Design and modelling of a reconfigurable parallel kinematic machine with customisable leg layout for enhanced performance

Parallel Kinematic Machines (PKMs) are well know for their high rigidity and precision, as several struts work collaboratively to achieve the desired activations, minimising accumulative errors and providing stiff configurations. As such PKMs have seen successful applications as machine tools, rapid positioners and high precision pointers for antennas and lasers. However, their wide application has been severely limited due to small workspaces (compared to the machine size), singularities in the workspace and complexity of implementation. For these reasons PKMs have been limited to operate in well prepared environments, such as those of workshops and laboratories. Nevertheless, currently robotic systems require to be highly adaptive to different task and scenarios. Adaptive PKMs or also known as reconfigurable PKMs represent a possibility to overcome the limitations of PKMs, as their modifiable architectures can be utilised to optimise their workspace, avoid singularities and in general enable systems capable to adapt to different environments. Furthermore, given the redundancy of actuators PKMs can be utilised as walking machine tools. Nevertheless, providing walking capabilities to a robotic system represents a complicated task. For this reason, most walking robots are based on fixed configurations that are selected according to desired environments and walking conditions, and their performance may be hindered if the ideal requirements are modified. For instance, mammal-inspired walking robots can transport heavy loads at high walking speeds thanks to their vertically-configured legs. Conversely, insect-inspired robots (e.g. hexapods) can traverse complicated terrains, as they can maintain three or more limbs in contact with the ground at any time, with higher stability coefficients than bipeds and quadrupeds. Thus, the number, distribution, and configuration of the legs play a crucial role in the performance of walking robots (e.g. speed, stability coefficients, actuation torque, etc.). Therefore, designing a Reconfigurable Parallel Kinematic Machine could bridge the current gap for PKMs , enabling its wider use as walking machine tools. Furthermore, enabling the reconfigurable capabilities PKMs could be exploited to walk on different terrains and optimise their performance during machining operations. In the research of this PhD thesis, the concept and design and test of a walking Reconfigurable Parallel Kinematic Manipulator (RPKM) is proposed. By presenting a reconfigurable kinematic architecture, the system is enabled to operate in many different configurations Furthermore, this thesis presents the kinematic and gait analysis for machining and walking operations. Special attention is presented to singularity avoidance, where an efficient methodology is presented to analyse the workspace and create singular-free paths. A novel methodology is presented to perform the safe walking of the proposed RPKM over different terrain conditions (e.g. flat and inclined terrain at different slopes). Furthermore, several configurations are implemented to show the advantages of a reconfigurable system with minimal walking errors