No description available.

This diploma thesis deals with design and implementation of the controllers of a mobile walking robot. The advantage of these controllers are their kinematics and geometrics independent representation, which allow to use them for different robot types and tasks. In this thesis the contact controller is designed, which minimizes residual forces and torques at the robot's center of gravity, and thereby stabilize robot's body. Next the thesis deals with a posture controller, which maximizes a heuristic posture measure to optimize posture of robot body. Because of this optimization, legs are moved away from their limits and therefore they have more working space for next move. Implementation of the chosen solution is made on the robot's MATLAB mathematical model. Controllers are composed into a control basis, that allows to solve general control tasks by simultaneous combination of contained controllers. The algorithm was created for that simultaneous activation and its operation was explained on flow charts.

Humans have a complex musculoskeletal arrangement which gives them great behavioural flexibility. As well as simply moving their legs, they can modulate the impedance of them. Variable impedance has become a large field in robotics, and tailoring the impedance of a robot to a particular task can improve efficiency, stability, and potentially safety. Locomotion of a bipedal robot is a perfect example of a task for which variable impedance may provide such advantages, since it is a dynamic movement which involves periodic ground impacts. This thesis explores the creation of two novel bipedal robots with variable impedance joints. These robots aim to achieve some of the benefits of compliance, while retaining the behavioural flexibility to be truly versatile machines. The field of variable impedance actuators is explored and evaluated, before the design of the robots is presented. Of the two robots, BLUE (Bipedal Locomotion at the University of Edinburgh) has a 700mm hip rotation height, and is a saggital plane biped. miniBLUE has a hip rotation height of 465mm, and includes additional joints to allow hip adduction and abduction. Rapid prototyping techniques were utilised in the creation of both robots, and both robots are based around a custom, high performance electronics and communication architecture. The human walking cycle is analysed and a simple, parameterised representation developed. Walking trajectories gathered from human motion capture data, and generated from high level gait determinants are evaluated in dynamic simulation, and then on BLUE. With the robot being capable of locomotion, we explore the effect of varying stiffness on efficiency, and find that changing the stiffness can have an effect on the energy efficiency of the movement. Finally, we introduce a system for goal-based teleoperation of the robots, in which parameters are extracted from a user in a motion capture suit and replicated by the robot. In this way, the robot produces the same overall locomotion as the human, but with joint trajectories and stiffnesses that are more suited for its dynamics.

This thesis discusses walking robots issues, their classification, management and construction. There are listed the most famous motion algorithms and their graphical representation. Examples of existing walking robots are also mentioned in this thesis. There are also described modifications of hexapod robot, its hardware and software. The robot is controlled through graphical user interface, which displays data from all sensors, visualises positions of all legs and allows the creation of user defined gaits and its simulations.

The Ph.D. thesis is focused on using the reinforcement learning for four legged robot control. The main aim is to create an adaptive control system of the walking robot, which will be able to plan the walking gait through Q-learning algorithm. This aim is achieved using the design of the complex three layered architecture, which is based on the DEDS paradigm. The small set of elementary reactive behaviors forms the basis of proposed solution. The set of composite control laws is designed using simultaneous activations of these behaviors. Both types of controllers are able to operate on the plain terrain as well as on the rugged one. The model of all possible behaviors, that can be achieved using activations of mentioned controllers, is designed using an appropriate discretization of the continuous state space. This model is used by the Q-learning algorithm for finding the optimal strategies of robot control. The capabilities of the control unit are shown on solving three complex tasks: rotation of the robot, walking of the robot in the straight line and the walking on the inclined plane. These tasks are solved using the spatial dynamic simulations of the four legged robot with three degrees of freedom on each leg. Resulting walking gaits are evaluated using the quantitative standardized indicators. The video files, which show acting of elementary and composite controllers as well as the resulting walking gaits of the robot, are integral part of this thesis.

The Ph.D. thesis is focused on using the reinforcement learning for four legged robot control. The main aim is to create an adaptive control system of the walking robot, which will be able to plan the walking gait through Q-learning algorithm. This aim is achieved using the design of the complex three layered architecture, which is based on the DEDS paradigm. The small set of elementary reactive behaviors forms the basis of proposed solution. The set of composite control laws is designed using simultaneous activations of these behaviors. Both types of controllers are able to operate on the plain terrain as well as on the rugged one. The model of all possible behaviors, that can be achieved using activations of mentioned controllers, is designed using an appropriate discretization of the continuous state space. This model is used by the Q-learning algorithm for finding the optimal strategies of robot control. The capabilities of the control unit are shown on solving three complex tasks: rotation of the robot, walking of the robot in the straight line and the walking on the inclined plane. These tasks are solved using the spatial dynamic simulations of the four legged robot with three degrees of freedom on each leg. Resulting walking gaits are evaluated using the quantitative standardized indicators. The video files, which show acting of elementary and composite controllers as well as the resulting walking gaits of the robot, are integral part of this thesis.

The work presented in this thesis aims to make contribution to the understanding and application of six-legged statically stable walking machines in both theoretical and practical levels. In this thesis five pieces of work, performed with and for the three-joint six-legged Robot-EA308, are presented: 1) Standard gaits, which include the well-known wave gaits, are defined and a stability analysis, in the sense of static stable walking, is performed on an analytical level. Various definitions are given / theorems are stated and proved. 2) A free gait generation algorithm with reinforcement learning is developed. Its facilities of stability improvement, smooth speed changes, and adaptation in case of a rear-leg deficiency with learning of five-legged walking are experimented in real-time on the Robot-EA308. 3) Trajectory optimization and controller design is performed for the protraction movement of a three-joint leg. The trajectory generated by the controller is demonstrated with the Robot-EA308. 4) The full kinematic-dynamic formulation of a three-joint six-legged robot is performed with the joint-torques being the primary variables. It is demonstrated that the proposed torque distribution scheme, rather than the conventional force distribution, results in an efficient distribution of required forces and moments to the supporting legs. 5) An analysis of energy efficiency is performed for wave gaits. The established strategies for determination of gait parameters for an efficient walk are justified using the Robot-EA308.