My research centres around bipedal locomotion and involves creating a novel biomorphic robotic platform which is capable of multimodal locomotion. I am especially interested in variable stiffness actuators and the optimal control of such joints. For my PhD I will be creating a biomorphic bipedal robot that mimics the human ability both to vary the stiffness of joints and store energy in compliant linkages. I will investigate neurologically inspired control methods based around reflexive modulation of rhythmical control signals using data gleaned from experiments and simulations. In parallel with this I will be looking at the representation and generation of optimal periodic movements with regard to energy expenditure, stability etc.
Traditionally bipedal robots have fallen into one of two categories: joint control, whereby rigid joints follow an exact pre-planned trajectory; and passive-dynamic, where minimal actuation is provided to a system mechanically tuned to oscillate and provide a naturally efficient walking behaviour. Whilst joint control robots are extremely versatile, and can perform any number of tasks, they are also very inefficient in their movements and generally vulnerable to external disturbances. Conversely, passive-dynamic walkers can produce very efficient locomotion at a single set speed, with some robustness to disturbances, but can perform no other tasks. By creating a robot capable of varying the effective joint stiffness it is aimed to create a platform which combines the advantages of both joint-control and passive dynamic approaches.
It is incredibly difficult to measure how the instantaneous stiffness of human joints changes as we walk. The creation of an anthropomorphic bipedal robot which can vary its joint stiffness explicitly will allow us to elicit more information about possible joint stiffness profiles in humans.
Prior to beginning my PhD project I completed the DTC neuroinformatics MSc with a masters thesis entitled "Age related effects on strategy switching ability in a virtual plus maze". Before this I gained an MEng in robotics from Heriot-Watt university.
Research press coverage: