Based on measured motion characteristics of the human shoulder we developed a kinematic model of the shoulder complex and designed a number of prototypes of the humanoid shoulder girdle. These prototypes included a spherical parallel mechanism. It was the first humanoid parallel shoulder girdle ever built. In order to design the mechanism, a profound study of kinematic singularities and sensitivity to the changes in kinematic parameters had to be done. One of the prototypes was actuated by pneumatic muscles with the goal to replicate the human motion. We also showed that a movement of a humanoid shoulder is produced by two coplanar movements, in the shoulder girdle and in the glenohumeral joint. In the area of redundant robots we developed an original method to solve the inverse kinematics, which enables to maximize the gradient in the secondary task level. This method increases the flexibility and adaptibility of a redundant robot. We additionally introduced the criterion of kinematic flexibility which measures the amount of redundancy in terms of the self-motion space. This criterion can be used either to identify the most suitable volume inside the workspace where the mechanism can solve different tasks or to design and optimise the mechanism itself. Redundant robots are usually designed to move in unstructured envirnonments. We developed different strategies which enable to control a redundant robot to move and solve the primary task in unknown space. In particular, we studied the obstacle avoidance problem. Experiments were carried out with a mobile manipulator possesing a sensory equipped mobile platform and a redundant planar 4R torque-controlled manipulator. We studied new methods to perceive the robot environment by a humanoid vision. We combined cameras of different resolution in order to detect, pursue, recognise and reach various objects of interest. We were able to establish an interaction between humans and humanoid robots that would not be possible by using standard computer vision systems. The vertical jump is a specific test used in sports to evaluate the status of an athlet. We studied the biomechanics of the vertical jump by using a dynamic model of the humanoid leg mechanism, muscles and tendons. It was shown that the height of the jump primarily depends on the elasticity of the whole system, especially of the Achilles' tendon. In our analyses special attention was given to the role of biarticular muscles. In the area of biomedical research we introduced a new approach of electrical stimulation of abdominal muscles. We demonstrated by experiments that electrical stimulation can temporarily replace mechanical ventilation and can contribute to a better distribution of air in lungs. The periodic abdominal muscle breathing activity-exercise load relationship and the phenomenon threshold were discovered. Results will be used for bio-feedback and relaxation or blocking of abdominal muscles during breathing. A neurophisiologyc investigation was carried out with the goal to develop an active machine which would help paraplegic patients in standing and mantaining equilibrium. Our focus in environmental medicine was on altitude acclimatization. We investigated a variety of physiological responses which reduce the magnitude of the relative work rate. We studied the manner in which acclimatization influences cardiorespiratory, metabolic and thermoregulatory function during exercise. Specifically, we studied the effects of both experimentally induced ischaemia and hypoxia on work capacity at altitude. We developed thermal foot manikin with gait simulator, which allows the determination of footwear insulation during static (standing) and dynamic (walking) conditions.