The generation of joint trajectories for humanoid robots is not an easy task because of complexity of humanoid robots. We propose to exploit the similarity between human and humanoid robot motion. We show how to transform human motion captured by optical tracking device into a trajectory for a robot. We propose an automatic approach to relate humanoid robot kinematic parameters to the kinematic parameters of a human performer.
COBISS.SI-ID: 18277927
We have addressed the problem of robot control that is capable of performing rhythmic tasks in a human-like way. Robots that can perform such tasks requires complex sensory systems and advanced control strategies. As an example we have selected a yo-yo. We have analysed how a human operates these objects and then developed models. We propose two control strategies: one based on predefined hand motion patterns, and the other generates the motion on-line depending on the prevailing situation
COBISS.SI-ID: 19814439
We investigated the kinematics of a human arm that we developed based on an enhanced model which was synthesized using motion measurements of healthy subjects. We proposed a method for solving redundancy of mechanism and avoidance or use of the singularity. The model of shoulder complex was used for the evaluation of reaching working space of the arm and for evaluation of arm functionality in rehablitiation and offers a possibiltiy of a direct comparison of an impaired and a healthy shoulder
COBISS.SI-ID: 20186407
Until recently, regulation of body temperature was assumed to be achieved by initiating responses to maintain deep body temperature close to a reference temperature. Our research has demonstrated that set-point theory of temperature regulation and of reciprocal inhibition do not adequately represent temperature regulation because they have neglected the contribution of non-thermal factors. We developed technology and a protocol for assessing behavioural thermoregulatory function in humans
COBISS.SI-ID: 19871527
We analyzed the conceptual difference between pseudo-inverse and minimal null-space based control algorithms. We proposed an efficient solution to computational instablity of MNS algorithms. We proposed a method for obstacle avoidance applying the force control or by reducing inertia in the null-space and alternative solutions for obstacle avoidance using vision system. The proposed algorithms improve the tobot performance and better responses for fast movements
COBISS.SI-ID: 20982567