First steps: the mechanics and control of velocity change in humans

Lead Research Organisation: University of Manchester
Department Name: Earth Atmospheric and Env Sciences

Abstract

Humans moving around on two legs appear to do this effortlessly most of the time. However, particularly as we get older, injuries due to falls are unfortunately very common. In order to understand how to reduce the frequency of falls, and to help prevent other injuries to our muscles, joints, bones, tendons and ligaments, we need to improve our understanding of how the locomotor system works. Locomotor research has concentrated primarily on what are called steady state activities such as standing in one place, or moving at a constant speed on a treadmill. However a lot of injuries occur at other times such as when starting, stopping or turning. These are the activities that are the focus of this research project. The trouble is that it is very hard to study these forms of locomotion because, by their very nature, they do not simply repeat themselves in an easily controlled fashion. We therefore plan to study people moving freely around the laboratory and responding to spoken commands (e.g. "turn left", "stop"). While they are doing this we will record what happens to their bodies in terms of the muscles that are doing the work, the movements that the limbs perform, and the forces that are applied to the outside world. However this only tells us half the story. We really want to know what is going on at a level where we can use the information to develop techniques that we can use in rehabilitation or training. That means that we need to know what is going on inside the body. This is very difficult to do experimentally so we are proposing instead to construct sophisticated computer models of locomotion that can be trained, using a supercomputer, to mimic the actions that we record from experimental subjects. In the computer model we can inestigate and control everything that is happening which lets us get a much more complete picture. In the computer we can simulate changes in muscle strength to mimic the changes associated with ageing, measure its effects, and directly test strategies and treatments that can help improve locomotor performance. This "in silico" approach fills the gap that is left by purely experimental approaches and allows us to get a much more complete understanding of locomotion for both practical and purely theoretical reasons.

Technical Summary

A locomotor system needs to be able to start from a stationary position, accelerate to desired velocities, progress at a constant velocity, change direction, and decelerate to a stop. A great deal of work has been carried out looking at the mechanics and control of steady state activities, however much less is known about how humans achieve controlled velocity change (gait initiation, gait termination, speed change, or turning) at either the experimental or theoretical level. Work on control has concentrated on externally measurable parameters for obvious ethical and practical reasons. Thus we know a great deal about the kinematics, EMG, and external forces associated with velocity change. However control is a largely internal activity and so there is a limit to what we can learn from external measurement. This is a complex biological process that depends upon the interaction of multiple components. The aim of this project is to use a systems biology approach to produce an in silico representation of the complete system and use this to run virtual experiments that will allow us to refine and test our understanding of how the components interact. This has recently become possible due to the development of high biofidelity forward dynamic simulations with global matching that can cope with both arbitrary movement patterns and also handle the complexities inherent in compliant actuation and passive energy stores. As a synthetic rather than an analytic technique it can identify the control processes that are able to both generate the required locomotor output and that also satisfy the physiological limitations of the system. Thus the in silico systems approach can fill in the gap left by experimental work and will allow us to understand human locomotor control at a level sufficient to provide useful strategies that can be used for rehabilitation, training, biomimetics, and to further understand the changes associated with development, senescence, and pathology.

Planned Impact

Steady state locomotion is a relatively rare occurrence in the real world due to unevenness in terrain, and the requirement to change velocity for starting, stopping, and to move round obstacles. In addition, injuries to the locomotor system often occur during these periods of velocity change. However research in locomotor biology has traditionally concentrated on steady state activities and research into these other aspects is relatively lacking. As such this project will be of great interest to workers in locomotor biology, sports science and clinical gait analysis. The software developed by the project is open source and this widens the general utility of the work by providing information about human locomotion to computational scientists and the robotics community.

However the application to injury means that it will have a much wider impact. Musculoskeletal injuries are particularly common among the elderly, or those participating in sport, with falls being a key risk factor. These falls are often associated with activities such as turning or stopping, and the risk of injury after a fall is much higher when turning is involved. This is a major problem with an ageing population and this project will provide essential basic information about the mechanics and control of normal locomotion that is highly relevant to helping minimise the risks of falling through identifying at risk individuals, designing exercise programmes to minimise risk, and improving rehabilitation support. Thus this project will have impact on clinicians and other stakeholders within the health industry.

In addition the science presented by this project, coupled with its obvious clinical application, will have large general public interest. Since we are addressing running as well as walking there are equally obvious sporting links and this will further heighten public interest. People are fascinated with how the human body works and analogies to the 'human machine' are a very effective way of communicating this information.
 
Description We have developed a technique for assessing body size and proportions by taking multiple photographs simultaneously and reconstructing a 3D model from the photographs. The experimental side of the work has generated a great deal of raw data and the analysis is progressing (but slowly because the award has now finished). We have produced 7 publications primarily based on the modelling side of the grant which has been very successful. We have developed new techniques to control simulated locomotion which have allowed us to address questions such as acceleration, multiobjective optimisation, and multiphysics simulations that have been successful and provide opportunities to expand our work further in future.
Exploitation Route Our initial output is a new method for assessing body proportions which should find a high degree of utility. The dataset we are amassing on normative walking will also be useful to the community. Our software is becoming more widely used by the general community provides a unique tool for people interested in machine learning in locomotor control.
Sectors Healthcare,Leisure Activities, including Sports, Recreation and Tourism,Culture, Heritage, Museums and Collections

URL http://www.animalsimulation.org
 
Description The simulations developed for this project has been used for a number of publications that have led to media attention. Our work on "Quadrupedal locomotor simulation: producing more realistic gaits using dual-objective optimization" was covered by Discovery Channel Canada, Engineering & Technology eandt.theiet.org, and phys.org amongst others.
First Year Of Impact 2018
Sector Other
Impact Types Cultural

 
Title GaitSym2014 
Description This is the latest version of our Multibody Dynamic Simulation tool. It has been updated to allow much easier custom model creation and better compatibility on multiple platforms. 
Type Of Technology Software 
Year Produced 2014 
Open Source License? Yes  
Impact The software was featured during a recent news story on locomotion simulation. 
URL http://www.animalsimulation.org
 
Title GaitSym2015 
Description Multibody dynamic simulator customised for biological use. 
Type Of Technology Software 
Year Produced 2015 
Impact The software was used to create part of the new Human Evolution exhibit at the Natural History Museum in London. 
URL http://www.animalsimulation.org
 
Title GaitSym2016 
Description :atest version of our in house multibody dynamic analysis package. 
Type Of Technology Software 
Year Produced 2016 
Open Source License? Yes  
Impact The software is now used in at least 4 laboratories around the world. 
URL http://www.animalsimulation.org