Human Running Research Articles

Comparative study of wireless sensors for measuring the energy consumption of human running

With the rapid development of technologies such as communications, embedded computing, and sensors, wireless sensor network technology has become more mature and its applications have become more and more extensive. The development of technology has brought great convenience to people's lives and work. At the same time, people's activity and exercise are greatly reduced, causing health problems such as physical weakness. In this regard, designing a research based on wireless sensors to measure human walking and running is of great significance to promote the development of people's physical health. This paper presents a design scheme for human exercise energy consumption detection, researches and designs the coordinator, router node and host computer monitoring system of wireless sensor network. In the human motion energy consumption detection system implemented in this paper, the communication effect between the coordinator and the router node is good, and there are fewer problems such as node loss and data loss. The host computer monitoring interface is simple and easy to operate, easy to use, and can run stably. This solution does not need wiring, reduces the point of failure, reduces the cost, and has good scalability, laying a foundation for the further development of the human motion detection system, and has a wide range of application prospects and practical values. This paper is based on two constant beam width algorithms to carry out the experimental research on the energy consumption of human walking and running, and analyze the results of the energy consumption experiment. Experimental results show that the two constant beamwidth beamforming algorithms have relatively good constant beamwidth effect, which basically verifies the correctness of the theoretical analysis results and can be applied to target noise measurement.

Plasticity of muscle synergies through fractionation and merging during development and training of human runners

Complex motor commands for human locomotion are generated through the combination of motor modules representable as muscle synergies. Recent data have argued that muscle synergies are inborn or determined early in life, but development of the neuro-musculoskeletal system and acquisition of new skills may demand fine-tuning or reshaping of the early synergies. We seek to understand how locomotor synergies change during development and training by studying the synergies for running in preschoolers and diverse adults from sedentary subjects to elite marathoners, totaling 63 subjects assessed over 100 sessions. During development, synergies are fractionated into units with fewer muscles. As adults train to run, specific synergies coalesce to become merged synergies. Presences of specific synergy-merging patterns correlate with enhanced or reduced running efficiency. Fractionation and merging of muscle synergies may be a mechanism for modifying early motor modules (Nature) to accommodate the changing limb biomechanics and influences from sensorimotor training (Nurture).

Synchronicity Rectangle for temporal gait analysis: Application to Parkinson’s Disease

Human walking and running gaits are represented by the same temporal model. The golden ratio ϕ - namely, the largest solution to the equation x2=1+x - is associated with all of the four eigenvalues of such a model. An eigenspace of dimension two, named symmetry plane, is determined. It contains an eigenspace of dimension one, named golden line, on which the ratio between stance and swing durations is equal to ϕ for walking and to ϕ−1 for running. The Synchronicity Rectangle is defined, whose sides are: the normalized distance of the gait from the symmetry plane and the normalized distance from the golden line of the gait projection on the symmetry plane. It constitutes a new comprehensive geometric representation of walking and running gaits. The area of such a rectangle computed for an averaged gait is illustratively shown to discriminate between patients affected by Parkinson’s Disease and healthy subjects: its mean value is significantly higher for patients affected by Parkinson’s Disease; high Sensitivity and moderate Specificity, along with relatively large likelihood and odds ratios, are obtained.

On the mechanical power output required for human running – Insight from an analytical model

In this paper the dynamics of human running on flat terrain and the required mechanical power output with its dependency on various parameters is investigated. Knowing the required mechanical power output is of relevance due to its relationship with the metabolic power. For example, a better understanding of the dependencies of required mechanical power output on weight, running and wind speed, step frequency, ground contact time etc. is very valuable for the assessment, analysis and optimization of running performance. Therefore, a mathematical model based on very few assumptions is devised. The purpose of the proposed model is to relate running speed and required mechanical power output as an algebraic function of the runner’s mass, height, step rate, ground contact time and wind speed. This is relevant in order to better understand the mechanical energy cost of locomotion, and how much it depends on which parameters. The first of the main energy dissipation mechanisms is due to vertical oscillation, i.e., during each step some of the potential energy difference gets transformed into heat. The second mechanism is due to the anterior ground reaction force during the first part of stance and the third is due to aerodynamic drag. With the approximations of constant running speed and a sinusoidal vertical ground reaction force profile one obtains closed algebraic expressions for the center of mass trajectory and the required mechanical power output. Comparisons of model predictions and reported performance data suggest that approximately a quarter of the ground impact energy is stored during the first part of ground contact and then released during the remaining stance phase. Further, one can conclude from the model that less mechanical power output is required when running with higher step rates and a higher center of mass. Non intuitive is the result that a shorter ground contact time is beneficial for fast runs, while the opposite holds for slow runs. An important advantage of the devised model compared to others is that it leads to closed algebraic expressions for the center of mass trajectory and mechanical power output, which are functions of measurable quantities, i.e., of step rate, ground contact time, running speed, runner’s mass, center of mass height, aerodynamic drag at some given speed, wind speed and heart rate. Moreover, the model relies on very few assumptions, which have been verified, and the only tuning parameter is the ratio of recovered elastic energy.

Are humans evolved specialists for running in the heat? Man vs. horse races provide empirical insights.

What is the central question of this study? Do available comparative data provide empirical evidence that humans are adapted to endurance running at high ambient temperatures? What is the main finding and its importance? Comparing the results of races that pit man against horse, we find that ambient temperature on race day has less deleterious effects on running speed in humans than it does on their quadrupedal adversary. This is evidence that humans are adapted for endurance running at high ambient temperatures. We debate whether this supports the hypothesis that early man was evolutionarily adapted for persistence hunting. Many mammals run faster and for longer than humans and have superior cardiovascular physiologies. Yet humans are considered by some scholars to be excellent endurance runners at high ambient temperatures, and in our past to have been persistence hunters capable of running down fleeter quarry over extended periods during the heat of the day. This suggests that human endurance running is less affected by high ambient temperatures than is that of other cursorial ungulates. However, there are no investigations of this hypothesis. We took advantage of longitudinal race results available for three annual events that pit human athletes directly against a hyper-adapted ungulate racer, the thoroughbred horse. Regressing running speed against ambient temperature shows race speed deteriorating with hotter temperatures more slowly in humans than in horses. This is the first direct evidence that human running is less inhibited by high ambient temperatures than that of another endurance species, supporting the argument that we are indeed adapted for high temperature endurance running. Nonetheless, it is far from clear that this capacity is explained by an endurance hunting past because in absolute terms humans are slower than horses and indeed many other ungulate species. While some human populations have persistence hunted (and on occasion still do), the success of this unlikely foraging strategy may be best explained by the application of another adaption - high cognitive capacity. With dedication, experience and discipline, capitalising on their small endurance advantage in high temperatures, humans have a chance of running a more athletic prey to exhaustion.

Using the spring-mass model for running: Force-length curves and foot-strike patterns

BackgroundThe spring-mass model is commonly used to investigate the mechanical characteristics of human running. Underlying this model is the assumption of a linear force-length relationship, during the stance phase of running, and the idea that stiffness can be characterised using a single spring constant. However, it remains unclear whether the assumption of linearity is valid across different running styles. Research questionHow does the linearity of the force-length curve vary across a sample of runners and is there an association between force-length linearity and foot-strike index/speed? MethodsKinematic and kinetic data were collected from twenty-eight participants who ran overground at four speeds. The square of the Pearson’s correlation coefficient, R2, was used to quantify linearity; with a threshold of R2 ≥ 0.95 selected to define linear behaviour. A linear mixed model was used to investigate the association between linearity and foot-strike index and speed. ResultsOnly 36–46 % of participants demonstrated linear force-length behaviour across the four speeds during the loading phase. Importantly, the linear model showed a significant effect of both foot-strike index and speed on linearity during the loading phase (p = 0.003 and p < 0.001, respectively). SignificanceThis study showed that the assumption of a linear force-length relationship is not appropriate for all runners. These findings suggest that the use of the spring-mass model, and a constant value of stiffness, may not be appropriate for characterising and comparing different running styles. Given these findings, it may be better to restrict the use of the spring-mass model to individuals who exhibit linear force-length dependence. It would also be appropriate for future studies, characterising stiffness using the spring-mass model, to report data on force-length linearity across the cohort under study.

Design simulation of a novel fluid based footstep energy harvesting system

With the ever-increasing demand for energy and a growing concern for the non-renewability of the energy sources as well as the environmental pollution, renewable energy harvesting systems have emerged as one of the energy technologies globally. In this context, this research work addresses a novel, clean and environmentally friendly renewable energy generation technology that harvests energy from human locomotion, like walking and running. The developed system is based on two square sized tiles structure paver in which two fluid bags are connected through flow control mechanisms including unidirectional valves and mini hydro generators that convert the human kinetic energy into electricity via fluid movements. All the design related simulations were performed in ANSYS software to find out the optimal parameters like fluid velocity, power output and fluid bag shape etc. Sequentially, it was experimentally observed that the energy harvesting paver can produce an average output power of 1.4 W per step during typical human walking and can power up a DC load of 390 Ω LED. The system is easily reproducible and can be installed with relative ease to power up street lamps, billboards and emergency lighting systems etc. Furthermore, the system is eco-friendly and cost-effective in comparison to other available energy harvesting paver systems. Hence, this novel fluid based footstep energy harvesting paver has considerable prospects as an effective renewable energy system.

Monitoring of human body running training with wireless sensor based wearable devices

With the rapid pace of today’s society, work pressure, less exercise time, people began to pay more attention to their health. Walking and running have become the first choice of moderate exercise for many young people. The recognition of human running state based on wireless acceleration sensor will play an increasingly important role in the fields of motion detection, energy consumption evaluation and health care. It is of great significance to develop and design a kind of wearable multi-functional wireless sensor which can monitor the running state of human body. In this paper, a wearable human body monitoring system based on wireless acceleration sensor technology is proposed for real-time monitoring of daily running volume of human body. Hardware and upper computer design: stm32f405 is used as the main control chip, and ma8451q is used to collect human motion data. In this paper, aiming at the problem that three kinds of motion states of human body are easy to be confused and difficult to distinguish, based on the in-depth study of the complex structure mode and self similarity characteristics of non-stationary acceleration signal, a method of human body motion state recognition based on single fractal and multi fractal is proposed. In this method, the fractal dimension and the generalized dimension are used as the feature variables, and the correlation judgment method is used to distinguish and recognize different motion states. Experiments show the validity and feasibility of single fractal and multifractal in walking and going up and down three kinds of motion state recognition. On the basis of multifractal motion state recognition, this paper combines fractal theory with wavelet multiresolution analysis, and proposes a matrix fractal human motion state recognition method based on wavelet transform. The fractal matrix based on wavelet transform quantifies the fractal characteristics of the component signals of walking and going up and down in different wavelet scales, and then describes the complexity and self similarity of the original acceleration signals. Experimental results show that the average recognition rate of walking, jogging and fast running can reach over 93% under the premise of less prior information.

Contribution of foot joints in the energetics of human running

The foot seems to demonstrate considerable power absorption and generation characteristics during running. These have been mainly accounted to the mechanics of the ankle joint, however, evidence suggests that joint kinetics have been overestimated by single-segment foot models. The scope of the present study was to estimate the energetics of the ankle-, chopart-, lisfranc- and hallux joint during heel-strike running. Power absorption and generation occuring at different segments of the foot of seven asymptomatic adults was modelled using a four-segment kinetic foot model. Participants ran barefoot with an average running speed 3.5 m/s along a 10 meter walkway. The peak power generation of the ankle, chopart, lisfranc, and hallux joint reached respectively an average of 13.9, 4.12, 1.08 and 0.32 Watt/kg. The Lisfranc joint showed poor power absorption compared to the other three joints. It was further demonstrated that the Ankle and Chopart joints seem to have both receptive and propulsive characteristics. The behavior of the Lisfranc joint complied almost exclusively with propulsive characteristics. Finally, it can be concluded that the midfoot accounts for approximately 25% of the total power absorption occuring at the foot joints and not 50% as initially hypothesized.

Improving the energy economy of human running with powered and unpowered ankle exoskeleton assistance.

Exoskeletons that reduce energetic cost could make recreational running more enjoyable and improve running performance. Although there are many ways to assist runners, the best approaches remain unclear. In our study, we used a tethered ankle exoskeleton emulator to optimize both powered and spring-like exoskeleton characteristics while participants ran on a treadmill. We expected powered conditions to provide large improvements in energy economy and for spring-like patterns to provide smaller benefits achievable with simpler devices. We used human-in-the-loop optimization to attempt to identify the best exoskeleton characteristics for each device type and individual user, allowing for a well-controlled comparison. We found that optimized powered assistance improved energy economy by 24.7 ± 6.9% compared with zero torque and 14.6 ± 7.7% compared with running in normal shoes. Optimized powered torque patterns for individuals varied substantially, but all resulted in relatively high mechanical work input (0.36 ± 0.09 joule kilogram-1 per step) and late timing of peak torque (75.7 ± 5.0% stance). Unexpectedly, spring-like assistance was ineffective, improving energy economy by only 2.1 ± 2.4% compared with zero torque and increasing metabolic rate by 11.1 ± 2.8% compared with control shoes. The energy savings we observed imply that running velocity could be increased by as much as 10% with no added effort for the user and could influence the design of future products.

Low leg compliance permits grounded running at speeds where the inverted pendulum model gets airborne

Animals typically switch from grounded (no flight phases) to aerial running at dimensionless speeds u^ < 1. But some birds use grounded running far above u^ = 1, which puzzles biologists because the inverted pendulum becomes airborne at this speed. Here, we combine computer experiments using the spring-mass model with locomotion data from small birds, macaques and humans to understand the relationship between leg function (stiffness, angle of attack), locomotion speed and gait. With our model, we found three-humped ground reaction force profiles for slow grounded running speeds. The minimal single-humped grounded running speed is u^ = 0.4. This speed value roughly coincides with the transition speed from vaulting to bouncing mechanics in bipeds. Maximal grounded running speed in the model is not limited. In experiments, animals changed from grounded to aerial running at dimensionless contact time around 1. Considering these real-world contact times reduces the solution space drastically, but experimental data fit well. The model still predicts maximal grounded running speed u^ > 1 for low stiffness values used by birds but decreases below u^ = 1 for increasing stiffness. For stiffer legs used in human walking and running, periodic grounded running vanishes. At speeds at which birds and macaques change to aerial running, we found periodic aerial running to intersect grounded running. This could explain why animals can alternate between grounded and aerial running at the same speed and identical leg parameters. Compliant legs enable different gaits and speeds with similar leg parameters, stiff legs require parameter adaptations.

Trunk and leg kinematics of grounded and aerial running in bipedal macaques.

Across a wide range of Froude speeds, non-human primates such as macaques prefer to use grounded and aerial running when locomoting bipedally. Both gaits are characterized by bouncing kinetics of the center of mass. In contrast, a discontinuous change from pendular to bouncing kinetics occurs in human locomotion. To clarify the mechanism underlying these differences in bipedal gait mechanics between humans and non-human primates, we investigated the influence of gait on joint kinematics in the legs and trunk of three macaques crossing an experimental track. The coordination of movement was compared with observations available for primates. Compared with human running, macaque leg retraction cannot merely be produced by hip extension, but needs to be supported by substantial knee flexion. As a result, despite quasi-elastic whole-leg operation, the macaque's knee showed only minor rebound behavior. Ankle extension resembled that observed during human running. Unlike human running and independent of gait, torsion of the trunk represents a rather conservative feature in primates, and pelvic axial rotation added to step length. Pelvic lateral lean during grounded running by macaques (compliant leg) and human walking (stiff leg) depends on gait dynamics at the same Froude speed. The different coordination between the thorax and pelvis in the sagittal plane as compared with human runners indicates different bending modes of the spine. Morphological adaptations in non-human primates to quadrupedal locomotion may prevent human-like operation of the leg and limit exploitation of quasi-elastic leg operation despite running dynamics.

Control of Vibrations of Common Pedestrian Bridges in Jordan Using Tuned Mass Dampers

Due to their inherent slenderness and lightness, pedestrian bridges are typically subtle to human induced vibrations. This study was conducted to investigate the effect of human induced vibrations on common simply supported steel footbridges in Jordan, especially those with natural frequencies between 2 to 4 Hz. Such fundamental frequency is close to human movement frequency of walking, jumping and running. The ETABS software was utilized to develop a finite element model (FEM) in order to identify the footbridge dynamic properties including frequency, natural period, stiffness and the critical length at which the induced-vibrations become of great effect. The bridge was excited with multiple human walking and running loading scenarios. The response of the footbridge was evaluated and compared with and without integrating tuned mass dampers (TMD) tuned with the fundamental natural frequency of the footbridge in order to optimize appropriate mass, stiffness and damping coefficient. It was observed that after attaching the TMD, the fundamental vibration frequency decreased to stable value less than human excitation frequencies. Also, acceleration, velocity, and displacement responses were decreased significantly to be within acceptable limits. Finally, a recommendation and guidelines are given for considering the induced-vibrations in the design of common footbridges in Jordan.

Comparison of Signal Processing Methods for Reducing Motion Artifacts in High-Density Electromyography DuringHuman Locomotion.

Objective: High-density electromyography (EMG) is useful for studying changes in myoelectric activity within a muscle during human movement, but it is prone to motion artifacts during locomotion. We compared canonical correlation analysis and principal component analysis methods for signal decomposition and component filtering with a traditional EMG high-pass filtering approach to quantify their relative performance at removing motion artifacts from high-density EMG of the gastrocnemius and tibialis anterior muscles during human walking and running. Results: Canonical correlation analysis filtering provided a greater reduction in signal content at frequency bands associated with motion artifacts than either traditional high-pass filtering or principal component analysis filtering. Canonical correlation analysis filtering also minimized signal reduction at frequency bands expected to consist of true myoelectric signal. Conclusions: Canonical correlation analysis filtering appears to outperform a standard high-pass filter and principal component analysis filter in cleaning high-density EMG collected during fast walking or running.

Design of an experiment for the biomechanical and thermal analysis of athletes during prolonged running exercise

Biomechanics and thermoregulation of human running are key aspects playing an important role in the training of professional athletes. In the case of a prolonged exercise, as occurs during the marathon race, the intensity and duration of the exercise (in concert with environmental conditions) can affect the thermoregulatory response and the running mechanics. The aim of this study is the design of a novel experiment able to simultaneously capture information concerning the biomechanics of lower limbs and the surface temperature map of the whole body during a prolonged running exercise on treadmill. Kinematics quantities, such as linear and angular displacement, velocity and acceleration of relevant body markers are recorded by a video system, while surface temperature evolution on time is detected by an infrared thermal camera. A group of five amateur athletes, with long experience in long-distance running competitions, was involved in this study. Results for the sample group, in terms of kinematic (displacements and angles) and thermal (local and total-body skin temperatures) quantities, revealed features related to individual response to the effort. Even though the relatively little number of athletes does not permit to infer general conclusions, the combined mechanical and thermal experiment is deemed to be a useful tool for the investigation of endurance running characteristics.

Running on the hypogravity treadmill AlterG® does not reduce the magnitude of peak tibial impact accelerations

Summary Background Hypogravity treadmills like the AlterG® have become a popular exercise tool especially in return-to-sports and return-to-competition rehabilitation programmes for distance runners, triathletes, footballers and further sports disciplines. By applying a slight positive air pressure in a chamber around the lower limbs, the effective bodyweight force is reduced, thereby reducing gravity load to the musculoskeletal system. So far it was unclear in how far that hypogravity unloading also affects peak accelerations at the tibiae, a common site of running-related overuse injuries. Material and Methods Fifteen well-trained male runners (peak oxygen consumption of 60 ± 4 ml min-1 kg-1) completed three incremental treadmill tests until volitional exhaustion in randomised order, two of which on the hypogravity treadmill AlterG® at 80% and 60% effective bodyweight, respectively, and the third on a conventional treadmill, i.e. at 100% gravity. Triaxial accelerations at the distal portions of both tibiae were captured throughout the three (hypo-)gravity conditions and all running speeds, along with the physiological cost of running in terms of heart rate, oxygen consumption and rating of perceived exertion. Results Mean peak tibial accelerations at impact and active push-off amounted to 12.9 ± 2.3 g0, 12.6 ± 1.9 g0 and 12.5 ± 2.3 g0 for 12–18 km h-1 at 100%, 80% and 60% effective bodyweight, respectively (g0 ≈ 9.81 m s-2). They were not reduced by hypogravity unloading within measurement uncertainty (p = 0.668). However, they exhibited a clear dependence on running speed (p elevated on the AlterG® (p Conclusions Distal tibial accelerations are not reduced in hypogravity running on the AlterG® treadmill in terms of absolute speeds, and are elevated in terms of relative speeds, i.e. at running speeds demanding the same physiological effort. Thus, precaution should be taken by clinicians and coaches when planning rehabilitation programmes for athletes with a recent injury background at the distal tibiae. The cause for the ineffectiveness of hypogravity running in reducing tibial acceleration load is most probably a subconscious energetic optimisation in human running motor control: For energetic reasons, the leaps of each step become flatter under hypogravity instead of steeper, whereas the latter would have intuitively been expected because of the lower effective acceleration of free fall. Level of Evidence: Ib

Ultrasound-derived changes in thickness of human ankle plantar flexor muscles during walking and running are not homogeneous along the muscle mid-belly region

Skeletal muscle thickness is a valuable indicator of several aspects of a muscle’s functional capabilities. We used computational analysis of ultrasound images, recorded from 10 humans walking and running at a range of speeds (0.7–5.0 m s−1), to quantify interactions in thickness change between three ankle plantar flexor muscles (soleus, medial and lateral gastrocnemius) and quantify thickness changes at multiple muscle sites within each image. Statistical analysis of thickness change as a function of stride cycle (1d statistical parametric mapping) revealed significant differences between soleus and both gastrocnemii across the whole stride cycle as they bulged within the shared anatomical space. Within each muscle, changes in thickness differed between measurement sites but not locomotor condition. For some of the stride, thickness measures taken from the distal-mid image region represented the mean muscle thickness, which may therefore be a reliable region for these measures. Assumptions that muscle thickness is constant during a task, often made in musculoskeletal models, do not hold for the muscles and locomotor conditions studied here and researchers should not assume that a single thickness measure, from one point of the stride cycle or a static image, represents muscle thickness during dynamic movements.

Variant and Invariant Spatiotemporal Structures in Kinematic Coordination to Regulate Speed During Walking and Running.

Humans walk, run, and change their speed in accordance with circumstances. These gaits are rhythmic motions generated by multi-articulated movements, which have specific spatiotemporal patterns. The kinematic characteristics depend on the gait and speed. In this study, we focused on the kinematic coordination of locomotor behavior to clarify the underlying mechanism for the effect of speed on the spatiotemporal kinematic patterns for each gait. In particular, we used seven elevation angles for the whole-body motion and separated the measured data into different phases depending on the foot-contact condition, that is, single-stance phase, double-stance phase, and flight phase, which have different physical constraints during locomotion. We extracted the spatiotemporal kinematic coordination patterns with singular value decomposition and investigated the effect of speed on the coordination patterns. Our results showed that most of the whole-body motion could be explained by only two sets of temporal and spatial coordination patterns in each phase. Furthermore, the temporal coordination patterns were invariant for different speeds, while the spatial coordination patterns varied. These findings will improve our understanding of human adaptation mechanisms to tune locomotor behavior for changing speed.

Research on virtual human running autonomous devices based on cyclic-coordinate descent algorithm

The traditional virtual human running research method completes the motion analysis under the ideal state, which lead to cannot obtain the accurate kinematic stability and the coincidence rate of the trajectory curve of the gravity center. To address these problems, the virtual human running research method based on cyclic-coordinate descent (CCD) algorithm is proposed in this paper. The skeleton driver is constructed based on CCD algorithm and the virtual human skeleton model is built. The running state of virtual human is analyzed through multi-joint movement, and the constraint equation of virtual human running is constructed to realize the data simulation of the running movement of virtual human. By using the skeleton model and constraint equation, the kinematics of virtual human is solved. The gravity center trajectory is calculated by using matrix transpose algorithm. Combining with the virtual human upper body movement and the parameters extraction of specified path running, the research of virtual human running based on CCD algorithm is realized. Experimental result show that the proposed virtual human running method improves the kinematic stability by 45.81% compared with the traditional method, and the coincidence rate of the gravity center trajectory is increased by 63.42%.

Springy ankle tether saves runners’ legs

If you're not a natural runner, it takes a lot of effort to get out and pound the streets; and that's not just my personal opinion. ‘Human running is inefficient’, says mechanical engineer Elliot Hawkes from the University of California, Santa Barbara, USA. Cycling through San Francisco's Golden Gate Park during his postdoc at Stanford University, Hawkes noticed that runners wasted energy each time they stopped and started their swinging legs. Back at Stanford, Hawkes discussed the problem with mechanical engineer Cole Simpson and together they came up with a simple way of recycling some of the energy that is usually squandered. Why not tie the legs together with an elastic band that could store energy when stretched to pull the foot forward at the beginning of the next stride?Intrigued by the suggestion, Stanford bioengineer Cara Welker built a computer simulation of an athlete running with a stretchy band linking their limbs, which suggested that human runners could benefit significantly from a stretchy tether. ‘At this point, I joined the team’, recalls human gait energetics expert Jessica Selinger, currently at Queens University, Canada, who worked with Hawkes, Simpson and Welker to design experiments to test how the band would store energy to pull the foot forward. ‘[We] spent months prototyping the device’, she says.Eventually, the team came up with a length of rubber surgical tubing, 25% of the length of the runner's leg, linking their shoes together, which Simpson says was initially disconcerting. ‘There's something tugging at your feet, but when you take it off, your legs feel heavy’, he smiles. Once the team of 19 volunteer athletes was content running with the new springy tether, the Stanford researchers set them going on a treadmill at about 6 miles per hour (2.67 m s−1) – the average running pace of 36 million users of the Strava social fitness network over 15 billion kilometres – with and without the rubber tube. The team also measured the amount of carbon dioxide exhaled by the runners to calculate how much energy they were using, in addition to filming some of the runners and recording the activity of the muscles in their legs to figure out how the stretchy link affected their running.Initially, the tube didn't seem to give the runners much of an advantage; however, by the end of the 2 day experiment, the tethered athletes were using 6.4% less energy than when running freely. ‘We were very surprised to achieve savings this high. That percentage is nearly the entire cost typically associated with swinging the legs’, says Selinger. The springy tether was saving the runners’ legs by pulling them back into action faster at the end of a stride than when they were running unaided, shortening their stride length and increasing their stride rate by 8%, up to 85–95 strides min−1.But how much could a springy ankle tether improve the finishing time of an average marathon runner in practice? ‘A runner, with a natural pace of 2.7 m s−1 and finishing time of 4 h 20 min, could expect a 6% improvement in economy with our device’, says Hawkes, which could see competitors knock an impressive 15 min off their finish time. And when the scientists tested the springy adaptation on athletes running around the Stanford campus, the team was impressed that all of the runners completed a 4 mile circuit and one even increased his pace to run 5 min 40 s miles.