Integration of the H3D-M400 detector and Spot robot for automated area
survey missions

Area survey and search/confirmation of nuclear/radioactive sources in public areas are challenging and time-consuming tasks. When performed manually, the task also has an associated risk of exposing the operators to unknown radiation. With the help of quadruped robots like the Boston Dynamics Spot, these repeated tasks could be automated efficiently, alleviating the radiation risks to inspectors. The Spot robot comes with its software development kit (SDK) for users to write custom code in Python to control the robot. In this study, a cadmium zinc telluride (CdZnTe) detector, namely M400 from H3D Inc. has been integrated with the Spot via a Spot-CORE payload computer, allowing the robot to capture gamma spectra of background radiation and nuclear/radioactive sources present in a facility. A machine-learning model was developed to use radiation maps from the detector to predict the direction of the radiation source. Two missions have been developed, area mapping and search. In the area mapping mission, the Spot robot walks in fixed steps along a pre-recorded path and performs gamma measurements with the M400 detector. In the search mission, a custom-built algorithm navigates the Spot robot to diverge from the path toward the source location; and once in a preset distance to the source, the Spot robot will collect a gamma spectrum for a long acquisition time, e.g., 5 minutes. In both missions, operators can also use the graphic user interface to direct the robot to a specific location to collect confirmation spectra. The paper will present the details of each mission along with the test results with a 133Ba radiation source in a simulated facility.