Air Quality (AQ) and the management of low-emission zones are critical issues in densely populated urban areas. In such environments, human activity significantly impacts AQ, prompting increased efforts to monitor it using a range of devices. Traditional Air Quality monitoring relies on regulated stations, which are often scarce due to high costs, leaving many areas unmonitored. Low-cost sensors offer a promising solution by enabling the higher-spatial-resolution monitoring of pollution levels. In this article, we present the results of a case study conducted in an urban setting where AQ is affected by human activity, particularly during Las Fallas, Valencia’s most renowned festival, which has been declared an Intangible Cultural Heritage of Humanity by UNESCO. The festival features widespread bonfires, firecrackers and large crowds, all of which contribute to worsening air pollution. In this context, we evaluate the performance of the off-the-shelf, low-cost ZPHS01B multisensor module in a real deployment. This module is capable of monitoring Temperature (T), Relative Humidity (RH), Particulate Matter (PM), CO, ${\mathrm{CO}}_2$, ${\mathrm{NO}}_2$, ${\mathrm{O}}_3$, ${\mathrm{CH}}_2$O and Volatile Organic Compounds. We analyze the features and properties of these sensors. In our deployments, the ZPHS01B module is connected to an ESP32 microcontroller and assembled into an AQ Internet of Things (IoT) node. We present AQ monitoring results from the festival and compare the measurements with those from regulated AQ monitoring stations, used as a reference. Additionally, we evaluate the power consumption of this AQ IoT node, providing its electrical operating characteristics and considering the use of duty cycles to reduce consumption while maintaining sensor stability. We conclude that this module offers promising capabilities for identifying pollution risk zones and opens the door to new research opportunities, particularly in efficient sensor calibration and AQ parameter prediction.

The evolution toward 6G will continue to leverage massive multiple-input multiple-output and millimeter-wave systems, which demand accurate angle-of-arrival (AoA) and angle-of-departure (AoD) estimation. While several deep learning models have demonstrated strong performance for this task, their accuracy, like that of most estimation methods, is often degraded by hardware non-idealities, which can be further exacerbated by time-varying operational factors such as component aging and adverse weather, among others. Building on a pre-trained U-Net architecture with demonstrated competitive performance for AoA/AoD estimation, we first propose an adaptation mechanism based on fine-tuning with impairment-augmented data. Specifically, we simulate hardware imperfections by introducing random phase errors in the antenna elements, ranging from mild fluctuations to severe signal distortions. The U-Net model with adaptation capabilities is then implemented on an NVIDIA Jetson Orin Nano device, a compact edge platform with heterogeneous computing resources. To this end, we design a co-execution strategy that performs AoA/AoD estimation (inference) on the CPU while simultaneously fine-tuning the model on the GPU, thus enabling continuous model adaptation to changing environmental or hardware conditions while preserving real-time inference performance. Experimental results show that impairment-aware fine-tuning effectively counters hardware degradation, particularly under significant phase impairments. In such scenarios, the fine-tuned model consistently preserves or even improves estimation accuracy, reducing the Root Mean Square Error (RMSE) by approximately 3.6% and increasing the Probability of Detection ( $$P_D$$ P D ) by up to 1 percentage point compared to the base model. Furthermore, a detailed energy-performance analysis demonstrates that while maximum frequency settings reduce training time by over 11 $$\times $$ × , they also increase power consumption by more than 5 $$\times $$ × , with optimal energy efficiency achieved at mid-range CPU and high GPU frequencies. This work establishes the feasibility of concurrent training and inference on resource-constrained heterogeneous hardware, paving the way for resilient and autonomous edge intelligence in future 6G systems.