An impedance control based framework for human-robot collaborative surface interaction and tooling

• Main Author: Sreekanth Kana
• Other Authors: Domenico Campolo
• Format: Thesis-Doctor of Philosophy
• Language: English
• Published: Nanyang Technological University 2020
• Subjects: Engineering::Mechanical engineering
• Online Access: https://hdl.handle.net/10356/137755
• Institution: Nanyang Technological University

The quality of industrial tooling tasks is highly influenced by the tool-surface interaction mechanism. Despite the advancement in the technologies, there exist tasks with cognitive aspects that are rather challenging for robots to handle autonomously, thus demanding the incorporation of human skill set. Under such settings, it is essential to provide a platform/framework which enables the user to be fast, flexible and perform minimal programming/re-programming (for varying task profiles) in the planning phase and engage in physical collaboration with the robot to smoothly interact with the surface in the execution phase. To that end, we develop an impedance control-based approach towards robot-surface interaction incorporating a varying degree of human-robot collaboration. Our main contribution is in the generality of the framework, which can be easily tailored to different possible scenarios such as tool-path planning, autonomous and collaborative (in presence of humans) real-time executions. The approach targets interaction with arbitrary surfaces especially in the context of triangular mesh surfaces, through the exploitation of the widely employed concepts in computer graphics (such as virtual proxy). The proposed approach not only incorporates a varying level of collaboration but also encapsulates different interaction (both contact and non contact) behaviours necessitated by applications across the various avenues of robotics, through modulation of a set of physical and virtual forces. The feasibility of our approach in the context of non-contact tasks has been backed up by simulated results. For contact tasks, experiments have been conducted under the settings of both automated and human-robot collaborative modes where the target was to trace a desired path on the surface of a free-from object. For automated tasks, the user was able to easily define a specific path for the robot to follow, with minimal programming. To trace the same path through human-robot collaboration, the algorithm imposed additional constraints (Guidance Virtual Fixtures) on the human motion thus not only confining the robot onto the desired path thus minimizing the subjective errors but also ensuring a safe work environment. We also tested the capability of our approach in terms of implementing region constraints (Forbidden Region Virtual Fixtures) thus constraining the human-robot motion into desired regions of interest. Three different set of trials have been conducted for each of the collaborative operation to study the effect of different impedance settings on the performance of the algorithm. As a special case of the proposed approach, we demonstrate an impedance-controlled guided approach towards collaborated curve tracing applications targeting industrial tooling tasks such as polishing and chamfering and so on, where the tool trajectory does assume an analytical form. Here, we exploit the known geometry of the workpieces to formulate a parametric representation for the tool trajectory. The proposed approach is validated through a set of experimental trials for both edge chamfering and edge polishing tasks (for straight edges). Two sets of trials are conducted for each of the tooling tasks, with and without employing the proposed approach. For the edge chamfering process, the measurements taken across the resulting surface with the help of a GapGun device show that the quality and consistency increase considerably by using the proposed technique. In the case of the polishing process, from visual observation, our approach has seen to be yielding better finishing of the surface compared to the one without the impedance-controlled guidance. The thesis is divided into two parts, where the first part covers collaborative tooling for analytical surfaces/curves. In the second stage, we introduce a generalised framework (for both analytical and free-form surfaces) for robot surface interaction (with human in the loop) with discrete surface models.