Control Schemes for Bilateral Teleoperation of Wheeled Mobile Manipulator Robots

Abstract

This thesis focuses on the development of advanced control schemes for bilateral teleoperation of Wheeled Mobile Manipulator (WMM) robots with haptic and visual modalities. WMM systems are particularly valuable because they bring together mobility and manipulation, enabling the extension of human capabilities and the leveraging of human intelligence over a distance. However, teleoperating such highly redundant robots through a single haptic device remains challenging due to mismatches in mechanical structures, disparate workspaces, varying degrees of freedom, operating in cluttered environments, and limited sensory feedback.

One key contribution of this thesis is the development of a novel switching scheme designed to address the workspace mapping challenge while avoiding transparency deterioration in the bilateral teleoperation system. The proposed method leverages force information, with or without force sensors, to ensure that interactions with the remote environment occur within the appropriate operation mode. Instead of relying on the operator or the haptic display's motion, as in conventional switching methods, the proposed approach leverages interactions with the environment to automatically transition from navigation to manipulation mode, enabling contact-based tasks without compromising transparency and maintaining fine-manipulation capabilities.

To enable effective remote task execution, control schemes must exploit the redundancy of the WMM robot while simultaneously enhancing operator performance. Building on whole-body motion and semi-automatic switching, a novel hybrid control scheme is developed to improve operator performance by reducing mental workload and task completion time. Moreover, the scheme incorporates essential constraints—including joint limits, safe physical interaction, and obstacle avoidance—to allow the robot to navigate and operate reliably within cluttered environments.

Another key area of research explored in this thesis focuses on resolving conflicts between operator commands and those generated by an optimisation-based constrained controller. Because the controller prevents the operator from issuing commands that may lead to collisions, a haptic shared control scheme is proposed to help the operator avoid potential conflicts. In this scheme, haptic cues act as warning signals—delivered through guidance or repulsive forces—so that when the operator visually observes deviations between the robot's motion and the commands provided through the haptic device, the assistive haptic feedback offers an additional explanation for why the robot cannot follow those commands.

In summary, this thesis presents control schemes that enhance the safety and ease of bilateral teleoperation of WMM robots by using haptic and visual feedback. It introduces an environment-based switching control scheme that automatically transitions from navigation to manipulation modes, ensuring transparency and fine-manipulation. Additionally, it proposes a hybrid control strategy that leverages the robot's redundancy while adhering to practical constraints in cluttered environments, with a focus on enhancing human operator performance. Lastly, it incorporates haptic cues to mitigate confusion when safety restrictions prevent the robot from following the operator's commands, ultimately improving task execution.