Abstract

The radiant systems consume less energy and provide better thermal comfort as compared to the conventional system, but the main drawback of radiant technology is inability to cater the latent load and condensation. This problem has engaged researchers to push their emphasis on Personalized Conditioning System (PCS). It forms a micro-climate region near the occupant. This study is aimed at PCS combined with a radiant conditioning system, which evaluates the performance of the PCS and compares it with different radiant cooling systems. A radiant cubicle was used to investigate the performance of the PCS for both heating & cooling modes. For heating mode, the Standard Effective Temperature (SET) was calculated with the help of thermal comfort surveys which were in a range of 23.4°C to 27.3°C. For cooling mode, four different cases compared the performance of the system which are: Conventional Air Conditioning System (CACS), Radiant Cooled Ceiling System (RCCS), Personalized Radiant Conditioning System (PRCS), and both PRCS-RCCS coupled. It was observed that PRCS alone reaches the thermal comfort criteria with low energy consumption. The energy-saving of RCCS-PRCS was achieved 38% when compared with CACS. It can be concluded that PRCS is a new step in the field of thermal comfort.

Abstract

Standard accuracy metrics have shown that Math Word Problem (MWP) solvers have achieved high performance on benchmark datasets. However, the extent to which existing MWP solvers truly understand language and its relation with numbers is still unclear. In this paper, we generate adversarial attacks to evaluate the robustness of state-of-the-art MWP solvers. We propose two methods Question Reordering and Sentence Paraphrasing to generate adversarial attacks. We conduct experiments across three neural MWP solvers over two benchmark datasets. On average, our attack method is able to reduce the accuracy of MWP solvers by over 40 percentage points on these datasets. Our results demonstrate that existing MWP solvers are sensitive to linguistic variations in the problem text. We verify the validity and quality of generated adversarial examples through human evaluation.

Abstract

Existing black box search methods have achieved high success rate in generating adversarial attacks against NLP models. However, such search methods are inefficient as they do not consider the amount of queries required to generate adversarial attacks. Also, prior attacks do not maintain a consistent search space while comparing different search methods. In this paper, we propose a query efficient attack strategy to generate plausible adversarial examples on text classification and entailment tasks. Our attack jointly leverages attention mechanism and locality sensitive hashing (LSH) to reduce the query count. We demonstrate the efficacy of our approach by comparing our attack with four baselines across three different search spaces. Further, we benchmark our results across the same search space used in prior attacks. In comparison to attacks proposed, on an average, we are able to reduce the query count by 75% across all datasets and target models. We also demonstrate that our attack achieves a higher success rate when compared to prior attacks in a limited query setting.

Abstract

We study an important and challenging task of attacking natural language processing models in a hard label black box setting. We propose a decision-based attack strategy that crafts high quality adversarial examples on text classification and entailment tasks. Our proposed attack strategy leverages population-based optimization algorithm to craft plausible and semantically similar adversarial examples by observing only the top label predicted by the target model. At each iteration, the optimization procedure allow word replacements that maximizes the overall semantic similarity between the original and the adversarial text. Further, our approach does not rely on using substitute models or any kind of training data. We demonstrate the efficacy of our proposed approach through extensive experimentation and ablation studies on five state-of-the-art target models across seven benchmark datasets. In comparison to attacks proposed in prior literature, we are able to achieve a higher success rate with lower word perturbation percentage that too in a highly restricted setting.

Abstract

The growing need for localization has created an array of alternative approaches to GNSS, based on Wi-Fi, Bluetooth, Ultra-Wideband, etc. Long Range Wide Area Network (LoRaWAN) is one such technology that has garnered tremendous attention due to its low power and long-range capabilities. Many attempts to achieve localization using LoRaWAN have been made till now, based on a variety of techniques such as Angle of Arrival (AoA), Time Difference of Arrival (TDoA), Received Signal Strength Index (RSSI). In this paper, we present a novel, standalone localization approach by developing a collaborative, TDoA-based methodology using LoRaWAN. The server determines the target node location by means of TDoA measurements from the target node to the gateways. The introduction of an additional stationary node allows the synchronization of the gateways without utilizing GNSS, either inbuilt or external. The cooperation between target node, synchronization node and server, is the innovative feature which makes this approach attractive for GNSS-free localization. Further, we explore the effects of timing resolution, time-on-air, and duty cycle constraints on the localization error. Finally, the distribution of error in a triangle of gateways situated approximately 8.6 kilometers apart is simulated in ideal Line of Sight (LoS) conditions, showing the maximum error to be around 23 meters

Abstract

Methods and apparatus for modulating light using a tunable light modulation device . The tunable light modulation devices comprises an elastomer structure including at least one elastomer layer , a compliant electrode network of con ducting fibers arranged on a first surface of the at least one elastomer layer , a patterned electric conductor arranged on a second surface of the at least one elastomer layer opposite the first surface . The patterned electric conductor includes a plurality of individually - addressable sections , and the com pliant electrode network is configured to compress the at least one elastomer layer in the presence of an electric field between the compliant electrode network and one or more of the individually - addressable sections of the patterned elec tric conductor to produce a voltage - dependent roughening of the at least one elastomer layer .

Abstract

Muscle weakening is a common consequence of stroke and can result in a reduction in physical activity of the affected body part. Therapy includes a range of physical exercises that will help the patients restore and build physical strength, endurance, flexibility, balance, and stability. In order to analyze and recognize the activity and movement of hands while performing these exercises, we have developed a 4 × 4 flexible pressure sensor matrix to quantize the performance and progress of a patient undergoing physiotherapy. We have also developed an artificial intelligence (AI) based algorithm to determine the accuracy of positioning by the patients. Experimental results demonstrate that the algorithm gives a mean error of 0.103 cm in detecting the position of load, compared to a mean error of 0.704 cm using mathematical analysis. With this system, a patient can be asked to move a weight to a particular location (as part of regular physiotherapy) and the pressure sensor matrix can be used to calculate error in positioning along with the time taken to complete the task. Advantages of the proposed pressure sensor matrix are cost effectiveness, facile fabrication, high sensitivity, robustness and flexibility

Abstract

An Electric Vehicle (EV) charger, also called Electric Vehicle Charging Station (EVCS) is an infrastructure element that supplies electrical energy for recharging EVs. The paper focuses on the design and fabrication of two-wheeler (escooter) electric vehicle charging equipment. Considering the future scenario of mass privatization of EV two-wheelers on Indian roads, the outline discusses the product design based on fabricability, affordability, and ability to mass manufacture. The proposed architecture follows the Level 2 charging standards (240 Volts), and is based on the open charge point protocol (OCPP). The suggested EVCS is constructed considering the safety prerequisites of system administrators, installers, consumers, government agencies and others. The design of EVCS links to three industries: equipment manufacturers, software industry and electric power networks. This paper presents design considerations by elaborating the hardware, software, and protocols followed to design the Level 2 charging standard EVCS. Index Terms—Electric Vehicle (EV), Electric Vehicle Charge Station (EVCS), Web Sockets, Open Charge Point Protocol (OCPP), Level 2 EV Charging

Abstract

This paper presents an ultra-low power CMOS voltage reference (CVR) which is free of amplifier and resistor. A current generator circuit is used to generate a supply independent current to bias the active load in the temperature compensation circuit. Drain-source voltage of two critical MOSFETs is made equal in the current generator circuit by using a feedback arrangement to ensure a high PSRR of -75dB for the reference voltage. Temperature compensation is achieved by using the complementary to absolute temperature (CTAT) nature of gatesource of a MOSFET operating in the subthreshold region and proportional to absolute temperature (PTAT) nature of conventional composite pair architecture. The proposed CVR is designed in TSMC 180nm technology. The circuit works desirably for a supply range of 0.85V to 2.3V while generating a reference voltage of around 0.68V. Maximum temperature coefficient (TC) of 184 ppm/˚C and minimum TC of 69 ppm/˚C are noted for mismatch and process variations in the Monte-Carlo simulation for 1000 samples. The circuit consumes only 46nW of power, which makes it suitable for ultra-low power applications.

Abstract

The present description relates to a method based on artifi cial intelligence to implement a wide range of microelec tronic circuits that can adapt by themselves to the usage conditions ( e.g. loading changes ) , manufacturing variances or defects ( e.g. process variations , device parameter mis matches , device model inaccuracies or changes , etc. ) , as well as environmental conditions ( e.g. voltage , temperature , interference ) in order to negate all or part of their effects on the circuit performance characteristics and achieve a very tight set of specifications over the wide range of conditions . Each microelectronic circuit is represented by a neural network model whose behavior is a function of the actual input signals , the usage and environmental conditions . An attached AI engine will infer from the model , the input signals , the usage conditions and the environmental condi tions and create the adaptive changes required to modify the microelectronic circuit's behavior to negate all or part of their effects on the circuit performance characteristics and to achieve a very tight set of specifications