Friday July 6 2012
'Real robot legs' hope: but can they get up the stairs?
BBC News has heralded the “most realistic robot legs” ever developed, in a headline that could raise hopes of sci-fi style exoskeletal limbs to help disabled people. Meanwhile, the Daily Mail has rather run away with the story, saying that researchers have created Wallace and Gromit style “‘wrong trousers’ that walk by themselves”.
The BBC says that US experts have developed the most “biologically accurate” robotic legs yet. These could help understanding of how babies learn to walk and have some role for spinal injury treatment. However, this is still only a ‘baby step’ towards an effective device that a person could use.
The research underpinning this headline involved engineers developing a robot that walks the way that humans walk. The robot has loaded, motor-controlled straps that mimic muscles in the legs, as well as a “central pattern generator”, which mimics the nervous system and reflexes.
The researchers say that walking relies on a complex system of muscles, nerves, balance and co-ordination, and that this system provides a valuable insight into the processes involved in the way people and animals walk. At this stage, the research will not help people who are paralysed or who have amputated limbs, but it could pave the way for the possibility of robotic controlled legs for people in the future.
Where did the story come from?
The study was carried out by researchers from the Department of Electrical and Computer Engineering, University of Arizona. Assessment of human subjects walking was funded by the National Institute of Health Research. The study was published in the peer-reviewed engineering journal Journal of Neural Engineering.
Headlines aside, the news coverage is representative of this research, and includes video footage of the robotic legs in motion.
What kind of research was this?
This was scientific research involving the development of a walking robot. The robot was constructed using Kevlar straps to model human muscles, with a computer message system that simulates the nerve pathways that signal to muscles, controlling their movement. Walking is a dynamic process involving interaction between the muscles, skeleton, the environment and the nervous system. The researchers say that this complex process means that if we can understand locomotion, it opens up the possibility of further understanding of how the brain works to create movement.
The researchers describe that in the lumbar region of the spinal cord, “the ‘central pattern generator’ produces rhythmic signals that interact with the biomechanics of the body to generate the step cycle”. They say that this central pattern generator uses feedback from several sources, including sensations from the feet, the nerves that sense muscle stretch and limb-loading in the legs, and the nerves that sense the position of hip joints.
The researchers developed a robotic model that is representative of the muscular movement, the sensory feedback and this central pattern generator involved in human walking.
What did the research involve?
The robot is a simplified model of the legs, based on previous research findings. The system comprises three joints (hip, knee and ankle) and nine muscles, including extensor muscles that extend, or straighten, the joint, and flexors that bend it. It includes three biarticular muscles, which are muscles that cross joints:
- the gastrocnemius, which spans the knee and ankle
- the rectus femoris, which spans the hip and knee
- the hamstrings, which also span the hip and knee
For each muscle in the robot, a special motor was attached to a bracket. A Kevlar strap was then buckled to the motor, with muscle contraction mimicked by rotating the motor to pull on the strap. A computer model designed to stimulate the central pattern generator produces signals to control each motor in the robotic legs. Each strap also holds a sensor that feeds back to the central pattern generator and measures the amount of tension or load produced. Other sensors provide feedback about contact with the ground and upon hip position. The results from these sensors are kept so that they can be compared to the way people walk.
What were the basic results?
Simply, the researchers demonstrated that they were able to make the robot walk normally. They also demonstrated that feedback to the central pattern generator from the sensors in the feet altered the signals coming from the generator and that this prevented toe stubbing and produced the correct ‘toe-off’ walking gait. When the researchers added a weight to the right robot ankle, the central pattern generator helped to stabilise the gait despite this physical disturbance. Without the central pattern generator, the right foot “dragged”.
Comparing their findings to two normal human subjects, they found that the joint angle movements in these people were comparable to the joint angles reported by the robot’s sensors. Other mechanics of movement, including timing of joint flexion, was similar between the humans and the robotic legs.
How did the researchers interpret the results?
The authors say that their research “represents a complete, if simplified, physical neurorobotic model of the lower human body”.
Engineers have developed a walking robot that models the way that humans walk. The robot has loaded straps that mimic muscles in the legs, as well as a “central pattern generator” mimicking the nervous system and reflexes. Compared to joint movements involved in normal human walking, the researchers have shown that the robot is a complete model for human walking.
The researchers say that this system could be of value in helping to understand the physiological processes involved in walking in animals and humans. The BBC quotes a UK expert as saying, “this work is exciting because the robot mimics control and not just movement”.
While this is exciting research, and the headlines may have conjured up images of cybermen – or in the Daily Mail’s case Wallace and Gromit’s “wrong trousers” – it doesn’t stand up to the hype. For example, while the researchers have made a seemingly excellent model of robotic walking, they have not demonstrated other lower limb functions, such as:
- sitting down
- standing up
- crouching or kneeling
- climbing stairs
Until such time as robotic legs are devised that can accurately and comfortably carry out these actions as well as walking, the therapeutic implications for paralysed people, or those who have amputated limbs, are extremely limited.
Analysis by Bazian. Edited by NHS Choices. Follow Behind the Headlines on twitter.