LoRaWAN has emerged as a leading LPWAN technology for emergency response systems due to its energy efficiency and wide coverage. However, its limited bandwidth poses significant challenges for transmitting time-sensitive images, which are crucial for first responders. A major constraint in such scenarios is that the sender - typically a resource-constrained edge device - cannot perform complex Machine Learning (ML) operations for image compression or enhancement prior to transmission. Conversely, the receiver, often a more capable gateway, has sufficient computational resources to handle ML-based image reconstruction. To address this asymmetry, we propose ZML-ChunkLoRa, an adaptive image transmission framework that segments images into optimally sized chunks based on real-time network conditions. Reduce sender-side processing while enabling high-quality ML-based reconstruction at the receiver. By offloading computation to the gateway, ZML-ChunkLoRa improves transmission efficiency without violating the strict energy and bandwidth restrictions of LoRaWAN. Experimental results demonstrate that our approach significantly improves image transfer speed and visual quality in time-critical, resource-limited first-responder environments.

Reliable, long-range mesh networks are critical for emerging Air-to-Ground (A2G) applications that require high-throughput, infrastructure-free communication. We present WiLong-Air, a long-range aerial mesh platform built using commercial off-the-shelf hardware and open-source software, which integrates 5 GHz IEEE 802.11 radios with high-gain omnidirectional antennas. Using 802.11s mesh mode with B.A.T.M.A.N. Advanced routing, WiLong-Air enables scalable aerial connectivity. Field evaluations across outdoor aerial experiments validate its effectiveness: sustained throughputs above 20 Mbps at 700 m in partial NLoS, and reliable video and data transmission from multiple aerial nodes. Furthermore, with six concurrent transmitters, increased jitter, contention, and loss were observed, motivating architectural innovations. Moreover, apart from communications- related hardware and software improvements, the video streaming parameters were also optimised. For instance, it was seen that while variable bitrate mode of video transport improves perceptual quality, constant bitrate modes support steadier link behaviour. These results position WiLong-Air as a robust platform for A2G mesh research.

Mesh networks enables quick on-demand infrastructure creation and help achieve vital last-mile connectivity. Wi-Fi integration can further enhance the potential and compatibility for the pervasive connectivity requirements in next-generation applications. Despite advancements, Wi-Fi standards still need efforts to meet long-range requirements. Accordingly, this paper proposes a customized handheld platform named WiLong for constructing long-range Wi-Fi Mesh networks in the unlicensed higher bands, 2.4 and 5 GHz. The proposed platform can offer versatile radio profiles by leveraging commercially available hardware and open-source software components, augmented with the addition of IEEE 802.11s mesh link and the B.A.T.M.A.N. advanced routing protocol. Experimental evaluations of the said platform have been conducted across various practical environments, including indoor/outdoor spaces, urban & open bay areas, multi-floor basement car parking, and dense mesh deployments. The experimental results underscore the effectiveness of the WiLong platform in improving long-range Wi-Fi performance while emphasizing the significance of multi-hop networking in demanding ground-to-ground scenarios. These findings highlight the platform’s robustness and versatility. In a selected urban route scenario, the WiLong platform achieved 80% voice and 63% video call bandwidth. Additionally, when configured with 2.4 GHz, the platform reached two floors below in a multi-floor basement car parking.

In IEEE 802.11-based access networks (Wi-Fi), the client remains connected to a far-poor Access Point (AP) rather than switching to a near-better AP. This scenario is termed a sticky client problem. This scenario can severely impact the performance of real-time applications. Several standards, such as 802.11k1v/r, are being developed to enhance Wi-Fi roaming capabilities. However, the sticky client problem is yet to be solved completely. This paper proposes Sticky less, a novel method that leverages machine learning to learn the home Wi-Fi network behavior to address the sticky client problem. Initially, Stickyless divides the deployment area of APs into distinct zones, generates training data, and subsequently trains the machine learning module. The Stickyless employs clustering models to recommend selecting the optimal AP within a specific zone by considering the application performance and quality metrics. To conclude, Stickyless assesses performance using a proposed scoring and cascading module. We also developed a prototype to evaluate the Stickyless performance, outperforming the existing methods. The proposed method improves the Wi-Fi roaming experience by reducing stickiness up to 40 %. Thereby, it improves the link quality of the client by an average of 19 % and decreases the packet error rate by up to 3.5 % compared to the existing approaches. We also experimented with popular gaming apps, and Stickyless reduced the latency by up to 7 -fold.

In current Wi-Fi technology trend, Mobile Hotspot (MHS) or Soft Access Point (S-AP) is an integral part of our day-to-day life. At any time, MHS could be enabled as Wi-Fi Hotspot in mobility devices (smart phone, tablet) with cellular backhaul network (3G/4G/5G) and provides Internet access to client devices such as laptop, TV etc. Unlike Wi-Fi Access Point, which is typically a powered device, MHS is enabled as battery-operated device. In addition, MHS consumes higher power and reported as one of the primary Voice of Customer (VoC) issue. Due to high power consumption, many customers are skeptical about MHS feature and its continuous usage. Apart from few literatures, there is no specific IEEE 802.11 standard for MHS and its power management. In this paper, we have proposed a Machine Learning (ML) based Intelligent MHS Power Save (I-MHSPS) algorithm using Wi-Fi parameters such as RSSI, SNR, TX power and channel condition. In addition, we have used other contextual parameters such as client behavior, battery level, application usage and internet backhaul to improve the accuracy of our algorithm. In I-MHSPS, we have proposed Intelligent Transmit Power Control (I-TPC): MHS TX power regulation based on client vicinity, Intelligent Ultra Power Save (I-UPS): Applying different system power level for MHS operation and Intelligent Low Power Encryption (I-LPE): Enabling low power encryption for short range MHS. In our first experiment with I-TPC idea has reduced power consumption by 10-16% approximately when compared to existing methodologies. In second experiment for I-UPS, we have applied different system power levels for MHS operation and achieved power saving around 22% without any performance degradation. Further, in third experiment, using I-LPE method, we have observed the power required for encryption of data packets reduced by 20%.

WiFi Mesh network is made up of Mesh Access Points (MAPs). MAPs perform two roles: Mesh Station (STA) to communicate with other MAPs, and Basic Service Set (BSS) AP to interact with WiFi client devices. Hence, MAPs typically broadcast two kinds of independent beacons periodically, which will be received by other MAPs and WiFi client devices. This could lead to multiple beacon collision scenarios. And these beacon collisions occur in Mesh STA or WiFi client devices when other MAPs are hidden (two hops away). We have proposed Unified Signaling Beacon Framework (USBF) concept to reduce the probability of these beacon collision substantially. The proposed USBF performs unification and synchronization between BSS AP and Mesh STA beacons in MAPs. Our simulation result with the proposed concept shows improvement of beacon processing time (reduced around 35%), and total energy consumption (reduced around 22%).