اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديد5G Handover using Reinforcement Learning
76
0
0.0
(
0
)
نشر من قبل
Vijaya Yajnanarayana Ph.D
تاريخ النشر
2019
مجال البحث
الهندسة المعلوماتية
والبحث باللغة
English
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
In typical wireless cellular systems, the handover mechanism involves reassigning an ongoing session handled by one cell into another. In order to support increased capacity requirement and to enable newer use cases, the next generation wireless systems will have a very dense deployment with advanced beam-forming capability. In such systems, providing a better mobility along with enhanced throughput performance requires an improved handover strategy. In this paper, we will detail a novel method for handover optimization in a 5G cellular network using reinforcement learning (RL). In contrast to the conventional method, we propose to control the handovers between base-stations (BSs) using a centralized RL agent. This agent handles the radio measurement reports from the UEs and choose appropriate handover actions in accordance with the RL framework to maximize a long-term utility. We show that the handover mechanism can be posed as a contextual multi-armed bandit problem and solve it using Q-learning method. We analyze the performance of the methods using different propagation and deployment environment and compare the results with the state-of-the-art algorithms. Results indicate a link-beam performance gain of about 0.3 to 0.7 dB for practical propagation environments.
قيم البحث
اقرأ أيضاً
We consider distributed caching of content across several small base stations (SBSs) in a wireless network, where the content is encoded using a maximum distance separable code. Specifically, we apply soft time-to-live (STTL) cache management policie
s, where coded packets may be evicted from the caches at periodic times. We propose a reinforcement learning (RL) approach to find coded STTL policies minimizing the overall network load. We demonstrate that such caching policies achieve almost the same network load as policies obtained through optimization, where the latter assumes perfect knowledge of the distribution of times between file requests as well the distribution of the number of SBSs within communication range of a user placing a request. We also suggest a multi-agent RL (MARL) framework for the scenario of non-uniformly distributed requests in space. For such a scenario, we show that MARL caching policies achieve lower network load as compared to optimized caching policies assuming a uniform request placement. We also provide convincing evidence that synchronous updates offer a lower network load than asynchronous updates for spatially homogeneous renewal request processes due to the memory of the renewal processes.
We study positioning of high-speed trains in 5G new radio (NR) networks by utilizing specific NR synchronization signals. The studies are based on simulations with 3GPP-specified radio channel models including path loss, shadowing and fast fading eff
ects. The considered positioning approach exploits measurement of Time-Of-Arrival (TOA) and Angle-Of-Departure (AOD), which are estimated from beamformed NR synchronization signals. Based on the given measurements and the assumed train movement model, the train position is tracked by using an Extended Kalman Filter (EKF), which is able to handle the non-linear relationship between the TOA and AOD measurements, and the estimated train position parameters. It is shown that in the considered scenario the TOA measurements are able to achieve better accuracy compared to the AOD measurements. However, as shown by the results, the best tracking performance is achieved, when both of the measurements are considered. In this case, a very high, sub-meter, tracking accuracy can be achieved for most (>75%) of the tracking time, thus achieving the positioning accuracy requirements envisioned for the 5G NR. The pursued high-accuracy and high-availability positioning technology is considered to be in a key role in several envisioned HST use cases, such as mission-critical autonomous train systems.
Joint communication and sensing allows the utilization of common spectral resources for communication and localization, reducing the cost of deployment. By using fifth generation (5G) New Radio (NR) (i.e., the 3rd Generation Partnership Project Radio
Access Network for 5G) reference signals, conventionally used for communication, this paper shows sub-meter precision localization is possible at millimeter wave frequencies. We derive the geometric dilution of precision of a bistatic radar configuration, a theoretical metric that characterizes how the target location estimation error varies as a function of the bistatic geometry and measurement errors. We develop a 5G NR compliant software test bench to characterize the measurement errors when estimating the time difference of arrival and angle of arrival with 5G NR waveforms. The test bench is further utilized to demonstrate the accuracy of target localization and velocity estimation in several indoor and outdoor bistatic and multistatic configurations and to show that on average, the bistatic configuration can achieve a location accuracy of 10.0 cm over a bistatic range of 25 m, which can be further improved by deploying a multistatic radar configuration.
The provision of both wireless and wired services in the optical access domain will be an important function for future passive optical networks (PON). With the emergence of 5th generation (5G) mobile communications, a move toward a dense deployment
of small cell antenna sites, in conjunction with a cloud radio access network (C-RAN) architecture, is foreseen. This type of network architecture greatly increases the requirement for high capacity mobile fronthaul and backhaul links. An efficient way of achieving such connectivity is to make use of wavelength division multiplexed (WDM) PON infrastructure where wireless and wired services may be converged for distribution. In this work, for the first time, the convergence of 5G wireless candidate waveforms with a single-carrier wired signal is demonstrated in a PON. Three bands of universally filtered orthogonal frequency division multiplexing (UF-OFDM) and generalized frequency division multiplexing (GFDM), are transmitted at an intermediate frequency in conjunction with a digital 10Gb/s pulse amplitude modulation (PAM) signal in the downlink direction. Orthogonal frequency division multiplexing (OFDM) is also evaluated as a benchmark. Results show, for each waveform, how performance varies due to the 5G channel spacing - indicating UF-OFDMs superiority in terms of PON convergence. Successful transmission over 25km of fibre is also demonstrated for all waveforms.
This letter studies a basic wireless caching network where a source server is connected to a cache-enabled base station (BS) that serves multiple requesting users. A critical problem is how to improve cache hit rate under dynamic content popularity.
To solve this problem, the primary contribution of this work is to develop a novel dynamic content update strategy with the aid of deep reinforcement learning. Considering that the BS is unaware of content popularities, the proposed strategy dynamically updates the BS cache according to the time-varying requests and the BS cached contents. Towards this end, we model the problem of cache update as a Markov decision process and put forth an efficient algorithm that builds upon the long short-term memory network and external memory to enhance the decision making ability of the BS. Simulation results show that the proposed algorithm can achieve not only a higher average reward than deep Q-network, but also a higher cache hit rate than the existing replacement policies such as the least recently used, first-in first-out, and deep Q-network based algorithms.
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
