Search

We tackle a multi-layer knowledge distillation problem between deep models with heterogeneous architectures. The main challenges of that are the mismatches of the feature maps in terms of the resolution or semantic levels. To resolve this, we propose a novel transformer-based multi-layer correlation knowledge distillation (TMC-KD) method in order to bridge the knowledge gap between a pair of networks. Our method aims to narrow the relational knowledge gaps between teacher and student models by minimizing the local and global feature correlations. Based on extensive comparisons with the recent KD methods on classification and detection tasks, we prove the effectiveness and usefulness of our TMC-KD method.

Artificial neural networks (ANNs) have shown several benefits over the traditional classification methods for radiation detector data, such as greater accuracy and the ability to classify neutron-photon combinations from piled-up events. These capabilities are of particular interest in applications involving intense radiation environments where large instantaneous detector count rates can lead to many piled-up events and subsequent information loss. The recovery of individual radiation detector pulses from piled-up data can improve the efficiency of classification systems, making them more attractive for field applications. This work extends the use of ANN systems with piled-up recovery to the real-time domain, with a focus on the hardware implementation. The ANN system is implemented on a Virtex-5 XC5VSX95T FPGA which collects pulse data at 250MHz and then processes and classifies pulses in the pipelined ANN at lower frequencies. The system’s performance is demonstrated by classifying pulses in real-time in a variety of scenarios including passive background, passive plutonium-beryllium (PuBe), and active PuBe. The results show that the system can provide accurate classifications in real-time while displaying the results clearly to the user. The system is shown to be capable of classifying pulses at a maximum rate of $1.11\times 10 ^{6}$ pulses per second, with a maximum latency of $7.7~\mu \text{s}$ , and an overall accuracy of 98.2%.

A robust and high-performance lateral control system is necessary for autonomous bus driving in complex downtown roads. In this paper, a Sliding Mode (SM) controller with an uncertainty observer is proposed to reduce the chattering phenomenon while ensuring robustness and maintaining the control performance. A stiffness gain is introduced in the cornering stiffness of rear wheels to compensate for the model uncertainty and external disturbance. For lane keeping and changing maneuvers, an optimal path planning algorithm is designed by considering dynamic constraints of the bus. The lane departure of bus is prevented with minimizing the lateral jerk through the sampling-based path regeneration process. The proposed lateral control system is compared to other conventional SM based controllers in the TRUCKSIM simulator. In an experiment with a real bus, lane keeping maneuver is conducted and compared with experienced human driver in the same driving course. Lane change maneuver is also tested with real bus to perform smooth lane transition while satisfying the lateral acceleration and yaw rate limitations.

To make mobile robot move considerately, an adaptive motion control method integrated potential field method is proposed. Risk map is firstly defined to show the magnitude of risks in the empty spaces around objects, then the considerate path is generated based on based on risk map. The robot is controlled by using integrated potential field method. The robot can automatically move with a considerate path to avoid disturbing others without make multiple moving rules, and areas with medium level risk are still passable, when there is no better choice. Besides, the robot will still make the users feel conformable by using effective interactions, like moving slowly to make sure safety or warning the users by sound. By comparing the experimental results with those controlled by conventional potential field method, the proposed method can control the robot move more adaptively and considerately with appropriate interactions.

Guide mobile robot have been researched and developed for decades. However, current robot cannot guide the user to the destination considerately, since the status of the user and the properties of the obstacles are not considered, especially for the guiding work of servicing visually impaired people. In this paper, a guide mobile robot with an easy-to-hold handle is designed, a generation method of spatial risk map is proposed to evaluate the influences of potential spaces of objects and a motion control method based on spatial risk map considering potential occupied spaces of objects is proposed. The users are successfully guided to the destination naturally without influencing other pedestrians by avoiding entering the potential spaces of objects automatically, and they can adjust their moving status in their own will since the considerate robot can adaptively adjust its guide status to adapt to the user. Through comparing guiding experiments of different tasks for 10 blindfolded users, the proposed guide mobile robot is proven considerate by guiding the users adaptively and make other pedestrians feel comfortable by avoiding entering the potential spaces of objects automatically.

Active interrogation (AI) is a promising technique to detect shielded special nuclear materials (SNMs). At the University of Michigan, we are developing a photon-based AI system that uses bremsstrahlung radiation from an electron linear accelerator (linac) as an ionizing source and trans-stilbene organic scintillating detectors for neutron detection. Stilbene scintillators are sensitive to fast neutrons and photons and have excellent pulse shape discrimination (PSD) capabilities. The traditional charge integration (CI) method commonly used for PSD analysis eliminates piled-up pulses and relies on a particle discrimination line to separate neutrons and photons. The presence of the intense photon flux during AI creates a significant number of piled-up events in the stilbene scintillator, thereby posing a great challenge to the traditional CI method. Identifying true single neutron pulses becomes challenging due to the presence of a pile-up cloud and overlapping neutron, photon and pile-up clouds in the PSD analysis. To mitigate the effect of pulse pile-up and identify true single neutron pulses from stilbene scintillators, an artificial neural network (ANN) system is developed. The developed ANN system identifies single neutron pulses and neutron-photon combinations from piled-up events. The results obtained from a 252Cf measurement in the presence of the intense photon flux show that the developed ANN system outperforms the traditional CI method. Since many piled-up events lie above the particle discrimination line, they get misclassified as neutrons by the traditional CI method resulting in 27% overestimation of the net neutron count rate during the linac pulse. The overall net neutron count rate (single and restored neutrons) during the linac pulse, estimated by the ANN system is 62.32% of the ground truth. Energy spectroscopy of the ANN attributed single neutron pulses further provides evidence on the detection of prompt fission neutrons from the 252Cf fission source.

There are many challenges in measuring capacity using metrics such as transactions per minute (TPM) and operation per second (OPS) for all server hardware, which are becoming increasingly obsolete due to the shortening of the lifecycle of hardware and the advent of microprocessors. Instead, the results of accredited performance measurements are used to measure these standards, which are further used as references in the capacity measuring methods employed by industries. Generally, in industries, the capacity of a web application server is defined by OPS, for which no clear transition criterion exists for calculating tpmC using an empirical verification method. Considering secure Unix to Linux (U2L) x86-based server migration, there are no methods to compare and verify the max-jOPS value of Standard Performance Evaluation Corp., which is an industry-recognized performance measurement standard, to the Unix-based benchmark tpmC. In this study, a scenario-based U2L migration was empirically verified by analyzing and comparing pre-to-post with the interpretation of a census statistical system log data, which was conducted on approximately 1.7M households over 21 days with 25,288 maximum concurrent users. We present the correlation through pre-to-post comparison and analysis of U2L for each census statistical system log data by measuring the maximum CPU utilization as U2L migration between heterogeneous CPUs. The correlation is applied to OPS using the tpmC value of TPC-C proportional equation and quantified as a derived conversion ratio of OPS to max-jOPS. Consequently, we formulated and normalized an arithmetic expression, resulting in a CPU core conversion ratio of 0.165 as facile from a Unix-based legacy platform to an x86-based server. Therefore, we propose a new server sizing model for secure U2L migration between heterogeneous CPU architectures, which results in an average of 14.3% improvement in data processing time.

Abstract This paper proposes a reclosing scheme using synchronism checking for utilization of battery energy storage system (BESS) in a distribution system. The algorithm disconnects the faulty phase and keeps the power supply from the BESS to the healthy phase. Synchronism checking between the main source side and the load side is applied to minimize the transients at the reclosing instant. The BESS at the faulty phase is reconnected after checking the successful reclosing. The distribution system including BESS and proposed reclosing scheme are modeled by the electromagnetic transients program (EMTP)/ATPDraw. The various simulations by varing the fault clearing time are conducted and the simulation results are discussed. Also, the relation between proposed reclosing scheme and reliability is discussed.

Through silicon vias (TSVs) are a critical technology for manufacturing three-dimensional stacked structure of semiconductor packages by forming holes that penetrate silicon wafers and vertically interconnect multiple wafers. Typically, TSVs are created by drilling via holes in wafers and filling their interiors using copper electroplating processes. Subsequently, the wafers are exposed to high-temperature environments during the back-end-of-line (BEOL) process. However, improper copper electroplating conditions can form defects, such as voids and seams, within TSVs, while the high temperature of the BEOL process induces copper protrusion phenomena. These defects and copper protrusion degrade the reliability of TSV. In this brief, copper protrusion behavior, which is a direct cause of reliability degradation in TSVs, was mitigated by experimentally exploring the seam defects that can occur during the TSV filling process. Subsequent annealing processes were applied to remove the seam defects based on the copper-grain growth. The copper protrusion height was analyzed based on the size of the seam defects and annealing temperature. From the proposed process in this brief, the copper protrusion heights of TSVs without and with seam defects were confirmed to be 1.531 and $1.289~\mu \text{m}$ , respectively, representing an improvement of approximately 15.81%. [ABSTRACT FROM AUTHOR]