses to input perturbations of the trunk. Without reflex control, the head and neck system produced a second-order underdamped response with a 5.2 dB resonant peak at 2.1 Hz. Adding the CCR component to the system dampened the response by approximately 7%. Adding the VCR component dampened head oscillations by 75%. The VCR also improved low-frequency compensation by increasing the gain and phase lag, creating a phase minimum at 0.1 Hz and a phase peak at 1.1 Hz. Combining all three components (mechanics, VCR and CCR) linearly in the head and neck system reduced the amplitude of the resonant peak to 1.1 dB and increased the resonant frequency to 2.9 Hz. The closed loop results closely fit human data, and explain quantitatively the characteristic phase peak often observed. Received: 15 April 1996 / Accepted in revised form: 1 July 1996     

Walking at the preferred stride frequency maximizes local dynamic stability of knee motion

Healthy humans display a preference for walking at a stride frequency dependent on the inertial properties of their legs. Walking at preferred stride frequency (PSF) is predicted to maximize local dynamic stability, whereby sensitivity to intrinsic perturbations arising from natural variability inherent in biological motion is minimized. Previous studies testing this prediction have employed different variability measures, but none have directly quantified local dynamic stability by computing maximum finite-time Lyapunov exponent (λ^Max), which quantifies the rate of divergence of nearby trajectories in state space. Here, ten healthy adults walked 45 m overground while sagittal motion of both knees was recorded via electrogoniometers. An auditory metronome prescribed 7 different frequencies relative to each individual's PSF (PSF; ±5, ±10, ±15 strides/min). Stride frequencies were performed under both freely adopted speed (FS) and controlled speed (CS: set at the speed of PSF trials) conditions. Local dynamic stability was maximal (λ^Max was minimal) at the PSF, becoming less stable for higher and lower stride frequencies. This occurred under both FS and CS conditions, although controlling speed further reduced local dynamic stability at non-preferred stride frequencies. In contrast, measures of variability revealed effects of stride frequency and speed conditions that were distinct from λ^Max. In particular, movement regularity computed by approximate entropy (ApEn) increased for slower walking speeds, appearing to depend on speed rather than stride frequency. The cadence freely adopted by humans has the benefit of maximizing local dynamic stability, which can be interpreted as humans tuning to their resonant frequency of walking.     

Predictive Simulation Generates Human Adaptations during Loaded and Inclined Walking

Predictive simulation is a powerful approach for analyzing human locomotion. Unlike techniques that track experimental data, predictive simulations synthesize gaits by minimizing a high-level objective such as metabolic energy expenditure while satisfying task requirements like achieving a target velocity. The fidelity of predictive gait simulations has only been systematically evaluated for locomotion data on flat ground. In this study, we construct a predictive simulation framework based on energy minimization and use it to generate normal walking, along with walking with a range of carried loads and up a range of inclines. The simulation is muscle-driven and includes controllers based on muscle force and stretch reflexes and contact state of the legs. We demonstrate how human-like locomotor strategies emerge from adapting the model to a range of environmental changes. Our simulation dynamics not only show good agreement with experimental data for normal walking on flat ground (92% of joint angle trajectories and 78% of joint torque trajectories lie within 1 standard deviation of experimental data), but also reproduce many of the salient changes in joint angles, joint moments, muscle coordination, and metabolic energy expenditure observed in experimental studies of loaded and inclined walking.     

Comparative efficiency of different regimens of locomotor training in prolonged space flights as estimated from the data on biomechanical and electromyographic parameters of walking

Biomechanical and electromyographic characteristics of locomotion were studied before and after a space flight on days 3, 7, and 10 after landing in 18 participants of prolonged space missions on board the International Space Station. It has been shown that microgravity causes significant changes in biomechanical and electromyographic characteristics of walking, such as a decrease in the amplitude of angular displacement in leg joints, a decrease in the double step length, and an increase in the electromyographic costs of locomotion. It has been also shown that interval locomotor physical training, such as alternation of running and walking, in prolonged space flights prevents an increase in the physiological costs of locomototions after a space flight and provides more efficient maintenance of the neuromuscular system’s performance after a flight. Cosmonauts who performed interval locomotor training had fewer changes in biomechanical and electromyographic characteristics of walking.     

Gait patterns of Japanese pedestrians

Gait patterns of adult Japanese were observed while walking on a sidewalk at Fukuoka city under natural conditions on clear days during December 1987. Pedestrians were recorded by means of video recorder for eleven 1-hr observations. From the measurements of the number of steps and the time required for each subject to traverse 50 m, walking speed was calculated for 479 women and 139 men, and step length and cadence for 292 women and 117 men. The following factors were taken into account in the data analysis: footwear, accompanist, time of day, estimated age and sex. Normal values for walking speed, step length, and cadence information on present-day adult Japanese were documented. The range for the total sample was from 49 to 119 m/min in walking speed, from 51 to 91 cm in step length, and from 95 to 160 steps/min in cadence. There were no significant differences in walking speed, step length, and cadence of the women between wearing flat shoes and heeled shoes. Walking velocity and step length were greater in single walking than in group walking. As compared with the values in the daytime (10:00-15:30) velocity, step length, and cadence tended to be higher in the morning (7:30-10:00) and lower in the evening (15:30-18:00). Older people walked at a slower speed with shorter steps than younger people. For younger pedestrians no significant sex difference was observed in walking speed. Women required a higher cadence than men because of their shorter step lengths. Older men walked faster than older women.     

Vertical displacement of the centre of mass during walking in people with diabetes and diabetic neuropathy does not explain their higher metabolic cost of walking

People with diabetes display biomechanical gait alterations compared to controls and have a higher metabolic cost of walking (CoW), but it remains unknown whether differences in the vertical displacement of the body centre of mass (CoM) may play a role in this higher CoW. The aim of this study was to investigate vertical CoM displacement (and step length as a potential underpinning factor) as an explanatory factor in the previously observed increased CoW with diabetes. Thirty-one non-diabetic controls (Ctrl); 22 diabetic patients without peripheral neuropathy (DM) and 14 patients with moderate/severe Diabetic Peripheral Neuropathy (DPN), underwent gait analysis using a motion analysis system and force plates while walking at a range of matched speeds between 0.6 and 1.6 m/s. Vertical displacement of the CoM was measured over the gait cycle, and was not different in either diabetes patients with or without diabetic peripheral neuropathy compared to controls across the range of matched walking speeds examined (at 1 m/s: Ctrl: 5.59 (SD: 1.6), DM: 5.41 (1.63), DPN: 4.91 (1.66) cm; p > 0.05). The DPN group displayed significantly shorter steps (at 1 m/s: Ctrl: 69, DM: 67, DPN: 64 cm; p > 0.05) and higher cadence (at 1 m/s: Ctrl: 117 (SD1.12), DM: 119 (1.08), DPN: 122 (1.25) steps per minute; p > 0.05) across all walking speeds compared to controls. The vertical CoM displacement is therefore unlikely to be a factor in itself that contributes towards the higher CoW observed recently in people with diabetic neuropathy. The higher CoW in patients with diabetes may not be explained by the CoM displacement, but rather may be more related to shorter step lengths, increased cadence and the associated increased internal work and higher muscle forces developed by walking with more flexed joints.     

Posture, gait and the ecological relevance of locomotor costs and energy-saving mechanisms in tetrapods

A reanalysis of locomotor data from functional, energetic, mechanical and ecological perspectives reveals that limb posture has major effects on limb biomechanics, energy-saving mechanisms and the costs of locomotion. Regressions of data coded by posture (crouched vs. erect) reveal nonlinear patterns in metabolic cost, limb muscle mass, effective mechanical advantage, and stride characteristics. In small crouched animals energy savings from spring and pendular mechanisms are inconsequential and thus the metabolic cost of locomotion is driven by muscle activation costs. Stride frequency appears to be the principal functional parameter related to the decreasing cost of locomotion in crouched animals. By contrast, the shift to erect limb postures invoked a series of correlated effects on the metabolic cost of locomotion: effective mechanical advantage increases, relative muscle masses decrease, metapodial limb segments elongate dramatically (as limbs shift from digitigrade to unguligrade designs) and biological springs increase in size and effectiveness. Each of these factors leads to decreases in the metabolic cost of locomotion in erect forms resulting from real and increasing contributions of pendular savings and spring savings. Comparisons of the relative costs and ecological relevance of different gaits reveal that running is cheaper than walking in smaller animals up to the size of dogs but running is more expensive than walking in horses. Animals do not necessarily use their cheapest gaits for their predominant locomotor activity. Therefore, locomotor costs are driven more by ecological relevance than by the need to optimize locomotor economy.     

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.