Rate Adaptation Algorithm Research Articles

Parallel and scalable AI in HPC systems for CFD applications and beyond

This manuscript presents the library AI4HPC with its architecture and components. The library enables large-scale trainings of AI models on High-Performance Computing systems. It addresses challenges in handling non-uniform datasets through data manipulation routines, model complexity through specialized ML architectures, scalability through extensive code optimizations that augment performance, HyperParameter Optimization (HPO), and performance monitoring. The scalability of the library is demonstrated by strong scaling experiments on up to 3,664 Graphical Processing Units (GPUs) resulting in a scaling efficiency of 96%, using the performance on 1 node as baseline. Furthermore, code optimizations and communication/computation bottlenecks are discussed for training a neural network on an actuated Turbulent Boundary Layer (TBL) simulation dataset (8.3 TB) on the HPC system JURECA at the Jülich Supercomputing Centre. The distributed training approach significantly influences the accuracy, which can be drastically compromised by varying mini-batch sizes. Therefore, AI4HPC implements learning rate scaling and adaptive summation algorithms, which are tested and evaluated in this work. For the TBL use case, results scaled up to 64 workers are shown. A further increase in the number of workers causes an additional overhead due to too small dataset samples per worker. Finally, the library is applied for the reconstruction of TBL flows with a convolutional autoencoder-based architecture and a diffusion model. In case of the autoencoder, a modal decomposition shows that the network provides accurate reconstructions of the underlying field and achieves a mean drag prediction error of ≈5%. With the diffusion model, a reconstruction error of ≈4% is achieved when super-resolution is applied to 5-fold coarsened velocity fields. The AI4HPC library is agnostic to the underlying network and can be adapted across various scientific and technical disciplines.

NERVE: Real-Time Neural Video Recovery and Enhancement on Mobile Devices

As mobile devices become increasingly popular for video streaming, it is crucial to optimize the streaming experience for these devices. Although deep learning-based video enhancement techniques are gaining attention, most of them cannot support real-time enhancement on mobile devices. Additionally, many of these techniques are focused solely on super-resolution and cannot handle partial or complete loss or corruption of video frames, which is common in the Internet and wireless networks. To overcome these challenges, we present NERVE, a novel approach in this paper. NERVE consists of (i) a novel video frame recovery scheme, (ii) a new super-resolution algorithm, and (iii) an enhancement-aware video bit rate adaptation algorithm. We implement NERVE on an iPhone 12, and it can support 30 frames per second (FPS). We evaluate NERVE in various networks such as 3G, 4G, 5G, and WiFi networks. Our evaluation shows that NERVE enables real-time video recovery and enhancement, and results in 24% - 83% increase in video Quality of Experience (QoE) in our video streaming system.

Retracted: Based on Fuzzy Measure Algorithm Message Adaptive Rate Algorithm of Internet of Things

Retracted: Based on Fuzzy Measure Algorithm Message Adaptive Rate Algorithm of Internet of Things

Rate Adaptation Algorithms in IEEE 802.11 Wireless Networks: A Comparative Study

Rate Adaptation Algorithms in IEEE 802.11 Wireless Networks: A Comparative Study

Environment-Aware Rate Adaptation Based on Occasional Request and Robust Adjustment in 802.11 Networks.

The IEEE 802.11 standard provides multi-rate support for different versions. As there is no specification on the dynamic strategy to adjust the rate, different rate adaptation algorithms are applied according to different manufacturers. Therefore, it is often hard to interpret the performance discrepancy of various devices. Moreover, the ever-changing channels always challenge the rate adaptation, especially in the scenario with scarce spectrum and low SNR. As a result, it is important to sense the radio environment cognitively and reduce the unnecessary oscillation of the transmission rate. In this paper, we propose an environment-aware robust (EAR) algorithm. This algorithm employs an occasional small packet, designs a rate scheme adaptive to the environment, and enhances the robustness. We verify the throughput of EAR using network simulator NS-3 in terms of station number, motion speed and node distance. We also compare the proposed algorithm with three benchmark methods: AARF, RBAR and CHARM. Simulation results demonstrate that EAR outperforms other algorithms in several wireless environments, greatly improving the system robustness and throughput.

Energy-efficient multisensor adaptive sampling and aggregation for patient monitoring in edge computing based IoHT networks

The need for remote healthcare monitoring systems that utilize limited resources’ biosensors is growing. These biosensors increase the amount of transmitted data across the Internet of Healthcare Things (IoHT) network. Therefore, it is necessary to decrease the transmitted data and make a decision at the edge gateway to save the energy of the biosensors and produce a quick response for the medical staff. This paper proposes an energy-efficient multisensor adaptive sampling and aggregation (EMASA) for patient monitoring in edge computing-based IoHT networks. In the edge-based IoHT network, EMASA operates on two levels: biosensors and the edge gateway. Each biosensor removes the redundant sensed data using the local emergency detection and sampling rate adaptation algorithms. In the edge gateway, it implements a machine learning-based Support Vector Machine (SVM) model to provide a suitable decision about the status of the monitored patient. We accomplished various examinations using real data from the patients’ biosensors. According to the simulation results, EMASA reduced the size of transmitted data from 93.5% to 99% and saved 78.35% of energy when compared to a previous study. It keeps the whole score with a good representation at the Edge gateway and provides accurate and fast decisions based on the patient’s condition.

An error based adaptive learning rate stochastic gradient descent algorithm in convolutional neural network

If the learning rate of convolutional neural network (CNN) is set improperly, the efficiency and accuracy of the algorithm will be greatly affected. To solve this problem, a learning rate adaptive algorithm is proposed to improve the traditional SGD: based on parameter prediction, the historical training error is used to update the learning rate. Under the condition of given initial learning rate, experiments on classical data sets prove the effectiveness of the above algorithms, and the adaptive learning rate stochastic gradient descent algorithm can keep the convergence of the network. Training accuracy is relatively stable; Shorter training time; And improve the learning accuracy.

A rate adaption algorithm for backscatter networks in mobile scenes

A rate adaption algorithm for backscatter networks in mobile scenes

A Rate Adaption Algorithm for Backscatter Networks in Mobile Scenes

A Rate Adaption Algorithm for Backscatter Networks in Mobile Scenes

QoE-Driven Adaptive Streaming for Point Clouds

With increasing popularity of virtual reality and augmented reality, application of point clouds is in critical demand as it enables users to freely navigate in an immersive scene with six degrees of freedom. However, point clouds usually comprise large amounts of data, and are thus difficult to stream in bandwidth-constrained networks. It is therefore important, yet challenging, to efficiently stream the resource-intensive point clouds, such that the user's quality of experience (QoE) is guaranteed on a high-level but with a low bandwidth consumption. To this end, we propose a QoE-driven adaptive streaming approach for the tile-based point cloud transmission, to maximize the user's QoE while reducing the transmission redundancy. By exploiting the perspective projection, we specifically model the QoE of a 3D tile as a function of the bitrate of its representation, user's view frustum and spatial position, occlusion between tiles, and the resolution of rendering device. Based on this QoE model, we then formulate the QoE-optimized rate adaptation problem as a multiple-choice knapsack problem, which allocates bitrates for different tiles under a given transmission capacity. It is equivalently converted to a submodular function maximization problem subject to knapsack constraints, and solved by a practical greedy-based algorithm with a theoretical worst-case performance guarantee. The proposed algorithm is able to achieve a near-optimal performance, but with a very low computational complexity. Experimental results further demonstrate superiority of the proposed rate adaptation algorithm over existing schemes, in terms of both user's visual quality and transmission efficiency.

Based on Fuzzy Measure Algorithm Message Adaptive Rate Algorithm of Internet of Things

In order to solve the problem that the existing LoRaWAN adaptive data rate control algorithm leads to low data transmission efficiency in the case of network congestion, a method combining a fuzzy logistic regression classifier and an improved adaptive data rate controller adjusting the avoidance time was proposed. The classifier could obtain the predicted congestion state by logistic regression learning. The data rate controller determined the data rate adjustment scheme according to the predicted congestion state. The experimental results showed that when the network congestion occurred in about 12s, the number of packet loss by the LoRaWAN default method was higher than that by the method in the research. The value of ADR_ MSG_CNT of the 15 source nodes in the method was 30 within 0–10 s, while the RCV_ACK_CNT of some nodes was 0. It proved that the method was more efficient than the original LoRaWAN adaptive data rate control algorithm.

A Collision Reduction Adaptive Data Rate Algorithm Based on the FSVM for a Low-Cost LoRa Gateway

LoRa (Long Range), a wireless communication technology for low power wide area networks (LPWANs), enables a wide range of IoT applications and inter-device communication, due to its openness and flexible network deployment. In the actual deployment and operation of LoRa networks, the static link transmission scheme does not make full use of the channel resources in the time-varying channel environment, resulting in a poor network performance. In this paper, we propose a more effective adaptive data rate (ADR) algorithm for low-cost gateways, we firstly analyze the impact of the different hardware parameters (RSSI, SNR) on the link quality and classify the link quality using the fuzzy support vector machine (FSVM). Secondly, we establish an end device (ED) throughput model and energy consumption model and design different adaptive rate algorithms, according to the different link quality considering both the link-level performance and the MAC layer performance. The proposed algorithm uses machine learning to classify the link quality, which can accurately classify the link quality using a small amount of data, compared to other adaptive rate algorithms, and the link parameter adaptation algorithm can maximize the throughput while ensuring the link stability, by considering the link-level performance and the MAC layer performance, compared to other algorithms. The results show that it outperforms the standard LoRaWAN ADR algorithm in both the single ED and the multi ED scenarios, in terms of the packets reception rate (PRR) and the network throughput. Compared to the LoRaWAN ADR in 32 multi-ED scenarios, the proposed algorithm improves the throughput by 34.12% and packets the reception rate by 26%, significantly improving the network throughput and the packets reception rate.

Optimal wireless rate and power control in the presence of jammers using reinforcement learning

Future wireless networks require high throughput and energy efficiency. This paper studies using Reinforcement Learning (RL) to do transmission rate and power control for maximizing a joint reward function consisting of both throughput and energy consumption. We design the system state to include factors that reflect packet queue length, interference from other nodes, quality of the wireless channel, battery status, etc. The reward function is normalized and does not involve unit conversion. It can be used to train three different types of agents: throughput-critical, energy-critical, and throughput and energy balanced. Using the NS-3 network simulation software, we implement and train these agents in an 802.11ac network with the presence of a jammer. We then test the agents with two jamming nodes interfering with the packets received at the receiver. We compare the performance of our RL optimal policies with the popular Minstrel rate adaptation algorithm: our approach can achieve (i) higher throughput when using the throughput-critical reward function; (ii) lower energy consumption when using the energy-critical reward function; and (iii) higher throughput and slightly higher energy when using the throughput and energy balanced reward function. Although our discussion is focused on 802.11ac networks, our method is readily applicable to other types of wireless networks.

A model for a practical evaluation of a DASH-based rate adaptive algorithm over HTTP

The proliferation of multimedia devices and user-generated content has driven a massive growth in Internet traffic for video streaming. There is a dire need to propose such solutions that can address multimedia streaming issues to achieve a user’s quality. Dynamic adaptive streaming over HTTP (DASH) improves the user’s quality through practical systems with limited bandwidth that enables the streaming media to run smoothly. This modern technology makes it easy to improve user perception and define the media presentation description (MPD) in terms of Uniform Resource Locator, content file, etc. The proposed adaptation algorithm attempts to determine the optimal solution to alleviate the conflict between maximizing the video quality and avoiding buffer stalls. We evaluate the proposed algorithm against alternative solutions such as ALDASH and FDASH for video content by taking into account the video bitrate, buffer level, and video bitrate switches for the single-user environment. A set of experiments have been conducted to investigate and analyze the benefits of our proposed algorithm. Our proposed algorithm is tested using the NS-3 network simulator, which shows that it improves video quality performance when compared to ALDASH and FDASH in terms of user satisfaction. Last but not least, the experimental results of our proposed algorithm provide high viewer quality in adaptive streaming as compared to its compititors.

Sensor-Assisted Rate Adaptation for UAV MU-MIMO Networks

Propelled by multi-user MIMO (MU-MIMO) technology, unmanned aerial vehicles (UAVs) as mobile hotspots have recently emerged as an attractive wireless communication paradigm. Rate adaptation (RA) becomes indispensable to enhance UAV communication robustness against UAV mobility-induced channel variances. However, existing MU-MIMO RA algorithms are mainly designed for ground communications with relatively stable channel coherence time, which incurs channel measurement staleness and sub-optimal rate selections when coping with highly dynamic air-to-ground links. In this paper, we propose SensRate, a new uplink MU-MIMO RA algorithm dedicated for low-altitude UAVs, which exploits inherent onboard sensors used for flight control with no extra cost. We propose a novel channel prediction algorithm that utilizes sensor-estimated flight states to assist channel direction prediction for each client and estimate inter-user interference for optimal rates. We provide an implementation of our design using a commercial UAV and show that it achieves an average throughput gain of 1.24\times and 1.28\times compared with the bestknown RA algorithm for 2- and 3-antenna APs, respectively

RAEN: Rate Adaptation for Effective Nodes in Backscatter Networks.

A backscatter network, as a key enabling technology for interconnecting plentiful IoT sensing devices, can be applicable to a variety of interesting applications, e.g., wireless sensing and motion tracking. In these scenarios, the vital information-carrying effective nodes always suffer from an extremely low individual reading rate, which results from both unpredictable channel conditions and intense competition from other nodes. In this paper, we propose a rate-adaptation algorithm for effective nodes (RAEN), to improve the throughput of effective nodes, by allowing them to transmit exclusively and work in an appropriate data rate. RAEN works in two stages: (1) RAEN exclusively extracts effective nodes with an identification module and selection module; (2) then, RAEN leverages a trigger mechanism, combined with a random forest-based classifier, to predict the overall optimal rate. As RAEN is fully compatible with the EPC C1G2 standard, we implement the experiment through a commercial reader and multiple RFID tags. Comprehensive experiments show that RAEN improves the throughput of effective nodes by 3×, when 1/6 of the nodes are effective, compared with normal reading. What is more, the throughput of RAEN is better than traditional rate-adaptation methods.

Tile-Based Panoramic Video Quality Assessment

As part of video transmission, video quality assessment plays an important role in improving the user’s visual quality and reducing transmission bitrate. However, the existing assessment schemes for the tile-based transmission mechanism cannot accurately predict the video quality. In this paper, we propose a tile-based panoramic video quality assessment method to optimize the transmission strategy for tile-based streaming. Considering the distortion of panoramic video projection, the inverse of gnomonic projection and bilinear interpolation are utilized to convert the planar tiles to sphere. Video Multimethod Assessment Fusion (VMAF) is applied to obtain the tile quality score. Besides, the user’s head movement and eye movement are combined to generate a tile weight map, and generate the final score of the panoramic video through weighted averaging. In addition, considering the impact of the quality fluctuation of the tiles within field of view, the tile quality loss model is established from multiple dimensions including the encoding parameter, viewport position and video content. Experimental results show that there is a high correlation between the predicted results and actual subjective scores, under two video quality assessment datasets. Furthermore, we integrate the method into the panoramic video transmission system. Compared with the tile-based adaptive rate algorithm, the proposed method degrades the video quality by only 3%, while reducing the bandwidth consumption by 16.99%.

AARF-HT: Adaptive Auto Rate Fallback for High-Throughput IEEE 802.11n WLANs

Wireless Local Area Network (WLAN) has been progressing rapidly. The IEEE 802.11n Physical (PHY) layer provides wider channel bandwidth, shorter guard interval, and up to four data streams. Therefore PHY 802.11n has a maximum of 128 data rate options from 6.5 Mbps to 600 Mbps. In addition, Medium Access Control (MAC) has been added Aggregate MAC Protocol Data Unit (AMDPU) scheme. If AMPDU is transmitted with a data rate corresponding to the channel conditions, then the probability AMPDU is received without error becomes increased. MAC determines the data rate used for transmitting AMPDU using a rate adaptation algorithm. Therefore some papers have proposed rate adaptation algorithms based on channel conditions. In this paper we propose a new rate adaptation algorithm that we call Adaptive Auto Rate Fallback for High Throughput (AARF-HT). Our development is done using NS-3 simulator version 3.26. AARF-HT algorithm performance is also tested through a number of simulations extensively. The simulation results show the data rate adaptation function based on the channel width, guard interval and the number of spatial streams in IEEE 802.11n WLAN has functioned well. The test results also show the AARF-HT algorithm resulted in higher throughput compared to the AARF algorithm.

Solution] Prepare your video for streaming with Segue

We identify new opportunities in video streaming, involving the joint consideration of offline video chunking and on-line rate adaptation. Due to a video’s complexity varyingover time, certain parts are more likely to cause performanceimpairments during playback with a particular rate adaptationalgorithm. To address such an issue, we propose Segue ,which carefully uses variable-length video segments, and augment specific segments with additional bitrate tracks. The keynovelty of our approach is in making such decisions basedon the video’s time-varying complexity and the expected rateadaptation behavior over time. We propose and implementseveral methods for such adaptation-aware chunking. Ourresults show that Segue substantially reduces rebufferingand quality fluctuations, while maintaining video quality delivered; Segue improves QoE by 9% on average, and by 22%in low-bandwidth conditions. Finally, we view our problemframing as a first step in a new thread on algorithmic anddesign innovation in video streaming, and leave the readerwith several interesting open questions.

Centralized Control System for Smart Street Lights Based on STM32 and LoRa

In order to solve the problem of waste of power and human resources in the traditional street lamp control system, this paper designs a smart street lamp centralized control system based on STM32 and LoRa technology. The centralized control system consists of a centralized control device and multiple single lamp control devices. The STM32F103 is the MCU of the centralized control device. The system uses the characteristics of low communication cost and long transmission distance of LoRa technology to realize the communication between the centralized control device and the single lamp control device. The centralized control device uses the Modbus RTU protocol to communicate with the sensor module, and adjusts according to environmental factors. The brightness of the street lamp realizes the "on-demand lighting" of the street lamp, and an adaptive rate algorithm is proposed to optimize the communication of the LoRa network and improve the working efficiency of the system. This centralized control system combines embedded and wireless communication technology and has advantages in cost, reliability, energy saving and intelligent control. It can be applied to street lamp lighting control in various environments. It has a high degree of intelligence and has a strong market potential.