RECONSTRUCTING MUSCLE ACTIVATION DURING NORMAL WALKING - A COMPARISONOF SYMBOLIC AND CONNECTIONIST MACHINE LEARNING TECHNIQUES

One symbolic (rule-based inductive learning) and one connectionist (ne ural network) machine learning technique were used to reconstruct musc le activation patterns from kinematic data measured during normal huma n walking at several speeds. The activation patterns (or . desired out puts) consisted of surface electromyographic (EMG) signals from the se mitendinosus and vastus medialis muscles. The inputs consisted of flex ion and extension angles measured at the hip and knee of the ipsilater al leg, their first and second derivatives, and bilateral foot contact information. The training set consisted of data from six trials, at t wo different speeds. The testing set consisted of data from two additi onal trials (one at each speed), which were not in the training set. I t was possible to reconstruct the muscular activation at both speeds u sing both techniques. Timing of the reconstructed signals was accurate . The integrated value of the activation bursts was less accurate. The neural network gave a continuous output, whereas the rule-based induc tive learning rule tree gave a quantised activation level. The advanta ge of rule-based inductive learning was that the rules used were both explicit and comprehensible, whilst the rules used by the neural netwo rk were implicit within its structure and not easily comprehended. The neural network was able to reconstruct the activation patterns of bot h muscles from one network, whereas two separate rule sets were needed for the rule-based technique. It is concluded that machine learning t echniques, in comparison to explicit inverse muscular skeletal models, show good promise in modelling nearly cyclic movements such as locomo tion at varying walking speeds. However, they do not provide insight i nto the biomechanics of the system, because they are not based, on the biomechanical structure of the system.