CaveX: A Low-Impact Robust Cave Mapping Hexapod Robot

Purpose In this paper, we present the Cave Exploration (CaveX) robot which is capable of low-impact cave exploration and three-dimensional (3D) mapping to aid scientific research, with continuing project development toward full autonomy. Design/Methodology/Approach The CaveX robot is a biologically-inspired hexapod which uses LiDAR (Light Detection And Ranging) sensing to explore and map cave systems while minimising its environmental footprint. It comprises five subsystems—sensing, control, power, structural and movement, and communications—which were derived from a functional analysis of desired system behaviour. A model-based systems engineering approach complemented a rigorous prototype design and construction process with ongoing stakeholder input. An onboard computer runs the Robot Operating System (ROS), controlling all movement and LiDAR scanning which is well-suited to low light conditions. An open-source code repository was implemented to control the robot’s movement, featuring four walking gaits and closed-loop feedback using the motors’ torque load. Six legs, each with three degrees of freedom, were mounted to a carbon fibre chassis to enable agile traversal of the highly variable cave terrain. Digital scan data files collected by the robot have been post-processed using a Simultaneous Localisation and Mapping (SLAM) algorithm, and may be integrated into existing 3D cave models. Findings The robot, as tested in real cave environments, successfully mapped confined areas for up to 30 min on various terrain types while operating on a 11.1 V LiPo battery. Originality The robot system aims to facilitate rapid deployment into regions of caves that are inaccessible or hazardous to humans, delivering robust, low-impact mapping capabilities to scientists whilst providing simple interfacing between robot and operator. The robot’s combination of 3D LiDAR scanning, small footprint and agile design enables it to reach areas other robots cannot, collecting new data points for research.