Analysis of Single Event Transients (SET) using Machine Learning (ML) and Ionizing Radiation Effects Spectroscopy (IRES)
Abstract : A methodology for automating the identification of single-event transients (SETs) through Ionizing Radiation Effects Spectroscopy (IRES) and machine learning (ML) is provided. IRES enhances the identification of SETs through statistical analysis of waveform behavior, allowing for the capture of subtle circuit dynamics changes.Automated identification of SETs is facilitated by a k-Nearest Neighbors (kNN) ML algorithm with IRES data. One-hundred thousand waveforms were measured from CMOS phase-locked loop (PLL) circuits irradiated at the Naval Research Laboratory’s Two-Photon Absorption (TPA) laser facility. Known SET signatures were used to train various kNN models based on statistical features derived from several standard circuit metrics and eight moment-generating functions. Results show that SETs can be automatically identified by the kNN models, with several features resulting in greater than 98% correct identification of SETs. The tradeoffs in ML-based anomaly detection, based on the size of available training sets, choice in signal metric, and the number of included statistical moment-generating functions are discussed, along with opportunities for the future development of specific event-type classification, in situ measurement, and real-time classification of data.
ML classification is used to perform an in-situ prediction of the TID level and circuit bias voltage with a single measurement. The TID response of VCO and PLL circuits designed and fabricated in a 130-nm CMOS technology is presented. In response to the TID radiation up to 300 krad(SiO2), the RF circuits (VCO and PLL) show a shift of 30 MHz in frequency. It is important to note that the behavior observed in the circuits used for this experiment is limited to the specific technology and circuit design. Other circuit designs and device fabrication parameters may exhibit different response to the similar TID radiation exposure described in this work.
The interactions of satellite devices or electronics with the charged particles can disturb the satellite operations by setting or resetting bits, creating transient signals that compete with legitimate signals within the systems, and by degrading device parameters.
? Sometimes these interactions can permanently damage the satellite and jeopardize the mission. For example, the HIPPARCOS satellite mission of the European Space Agency (ESA) was terminated due to radiation effects on satellite components resulting difficulties in satellite communication between the ground and the onboard computer
An approach to identify TID radiation effects in the presence of noise in RF circuits with very high accuracy is achieved through the proposed IRES technique. The IRES technique is also utilized for identification of spurious transient events in PLL circuit. IRES is based on RF-DNA fingerprinting, a technique for identification of different devices based on statistical features extracted from their transmitted waveforms. The IRES technique for TID uses only two statistical features (mean and standard deviation of the instantaneous frequency) to identify the operational health of the circuit. LD models were used to classify the TID radiation level as well input bias voltage of the VCO and PLL
? IRES is based on Radio Frequency-Distinct Native Attribute (RF-DNA) fingerprinting for the identification of essential statistical features of erroneous signals within a device, circuit, or system. RFDNA fingerprinting is a waveform-based technique used for augmenting existing wireless network security mechanisms.
? The PLL model was designed with Verilog-A to match the circuit dynamics of the DUT used for experimental validation.
? SET simulations were performed with the Cadence Spectre circuit simulator.
|