We present a practical verification method for safety analysis of the autonomous driving system (ADS). The main idea is to build a surrogate model that quantitatively depicts the behavior of an ADS in the specified traffic scenario. The safety properties proved in the resulting surrogate model apply to the original ADS with a probabilistic guarantee. Given the complexity of a traffic scenario in autonomous driving, our approach further partitions the parameter space of a traffic scenario for the ADS into safe sub-spaces with varying levels of guarantees and unsafe sub-spaces with confirmed counter-examples. Innovatively, the partitioning is based on a branching algorithm that features explainable AI methods. We demonstrate the utility of the proposed approach by evaluating safety properties on the state-of-the-art ADS Interfuser, with a variety of simulated traffic scenarios, and we show that our approach and existing ADS testing work complement each other. We certify five safe scenarios from the verification results and find out three sneaky behavior discrepancies in Interfuser which can hardly be detected by safety testing approaches.

Transfer learning has been highlighted as a promising framework to increase the accuracy of the data-driven model in the case of data sparsity, specifically by leveraging pretrained knowledge to the training of the target model. The objective of this study is to evaluate whether the number of requisite training samples can be reduced with the use of various transfer learning models for predicting, for example, the chemical source terms of the data-driven reduced-order modeling (ROM) that represents the homogeneous ignition of a hydrogen/air mixture. Principal component analysis is applied to reduce the dimensionality of the hydrogen/air mixture in composition space. Artificial neural networks (ANNs) are used to regress the reaction rates of principal components, and subsequently, a system of ordinary differential equations is solved. As the number of training samples decreases in the target task, the ROM fails to predict the ignition evolution of a hydrogen/air mixture. Three transfer learning strategies are then applied to the training of the ANN model with a sparse dataset. The performance of the ROM with a sparse dataset is remarkably enhanced if the training of the ANN model is restricted by a regularization term that controls the degree of knowledge transfer from source to target tasks. To this end, a novel transfer learning method is introduced, Parameter control via Partial Initialization and Regularization (PaPIR), whereby the amount of knowledge transferred is systemically adjusted in terms of the initialization and regularization schemes of the ANN model in the target task.

This paper presents a systematic design approach for developing a semiflexible multiple-input–multiple-output antenna system operating in the millimeter wave frequency spectrum, specifically designed for body-worn applications in biotechnologies. The designed antenna features dual flower-shaped antenna radiators placed in a spatial diversity configuration. Strategic modifications have been implemented by integrating dual crescent-shaped slots in the ground layer to attain the targeted frequency band of 25.7–30.6 GHz. Later, the upper edge of the ground plane is truncated in order to achieve circularly polarized radiation characteristics at 29.4 GHz with 3 dB ARBW of 0.6 GHz (29.1–29.7 GHz). The realization of circular polarization in the antenna geometry is validated through the analysis of characteristic mode theory. A maximum gain of 5.6 dBi is attained along with a port isolation of >30 dB. The proposed antenna undergoes analysis to assess its performance in the bending conditions and specific absorption rate, besides validation of diversity metrics encompassing envelope correlation coefficient, diversity gain, channel capacity loss, total active reflection coefficient, and mean effective gain has also been conducted. Finally, the proposed antenna structure is fabricated, and its performance is validated and subsequently compared with that of its simulated counterpart.

In this paper the performance of an internal strut-based thrust vector control (TVC) system has been studied at different expansion conditions of propulsion nozzle. The TVC system uses a cylindrical strut inserted through the diverging wall of a supersonic nozzle. This TVC system can be construed as an alternative to secondary injection TVC method. The nozzle had an expansion ratio of 1.545 and nozzle pressure ratio (NPR) of 6.61 for optimum expansion. Numerical simulations were performed at over-expansion (NPR = 3.94) and under-expansion (NPR = 7.89) conditions for four strut locations (xs) and five strut heights (hs). The strut location from the nozzle throat corresponded to 33%, 50%, 66.7% and 80% of the diverging length (Ld) of the nozzle. The schlieren images of the nozzle exhaust and nozzle internal wall pressure distribution from experiments were compared with the results from numerical simulation and the agreement was quite good. Computational results show that introduction of the strut caused a maximum total pressure loss of 1.5% at its maximum height. The calculations also show that $ \pm $4${}^{\circ}$ thrust deflection angle could be achieved using combinations of strut location and strut height over a range of nozzle operational conditions. Thrust vectoring performance of strut insertion TVC was evaluated using a parameter called vectoring performance index (VPI) defined as thrust deflection angle per unit percentage of pressure loss. The maximum VPI was observed when xs=0.5Ld at ${\bar{h_s}} = 0.429$ in both over-expansion and under-expansion conditions. The study reveals that an internal strut based TVC has a good future potential to be developed as an alternate TVC system obviating the requirement of carrying a fluid tank for a system like secondary injection TVC.