Y. Moline a, F. Carrel a, G. Corre a, C. Lynde a
a CEA, DRT, LIST, DM2I, SCI, Laboratoire Capteurs et Architectures Electroniques Laboratoire Capteurs et Architectures Electroniques, F-91191 Gif sur Yvette, France
This paper introduces a new approach for radiation localization and its implementation in the context of the CEA project called RADLOC (RADiation LOCator). The RADLOC development tends to ease the radioactive source localization by the use of a complete handheld direction-sensitive sensor based on user movements.
Localization of radioactive sources is currently a main issue addressing different fields, including nuclear industry, nuclear decommissioning as well as homeland security. The objective of the RadLoc project is to propose a system capable of quickly locating a radioactive source. When talking about localization, the first step is to go in the right direction. Existing systems can be divided in two categories. The first one, focused on the sensor part, is composed by expensive solutions, which require a high level of expertise to be handled. Most recent examples are based on multiple sensors , ,  to achieve coincidence processing or covers 360 degrees measurements . The second one, most commonly encountered, is based on the use of ictometer-type detector (e.g. Geiger-Müller) by observing the count rate displayed on a screen, a visual signal (LED) or audible signal (buzzer). With the detector in hand, the user scans the surrounding space in order to vary the count rate indicated and thus obtain an initial indication of direction. In this case of concrete application, the ability of the user to gauge the evolution of the signal (visual or audible) over time and to take an arbitrary direction is necessary for the method to work. However, these systems never benefit from the movement that naturally occurs when brought by the handler. Based on this idea, this paper proposes a low-cost and lightweight measurement system able to perform the same function automatically by adding the notion of movement, resulting on a direction-sensitive detector able to find the closest minimum distance between the starting point and the location of the radioactive source.
The RADLOC system is built around four orthogonal sensors, a four channel ad-hoc digital pulse processing (DPP) board, an embedded PC, an Inertial Measurement Unit (IMU), a screen and all required voltage supplies. Each measurement channels is composed by an YSO(CE) scintillator (0.28x0.28x0.83 inch) coupled with SiPM, directly connected to Analog to Digital Converters (ADC) sampled at 250 MHz on the DPP board. A Field Programmable Gate Array (FPGA) achieves sampled data acquisition and pulse processing. Configurable DPP processing units such as smoothing, Baseline Restoration (BLR) or fast timing discrimination for pulse feature extraction are performed (count rate, pulse height or width). Then, features are transmitted to a microcontroller for frame building and serial communication with the embedded PC, a Raspberry Pi. Then, the Raspberry Pi carries out the angular position reading with an IMU and displays on the screen the direction to take following the example of a compass. For the first version of the prototype, two measurement criteria are used to make an elementary decision-taking algorithm and plot the right direction. The first is the difference of counting rates between each measurements channels, with calculation of the barycentre. The second is the counting rate variation between the detectors during the user's movements. These criteria can ideally benefit from the shielding effect between the detectors aligned with the radioactive source. Different count rates granularities are used to allow the system an automatic adaptation to the activity i.e. the proximity of the source in order to process more statistics or make the display of the direction more reactive.
Preliminary results obtained with the first prototype are very promising for radioactive sources at very short range (<2m). The entire DPP architecture is flexible enough to be used with many other sensors. Therefore, with an objective of performing localization at least 5 meters away, Monte-Carlo simulation are accomplished to design a system with larger sensors while respecting the constraints of portability and will be detailed in the full paper.
 C. B. Hoshor et al., “Real-time neutron source localization and identification with a hand-held, volumetrically-sensitive, moderating-type neutron spectrometer,” Nucl. Instruments Methods Phys. Res. Sect. A Accel. Spectrometers, Detect. Assoc. Equip., vol. 866, no. May, pp. 252–264, Sep. 2017, doi: 10.1016/j.nima.2017.05.033.
 D. E. Newman, “Radiation Detector With Two-dimenstional Directionality,” 2018.
 E. D. Greenlee, “Real Time Fusion of Radioisotope Direction Estimation and Visual Object Tracking,” University of Tennessee, 2018.
 E. M. Becker, “A Direction-Sensitive Radiation Detector for Low-Altitude, UAV-based Radiological Source Search,” Oregon State University, 2016.