Abstract This paper presents an AI-assisted inverse design methodology for a compact, ultra wideband microstrip bandpass filter (BPF) that eliminates the need for conventional resonator topologies. Unlike traditional approaches relying on stepped-impedance resonators, stubs, or defected ground structures,... This paper presents an AI-assisted inverse design methodology for a compact, ultra wideband microstrip bandpass filter (BPF) that eliminates the need for conventional resonator topologies. Unlike traditional approaches relying on stepped-impedance resonators, stubs, or defected ground structures, the proposed design employs a pixelated conductive region optimized directly from electromagnetic performance objectives using artificial intelligence. The filter architecture consists of two 50- microstrip feedlines coupled through a 10 mm × 10 mm square domain discretized into a 20 × 20 binary pixel grid, fabricated on a Rogers RO4350B substrate. A convolutional neural network surrogate model accelerates the optimization process by over two orders of magnitude, enabling efficient exploration of the topology space while enforcing fabrication constraints such as minimum feature size and metal continuity. Full-wave simulations demonstrate a continuous passband from 3 to 10 GHz (fractional bandwidth >100%), in-band insertion loss below 1.2 dB, return loss better than 15 dB, and stopband rejection exceeding 40 dB. Parametric sensitivity analyses confirm robustness against ±5% variation in pixel fill ratio, ±3% deviation in substrate permittivity, and typical feedline width tolerances highlighting practical manufacturability. The AI-discovered geometry intrinsically integrates resonance generation, impedance matching, and transmission zero formation within a single planar region, achieving high performance in an electrically compact footprint (<0.25g × 0.25 g). This work establishes inverse electromagnetic design as a scalable and powerful alternative to classical synthesis techniques, with potential extensions to multi band and reconfigurable RF front end components. Read More Read Less

Abstract This work establishes a principled framework for contextual intelligence in autonomous scene understanding, moving beyond isolated object detection toward holistic environmental interpretation. Leveraging the nuImages dataset, we quantify semantic dependencies through four integrated components: (1) object– object relationship... This work establishes a principled framework for contextual intelligence in autonomous scene understanding, moving beyond isolated object detection toward holistic environmental interpretation. Leveraging the nuImages dataset, we quantify semantic dependencies through four integrated components: (1) object– object relationship graphs revealing statistically significant co occurrences (e.g., pedestrian crosswalk correlation of 0.91); (2) conditional probability models demonstrating how contextual cues modulate object presence sidewalks elevate pedestrian likelihood to 0.70 versus a marginal baseline of 0.34, while highways suppress it to 0.08; (3) a multi dimensional Scene Complexity Index (SCI) that stratifies environments by semantic density and interaction potential, yielding medians of 8.1 (urban), 5.2 (residential), and 2.3 (highway); and (4) context aware risk scoring that quantifies perception uncertainty across scene typologies (urban: 0.53; residential: 0.33; highway: 0.08). Critically, we establish a near deterministic positive correlation between scene complexity and perception risk (r = 0.977, R² = 0.954, p < 0.001), validating that environmental semantics directly govern detection uncertainty. These findings substantiate a paradigm shift toward context adaptive perception architectures that proactively allocate computational resources, adjust detection thresholds, and anticipate hazards based on environmental typology. By formalizing contextual priors through statistical models rather than heuristic rules, our framework enhances robustness in safety critical autonomous applications particularly in complex urban environments where semantic ambiguity and multi agent interactions dominate failure modes. Read More Read Less

Abstract The increasing penetration of high power residential appliances and renewable energy sources has intensified peak demand challenges in modern power systems. Demand Side Management (DSM), particularly through direct load control (DLC), is a key enabler of smart... The increasing penetration of high power residential appliances and renewable energy sources has intensified peak demand challenges in modern power systems. Demand Side Management (DSM), particularly through direct load control (DLC), is a key enabler of smart grids for mitigating peak loads while maintaining user comfort. This paper presents a fuzzy logic based Fuzzy Decision Maker (FDM) for optimal scheduling of timeshiftable residential loads. Unlike conventional approaches that rely solely on price signals or individual load characteristics, the proposed FDM jointly considers appliance demand power and the predicted aggregate residential load profile to determine an optimal delay time. The system emulates expert operator reasoning through interpretable linguistic rules and operates under a DLC paradigm with customer consent and incentives. Simulation results for a residential community demonstrate peak load reduction and a lower Peak to Average Ratio (PAR), thereby validating the effectiveness and scalability of the proposed approach for smart grid applications. Read More Read Less