WP2 Control Systems Design
Action 2-1: Study on stability and control methods in animals and humans. (UTH). Humans and animals will be guided to walk and run along a 15m track with two force plates. The track will be instrumented with wooden obstacles with heights up to 2.5 cm before and after the two force plates. Moreover, the second force plate will be set on four different heights (0 cm, 5 cm, 10 cm, 15 cm). The participants will be visually aware of the whole walking track and will be allowed to perform several practice trials along the walkway. Then the participants will be guided to trials on the walkway with different step heights and different track compliance. The trials will be successful if the participants hit both force platforms during the touchdowns. Limb and body kinematics will be recorded using our 10- camera Vicon motion analysis. Multiple markers will be positioned on each segment to obtain realistic 3D motions. Moreover the ground reaction forces will be measured with the aforementioned force plates. From the kinematic and kinetic analysis, the leg contact angle and leg stiffness will be calculated during gait at different track configurations. In addition, the role of leg muscles and how they are linked to kinematic and dynamic parameters will be investigated. To achieve this goal the purchase of an Electromyographic (EMG) system is essential. EMG, kinematics and dynamic parameters will be recorded simultaneously.
Action 2-2: Stability and control methods for four-legged robots (NTUA). Alternative control methods will be developed, which will successfully enable the robot to move statically and dynamically in a stable and repeatable manner. The control methods will be validated and evaluated analytically and computationally (Matlab), using the model developed in WP1. During this procedure, the robot parameters will be modified to determine which control method is more suitable for different robot setups. This way the optimum setup and control method can be determined with respect to the robot’s main functionality and task. For example, if the main task for a four-legged robot is to run as fast as possible, its parameters and control algorithm will be quite different compared to a quadruped oriented towards maximum stability on rough terrain. Depending on the outcome of this study, the possibility of changing the robot’s parameters according to its primary task will be studied. The control parameters include the DOFs (locking or unlocking a joint), stiffness for joints (e.g. changing a springs’ pre-compression), leg length and other which will be determined during the project.
Action 2-3: Control methods for multi-legged robots (FORTH). The computational models of multi-legged robots developed in WP1 will be used for the design of motion control strategies for locomotion in unstructured environments. As a first step, open-loop strategies will be developed, focusing on the generation of gaits observed in biological organisms with morphology similar to the robots under consideration. This will be based on distributed control architecture, e.g., employing networks of nonlinear oscillators inspired by biological central pattern generators (CPGs). Subsequently, the exploitation of sensory information will be investigated, both for robust adaptation to terrain irregularities, as well as for implementing different reactive behaviours. These efforts will be supported by comprehensive simulations, employing the computational tools developed in WP1.
Action 2-4: Control methods for biped robots (FORTH). The study of bipedal locomotion control will first consider open-loop strategies, accomplishing stable movement in planar surfaces. Subsequently, closed loop schemes will be considered, that will enable robust locomotion in a planar surface and in uneven terrain. The computational tools will be extended to include simulated complex surfaces that will allow testing the control schemes in environments with varying stiffness, frictional forces, etc. Taking advantage of the anatomical characteristics of the human body, novel control algorithms will be designed, that coordinate lower and upper limb movement, improving robot stability when moving on uneven terrain (e.g., opening hands wide to enhance balance).