Research

For manipulation, we formulated the framework so that the constrained controller optimises the contact force, and the unconstrained controller performs the desired task while been compliant to human interactions. We borrow the concept from robot grasping so that this is achieved by divided the control space into internal/and external forces.

For locomotion, we need to consider how to control swing foot position, balance, and maintaining the contacts of the stance legs. For this, we divided the controller into three separate controllers. During walking, the task of the swing leg is to move the foot position to a designated target. It is important to have Cartesian impedance so it is compliant to uneven terrain or unpredictable obstacles The torso position is crucial because it is highly relevant to the balance of the robot. For the stance leg, we need to find out how much ground reaction force is needed to maintain contacts such that the robot can resist any external forces, which may be from human interaction or the environment where the robot is acting on. For this, we use constraint optimisation to find the solution

In this work, we propose a probabilistic-based approach for detecting transitions in hybrid control systems with limited sensing. We empirically validate our approach while detecting contact transitions in a hand-over scenario where a human operator brings a large object and hands it over to a pair of robotic arms.

In this work, our aim is to develop a method such that some previously learnt behaviours can be adapted to new task in an appropriate way. In particular, we consider learning the null space projection matrix of a kinematically constrained system, and see how previously learnt policies can be adapted to novel constraints. In this example, we teach the robot hand to operate a remote control.

Legged systems need to optimize contact force in order to maintain contacts. For this, the controller needs to have the knowledge of the surface geometry and how slippery is the terrain. In this paper, we propose an online method to estimate the surface information via haptic exploration.

What defines a human movement is an overloaded question with contradictory interpretations. One way to arrive at a solution is to employ optimisation strategies that somehow capture the essence of gait. However, this is not trivial since the motion that we observe are masked by intra-personal and inter-personal variations.

We propose a novel method for generalising the consistent characteristics of human movement subject to variations in environment, embodiments, and behaviours. This is achieved by reconstructing an unconstrained policy without explicit knowledge of the variations.

One of our objectives is to measure the difference between gaits. A potential application in gait rehabilitation, for instance, is to determine how much the device should correct the subject. The principle is to compare the subject’s gait with an appropriate reference gait which is expected to be normal and use their difference in a feedback controller. For this, we need a way to compare the difference between two gaits.