Arizona State University Orange Mall Green Infrastructure Project Methods

Abstract

A fundamental challenge of autonomous driving is maintaining the vehicle in the center of the lane by adjusting the steering angle. Recent advances leverage deep neural networks to predict steering decisions directly from images captured by the car cameras. Machine learning-based steering angle prediction needs to consider the vehicle's limitation in uploading large amounts of potentially private data for model training. Federated learning can address these constraints by enabling multiple vehicles to collaboratively train a global model without sharing their private data, but it is difficult to achieve good accuracy as the data distribution is often non-i.i.d. across the vehicles. This paper presents a new confidence-based federated distillation method to improve the performance of federated learning for steering angle prediction. Specifically, it proposes the novel use of entropy to determine the predictive confidence of each local model, and then selects the most confident local model as the teacher to guide the learning of the global model. A comprehensive evaluation of vision-based lane centering shows that the proposed approach can outperform FedAvg and FedDF by 11.3% and 9%, respectively.

Experimental Study and Parameter Optimization of Oily Sludge Treatment Through Ozonation Technology

“…[37,38] The direct oxidation of ozone molecules with organic matter is highly selective, O 3 attacks hydrocarbons via molecule addition, which leads to bond breaking. [39] Alkanes can be oxidized by O 3 into alcohols, ketones, and carboxylic acids. Ketones and carboxylic acids are attacked by * OH and react with O 2 to form CO 2 and H 2 O.…”

MVDC Railway Traction Power Systems; State-of-the Art, Opportunities, and Challenges

“…The complexities involved when implementing bidirectional AC/DC converters from CHB and CNPC inspired the emergence of MMCs [71]. MMCs are characterized by increased modularity, scalability to any power or voltage level, filter-free high-power-quality operation, transformerless operation, high efficiency, fault tolerance capability, the absence of DC link capacitance, and lower electromagnetic interference [73,74]. Figure 6 shows a generalized schematic diagram of a three-phase MMC.…”

“…Consequently, during DC faults, HB SMs cannot block fault currents on the AC side. Thus, opportunities for other SMs such as full bridge (FB), unipolar full bridge (UFB), clamped double circuits (CDC), and the three-level converter circuit (3LCC)-three-level NPC or three-level FC-as well as the five-level cross-connected (5LCC) SM arose to improve the MMC's fault-blocking capability [71][72][73][74]. Table 6 illustrates the comparison of the salient features of the MMC SM topologies in terms of rating, operation, losses, design, and control complexity, as well as DC fault-blocking capability.…”

“…Over the years, appropriate control/modulation schemes have emerged-as reviewed in [73,75]-to address MMC operational issues. Additionally, other issues such as modelling, SM topologies, fault diagnosis/tolerance, and protection have been investigated widely towards realizing an efficient MMC-based system [74][75][76].…”

“…A lot of research [73,76,81] has been undertaken on MMC modelling, topologies, control schemes, modulation strategies, fault diagnosis/tolerance, protection, and related challenges. From the analysis, some of the challenges include the unavailability of accurate and efficient SM/MMC models and limited insight into MMCs' internal dynamics, reliability, and DC protection issues-research on which is still in progress [74,75,82].…”