Om Kathalkar

Industrial autonomous mobile robots increasingly depend on reliable wireless connectivity for real-time control and data streaming, yet existing path planning methods fail to account for the complex interplay between spatial geometries and material-dependent signal propagation across heterogeneous facility environments. This paper presents a spatial-aware transformer framework for network quality prediction that explicitly encodes both geometric layouts and electromagnetic properties (permittivity, conductivity, reflection coefficients, absorption coefficients) of building materials to enable accurate signal strength, throughput, and latency predictions across diverse industrial facilities. The framework represents environments through declarative YAML specifications capturing wall geometry and validated ITU-R material properties, eliminating exhaustive physical site surveys. Rigorous validation against the iV2I+ real-world measurement dataset established Sionna ray-tracing simulation fidelity with R2 = 0.87 for SNR predictions, providing confidence in physics-based ground truth for supervised learning. Comprehensive evaluation across 8,840 test paths spanning concrete, wood, and metal environments with varying geometric configurations demonstrated SNR prediction RMSE of 2.31 dB with R2 = 0.891, representing 19.5% improvement over state-of-the-art baselines. Critically, zero-shot evaluation on mixed-material scenarios and held-out geometric layouts—combining spatial and material configurations never jointly observed during training— yielded only 10% performance degradation, eliminating facility-specific retraining requirements. Ablation studies confirmed spatial-material encoding’s essential contribution with 47% accuracy improvement over encoding-agnostic variants, while computational efficiency analysis revealed 73 ms inference time achieving 31× speedup compared to graph neural network baselines. The framework enables train-once-deploy-everywhere paradigm for industrial system integrators, with single trained models serving warehouses, manufacturing floors, and assembly areas without site-specific customization.

Autonomous Mobile Robots (AMRs) in industrial environments require reliable wireless connectivity for coordination, control, and safety operations. Traditional path planning algorithms focus solely on geometric constraints, often leading robots through areas with poor network coverage that can compromise mission-critical operations. This paper presents a comprehensive framework for network-aware path planning that incorporates wireless network quality metrics as path constraints. We validate our Sionna-based ray-tracing simulations against real-world measurements from the Hernangomez et al. iV2I+ dataset, achieving strong correlation between simulated and real-world measurements (R² = 0.87 for SNR, 0.82 for throughput). Using this validated simulation framework, we implement three novel path planning algorithms: A* with network constraints, conditional variational autoencoder (CVAE)-based neural path planning, and Graph neural network-based multi-path (GraphMP) planning. Our evaluation demonstrates trade-offs between network quality requirements and path efficiency, with CVAE achieving 95.2% constraint satisfaction and GraphMP showing 23% shorter planning times. We provide practical guidelines for selecting appropriate algorithms and network quality thresholds based on application requirements, enabling more reliable AMR operations in industrial settings.

Air pollution represents one of the most critical environmental and public health challenges globally, with traditional sensor-based monitoring systems facing significant scalability and economic constraints. Image-based air quality estimation has emerged as a promising alternative, leveraging the visual characteristics of atmospheric pollutants in traffic scenes. However, existing methods suffer from limited cross-city generalization and inadequate exploitation of multi-view perspectives. We present AQIFormer, a novel transformer-based ensemble architecture that addresses these fundamental limitations through innovative dual-view integration, weather-aware attention mechanisms, and comprehensive multi-task learning. Our approach uniquely combines front and rear traffic imagery with meteorological parameters to achieve robust air quality classification across diverse urban environments. Extensive evaluation on a dataset of 26,678 synchronized front–rear image pairs demonstrates 89.96% accuracy, representing a 14.96% improvement over state-of-the-art methods. Most importantly, our model maintains exceptional cross-city generalization capabilities, achieving 81.67% accuracy on an independent dataset collected in Nagpur, India, with only 8.29% performance degradation using few-shot adaptation and minimal training samples.

Air quality estimation through sensor-based methods is widely used. Nevertheless, their frequent failures and maintenance challenges constrain the scalability of air pollution monitoring efforts. Recently, it has been demonstrated that air quality estimation can be done using image-based methods. These methods offer several advantages including ease of use, scalability, and low cost. However, the accuracy of these methods hinges significantly on the diversity and magnitude of the dataset utilized. The advancement of air quality estimation through image analysis has been limited due to the lack of available datasets. Addressing this gap, we present TRAQID - Traffic-Related Air Quality Image Dataset, a novel dataset capturing 26,678 front and rear images of traffic alongside co-located weather parameters, multiple levels of Particulate Matters (PM) and Air Quality Index (AQI) values. Spanning over multiple seasons, with over 70 hours of data collection in the twin cities of Hyderabad and Secunderabad, India, the TRAQID offers diverse day and night imagery amid unstructured traffic conditions, encompassing six AQI categories ranging from “Good” to “Severe”. State-of-the-art air quality estimation techniques, which were trained on a smaller and less-diverse dataset, showed poor results on the dataset presented in this paper. TRAQID models various uncertainty types, including seasonal changes, unstructured traffic patterns, and lighting conditions. The information from the two views (front and rear) of the traffic can be combined to improve the estimation performance in such challenging conditions. As such, the TRAQID serves as a benchmark for image-based air quality estimation tasks and AQI prediction, given its diversity and magnitude.