Increasing demands on railway structures have led to a need for new cost-effective maintenance strategies in recent years. Current dynamic railway track monitoring systems are usually based on the analysis of axle-box accelerations to automatically detect track singularities and defects. These methods rely on hand-crafted feature extraction and classifiers for different tasks. However, the low performance shown in previous literature makes it necessary to complement these analyses with in-situ inspections. Very recent works have proposed the use of deep learning systems that allow extracting more generalizable features from time–frequency spectrograms. However, the lack of specific public domain datasets and the finite number of track singularities in a railway structure have limited the development of deep learning based systems. In this paper, we propose a method capable of outstanding in low-data scenarios. In particular, we explore the use of supervised contrastive learning to cluster class embeddings nearly in the encoder latent space, which is used during inference for prototypical distance-based class assignment. We provide comprehensive experiments demonstrating the performance of our method in comparison to previous literature for detecting worn-out crossings.

Power-line communications are employed in home networking to provide easy and high-throughput connectivity. The IEEE 1901, the MAC protocol for power-line networks, employs a CSMA/CA protocol similar to that of 802.11, but is substantially more complex, which probably explains why little is known about its performance. One of the key differences between the two protocols is that whereas 802.11 only reacts upon collisions, 1901 also reacts upon several consecutive transmissions and thus can potentially achieve better performance by avoiding unnecessary collisions. In this paper, we propose a model for the 1901 MAC. Our analysis reveals that the default configuration of 1901 does not fully exploit its potential and that its performance degrades with the number of stations. Based on analytical reasoning, we derive a configuration for the parameters of 1901 that drastically improves throughput and achieves optimal performance without requiring the knowledge of the number of stations in the network. In contrast, 802.11 requires knowing the number of contending stations to provide a similar performance, which is unfeasible for realistic traffic patterns. We confirm our results and enhancement with testbed measurements, by implementing the 1901 MAC protocol on WiFi hardware.

In the past years, researchers have been advocating for flexible 802.11 devices that dynamically adapt to the varying network conditions, looking for efficient alternatives to the 802.11 standard MAC. In this work we demonstrate that this flexibility is readily available at the MAC level, and its operation can be tuned by re-programming the firmware inside the wireless chipsets that are built on relatively generic hardware modules. We show this by implementing the new amendment IEEE 802.11aa in legacy equipments by simply coding the frame exchange schemes at the firmware level. Nevertheless, we claim that the lack of flexibility in the way these modules interact results in a bottleneck that severely degrades performance. In our work, we prove this inefficiency of the 802.11 hardware architecture that hinders high throughput features, as in our case study of 802.11aa reliable multicast. To solve this problem, we provide new directions for the revision of the current hardware architecture and propose a new vision for the future design of wireless chipsets.

Handling voice traffic in existing WLANs is extremely inefficient, due to the large overhead of the protocol operation as well as the time spent in contention. In this paper, we propose a simple scheme (VoIPiggy) to improve the efficiency of WLANs with voice traffic. The key idea of the mechanism is to piggyback voice frames onto the MAC layer acknowledgments, which reduces both the frame overhead and the time wasted in contention. To quantify the gains of our proposal, we first study its performance by means of a capacity and delay analysis of a WLAN operating under the VoIPiggy mechanism. Then, we present an implementation of the mechanism using commercial off-the-shelf devices, which involves programming at the driver and firmware levels. The performance of the proposed scheme is evaluated in a large-scale testbed consisting of 30 devices. Our extensive measurements, which are comprised of different network conditions in terms of number of active nodes, traffic load and transmission rates, confirm that the experimental results match the analytical ones, and show a dramatic performance improvement for both “voice only” and “voice and data” scenarios.

Video traffic is foreseen to account for the majority of the Internet traffic in the near future. While the demand of video transmission keeps growing, the vast majority of wireless equipment deployed in the home environment, based on IEEE 802.11, cannot satisfy the amount of bandwidth that the video applications require. In order to cope with the increasing demand of multimedia traffic, the IEEE 802.11aa Task Group has recently standardized new mechanisms to allow efficient and robust transmission of multicast flows in Wireless LAN. However, the standard leaves open the choice of which one to use for a given scenario. In this paper, we explore the new mechanisms introduced by the 802.11aa Task Group, providing insights of the new choices for handling group addressed frames, by carrying out extensive simulations. Our results highlight the various trade-offs each mechanism has in terms of robustness, resource consumption and complexity, and provide a set of recommended guidelines for their use.

The estimation of thermal parameters belonging to active fires, such as fire radiative power (FRP), is a recommended method to determine fire severity. A review of the state of the art provides several semiempirical methods to estimate the FRP. On the other hand, the Sentinel-3/SLSTR instrument has two spectral channels dedicated to active-fire observation, which allows fire monitoring in a wide variety of conditions not affected by sensor saturation. This contribution analyzes these technical approaches, which are to be applied on the future Sentinel-3/Sea and Land Surface Temperature Radiometer (SLSTR) sensor in order to clarify its requirements and restrictions.

Supporting voice traffic in existing WLANs results extremely inefficient, given the large overheads of the protocol operation and the need to prioritize this traffic over, e.g., bulky transfers. In this paper we propose a simple scheme to improve the efficiency of WLANs when voice traffic is present. The mechanism is based on piggybacking voice frames over the acknowledgments, which reduces both frame overheads and time spent in contentions. We evaluate its performance in a large-scale testbed consisting on 33 commercial off-the-shelf devices. The experimental results show dramatic performance improvements in both voice-only and mixed voice-and-data scenarios.

Beacon and event-driven warning messages are two important messages that are used by many Intelligent Transportation Systems (ITS) applications for improving road safety and traffic efficiency. Beacon messages are regularly broadcast by each vehicle to notify other vehicles of its presence and its status. Event-driven warning messages are emitted when a vehicle or the road infrastructure discerns a critical situation. When congestion occurs, it is necessary to reduce the beacon load to keep a certain amount of the available bandwidth for the event-driven messages because they carry time- critical information of high importance. This paper considers three beacon congestion control algorithms that control the beacon load below a certain threshold by adjusting the transmit rate and/or transmit power for the beacon messages. We perform a large set of simulations to evaluate these algorithms. Our simulation results reveal promising results for these algorithms.