Path Planning & Measurement Registration for Robotic Structural Asset Monitoring

In recent years, robotic devices have been explored across a wide range of disciplines as delivery mechanisms for multiple sensing technologies applied to structural asset monitoring. Typically the drivers for moving to automated delivery relate to operation in environments that would be hazardous for human operators (working at height, on live plant, or radioactive environments). The exact measurement requirement varies dependent on application, for example specific localised measurements may be required at a known problem area, or a more general wide area survey may be undertaken. Repeated measurements may be required to quantify the evolution of structural condition/ damage in a structure. In all cases, a common requirement is for efficient path planning for the robotic vehicle, combined with accurate registration of the position on the structure that the measurement was undertaken, the latter of particular importance when considering repeated measurements. Since 2003, The University of Strathclyde has been involved with research as part of the UK Research Centre for Non Destructive Evaluation (RCNDE) to develop robotic deployment of NDE measurement probes. The use of multiple inspection vehicles (coupled with different sensing modalities), allows for a flexible, reconfigurable, and adaptive approach to remote health monitoring. The move to increased levels of autonomy for robotic delivery of inspection for asset monitoring, demands a structured approach to path planning and measurement data presentation that greatly surpasses the more ad‐hoc approach typically employed by remotely controlled, but manually driven robotic inspection vehicles. The authors describe a traditional CAD/CAM approach to motion planning (as used in machine tool operation) which has numerous benefits including the provision for detailed structural layout to be built into the path planning, and the subsequent visual overlay of measurement data back into the CAD model. Initial trials using a steel plate with varying thickness defects have been investigated. A custom robot incorporating an ultrasonic wheel probe as rear wheel has been used to perform an ultrasonic thickness scan of the sample – the results have been compared to those obtained using a high resolution wheel probe scan of the sample. The robot control was generated from an initial CAD model of the sample, and the integration of post‐processor output (in this case G code), was used as the input for our custom robotic control software to generate both simulation and experimental motion control of our robotic platform.