In this paper, we propose a target Direction of Arrival (DOA) estimation algo-rithm through the triangular array antenna to improve the perception accuracy of intelligent vehicular millimeter-wave radar. Firstly, we utilize Total Least Square- Estimation of Signal Parameters to obtain the solution set of target azimuth with the geometrical advantages of the triangular array. Subsequently, the preliminary optimal target azimuth can be derived based on the power spectrum function of the improved Multiple Signal Classification algorithm, which is used to estimate the DOA solution set. Finally, we could estimate the optimal azimuth parameters accurately through searching the preliminary optimal target azimuth. The pro-posed algorithm avoids a wide range of power spectrum search, and reduces the error of azimuth estimation. Compared with both TLS-ESPRIT and MMUSIC algorithm, the spatial resolution is further improved. The evaluation results indi-cate that the proposed algorithm reduces the root mean square error and the com-putational complexity, and meanwhile improves the target azimuth estimation ac-curacy.
Authors: Xiaoyu Du (Henan University), Guoping Jiang (Nanjing University of Posts and Telecommunications), Chong Han (Nanjing University of Posts and Telecommunications), Chunsong Wang (Henan Universtiy), Yi Zhou (Henan University),
Hide Authors & Abstract

RFID technology is playing an increasingly more important role in the Internet of Things, especially in the dense deployment model. In such networks, in addition to communication, nodes may also need to harvest energy from the environment to operate. In particular, we assume that our network model relies on RFID sensor network consisting of WISP devices and RFID exciters. In WISP, the sensors harvest ambient energy from the RFID exciters and use this energy for communication back to the exciter. However, as the number of exciters is typically small, sensors further away from an exciter will need longer charging time to be able to transmit the same amount of information than a closer by sensor. Thus, further away sensors limit the overall throughput of the network. In this paper, we propose to use a multi-modulation scheme, which trades off power for transmission duration. More specifically, in this scheme, sensors closer to the exciter use a higher-order modulation, which requires more power than a lower-order modulation assigned to further away sensors, for the same bit error rate of all the sensors’ transmissions. This reduces the transmission time of the closer sensors, while also reducing the charging time of the further away sensors, overall increasing the total network throughput. The evaluation results show that the RFID sensor network with our multi-modulation scheme has significantly higher throughput as compared with the traditional single-modulation scheme.
Authors: Zijing Tian (University of Texas at Dallas), Zygmunt Haas (University of Texas at Dallas, Cornell University),
Hide Authors & Abstract

Massive MIMO system greatly enriches the randomness of the secret keys in physical layer, and increases the rate of key generation. However, it is not practi-cal to obtain full channel state information for key generation in actual communi-cation scenarios due to a large number of additional signaling overhead. In this paper, we proposed a feasible physical layer key generation scheme by using the beam information as the random source. The procedure for key generation is de-signed based on the current beam management mechanism in 5G NR. Therefore, the secret key is synchronously generated in the process of two-stage beam man-agement between the gNB and the UE before data transmission, and the addition-al signaling overhead for key generation is little. Furthermore, to cope with the non-uniform distributed characteristics of the beams, we adopt Huffman code in the encoding of beam index, thereby improving the efficiency of the key genera-tion. Simulation results show that the proposed scheme can achieve a mutual in-formation per bit as high as 0.97, which is 2% to 3% better than that of equal length coding. Furthermore, bit disagreement rate can be less than 1% in a harsh communication environment with a signal-to-noise ratio of -10dB.
Authors: Hao Gao (Beijing University of Posts and Telecommunications), Yanling Huang (Beijing University of Posts and Telecommunications), Danpu Liu (Beijing University of Posts and Telecommunications),
Hide Authors & Abstract

The unmanned aerial vehicle (UAV) based aerial base station (BS) has emerged as a feasible solution to the high traffic demands of the future wireless networks. It is essential that UAV can integrat with non-orthogonal multiple access (NOMA) to support the massive connection service. In this paper, we study the problem of weighted sum rate maximization in the downlink for NOMAbased UAV networks. This is a non-convex optimization problem, which is intractable to be directly solved using the convex optimization method. To deal with the problem, we propose an efficient iterative algorithm via variable substitution and relaxation methods, then construct the framework of alternating optimization to solve the problem joint placement and power allocation optimization. The simulation results show that the proposed algorithm performs better than other schemes.
Authors: zhengqiang wang (Chongqing University of Posts and Telecommunications), Hao Zhang (Chongqing University of Posts and Telecommunications), Xiaoyu Wan (Chongqing University of Posts and Telecommunications, P. R. China), Zifu Fan (Chongqing University of Posts and Telecommunications, P. R. China), Xiaona Yang (Huaxin Consulting Company Ltd., Hangzhou 310014, Zhejiang, P. R. China), Yuanmao Ji (Senior Instructor, Ericsson (China) Communication Co., Ltd),
Hide Authors & Abstract

This paper considers an unmanned aerial vehicle (UAV)-enabled wireless power transfer (WPT) system, in which a UAV hovers in a given flying altitude to transfer energy to more than two energy receivers (ERs) on the ground.We consider to maximize the sum energy and weighted sum energy of ERs by UAV placement optimization. As the sum energy and weighted sum energy maximization problem are sum of ratio problems, which are generally NP-hard. It is difficult to give the optimal location of the UAV for those two problems. To tackle those problems, we adopt a novel quadratic transform technique to transfer to an equivalent problem. Based on the equivalent problem, we propose an iterative coordination update algorithm in a closed-form expression, which can converge to the stationary point of the sum energy maximization problem or even the global optimal solution under a sufficient condition of the flight altitude. Simulation results show that the proposed algorithm can achieve nearly the same weighted sum energy for ERs and reduces more than 90% complexity compared to the two-dimensional (2D) exhaustive search method.
Authors: zhengqiang wang (Chongqing University of Posts and Telecommunications), Yang Liu (Chongqing University of Posts and Telecommunications), Hao Zhang (Chongqing University of Posts and Telecommunications), Xiaoyu Wan (Chongqing University of Posts and Telecommunications, P. R. China), Zifu Fan (Chongqing University of Posts and Telecommunications, P. R. China),
Hide Authors & Abstract

Recently, the mobile crowd sensing (MCS) technique is believed to be an important role in multi-source data acquisition tasks. With devices or people with different sensing abilities in the cities, we can easily split and distribute the complex task in an appropriate way so that those devices or people can be stimulated to collect data within different scopes individually, while the results of them can be analyzed and integrated collaboratively to fulfill that complex task. However, in typical centralized architecture, the latency brought by unstable and time-consuming long-distance network transmission limits the development of MCS. The multi-access edge computing (MEC) technique is now regarded as the key tool to solve this problem. By establishing a service provisioning system based at the edge of the network, the latency can be reduced and the analysis or integration can also be conducted in time with the help of corresponding services deployed on nearby edge servers. However, as the edge servers are resource-limited, the sensing abilities vary among devices or people, and the budget of fulfilling a task is determined, we should be more careful in task assignment and service deployment. In this paper, we investigate the relationship between the task quality and the cost in the MEC-based MCS system, propose an analysis framework of it, and conduct comprehensive experiments to evaluate the performance of our approach.
Authors: Zhengzhe Xiang (Zhejiang University City College), Shuiguang Deng (Zhejiang University), Yuhang Zheng (Zhejiang University), Dongjing Wang (Hangzhou Dianzi University), Cheng Zhang (Zhejiang University), Yuanyi Chen (Zhejiang University City College), Zengwei Zheng (Zhejiang University City College),
Hide Authors & Abstract

In this paper, we proposed a novel MIMO channel measurement system architecture for 5G wireless communication based on the delay lines in combination with switches, and we implement propagation graph modeling to simulate the channel measurement procedure. Channel sounders equipped with multiple-element antenna arrays in the transmitter (Tx) and receiver (Rx) usually perform a measurement in two ways: switched channel measurement and parallel channel measurement. The latter usually needs multiple Txs/Rxs which leads a high cost, while the former with a high-speed radio-frequency switch at the transmitter and receiver has a lower cost but it is difficult to realize beamforming due to the Tx/Rx antennas do not transmit/receive signals simultaneously. By adding delay lines to the switched MIMO channel measurement system, the delay in different time slot at every Tx antennas can be compensated so that the multiple Tx antennas (empowered by only one Tx) transmit signals simultaneously. Furthermore, adding phase shifters after delay lines makes it easy to change the phase of each signal, which provides a convenient way for beamforming. The feasibility of the proposed method is preliminarily validated through simulations based on propagation graph modeling, the evaluation of the results is conducted by calculating the channel impulse response (CIR) or power delay profile (PDP) and estimating the direction of arrival (DOA) using Multiple Signal Classification (MUSIC) algorithm.
Authors: Shengnan Xu (Tongji University),
Hide Authors & Abstract

Indoor automatic localization technology is very important for the Internet of Things. With the development of wireless technology and the iversification of location service requirements, especially in complex indoor scenarios, users are increasingly demanding location-based ervices. Traditional Global Positioning System location technology is difficult to solve some positioning problems in indoor environments, and WiFi is now available in most indoor environments.Therefore, using WiFi for positioning is a very cost-effective method. However, WiFi-based indoor positioning requires a large amount of data, so we can use artificial intelligence methods to analyze the data and obtain a positioning model. The traditional indoor positioning methods based on WiFi signals have some problems such as long positioning time and poor accuracy.In order to solve the above problems, this paper proposes an indoor localization method based on Deep Neural Networks for WiFi fingerprint. In particular, a DNN-based WiFi-RSSI positioning method is proposed for indoor automatic localization.Besides, in the process of DNN training, a joint training method based on unsupervised learning and supervised learning is adopted and the special loss function is defined.Extensive experiments are carried out in both the UJIIndoorLoc public database and a real scenario, and a thorough comparison with several existing approaches indicates that the proposed scheme improves the localization accuracy on average.
Authors: Bing Jia (Inner Mongolia University), Zhaopeng Zong (Inner Mongolia University), Baoqi Huang (Inner Mongolia University), Thar Baker (SLiverpool John Moores University),
Hide Authors & Abstract

Current cellular mobile network should satisfy the service requirements of the User Equipment (UE) applications through Radio Resource Management (RRM) mechanisms as much as possible. In order to improve the resource utilization rate and Quality of Experience (QoE) for downlink Real-Time (RT) services in 5G system. In this paper, based on the Modified Largest Weighted Delay First (M-LWDF) algorithm, a slicing-oriented resource scheduling algorithm-S-MLWDF is proposed with using 5G network slicing technology. S-MLWDF takes RB groups as the basic units of RA (resource allocation) and takes slices as the allocation object. During the process of in-slice scheduling, on account of the Channel Quality Indication (CQI) ob-tained from Base Station (BS) feedback and the allocation of RBs over time, the generated weighting factor can guarantee the edge users to get equal scheduling opportunities. Meanwhile, the modified queue delay and HARQ retransmission packets delay can solve the problem of surge in Packet Loss Rate (PLR) near the delay threshold. The simulated results show that the per-formance of the proposed algorithm is better than the traditional scheduling algorithms. Especially compared with M-LWDF, the fairness and PLR of S-MLWDF are optimized by about 10% and 16.3%, which can better meet the needs of users.
Authors: Shanwei Wang (School of Communication and Information Engineering Chongqing University of Posts and Telecommunications), Bing Xi (School of Communication and Information Engineering Chongqing University of Posts and Telecommunications), Zhizhong Zhang (School of Communication and Information Engineering Chongqing University of Posts and Telecommunications), Bingguang Deng (School of Communication and Information Engineering Chongqing University of Posts and Telecommunications),
Hide Authors & Abstract

The explosive increasing of high bandwidth-consumption traffic puts great pres-sure on the Internet. Many popular applications springing up work in a manner of one-to-many or many-to-many communication, such as TikTok, Instagram, Ten-cent conference, and numbers of interactive games. Due to the scalability and ap-plicability problems, existing multicast schemes, e.g. IP multicast, are not widely implemented. Instead, most of those traffic is transmitted through the Internet in unicast, which results in vast redundant traffic in backbone networks. In this pa-per, we propose a unicast/multicast jointed routing mechanism in software de-fined networks, SDUM. We devote to achieve unicast data distribution following a dynamic multicast tree, which is managed by centralized control and application plane. Other than OpenFlow protocol, this mechanism doesn’t require any specif-ic multicast protocols or software. The network can be a virtualized network with distributed OpenFlow devices interconnected by legacy routers. The evaluation results confirm the efficiency of the proposal in the number of control messages, signaling overhead, occupation of flow table entries, and qualitative comparison.
Authors: Shimin Sun (Tiangong University), Wentian Huang (Tiangong University), Xinchao Zhang (Tiangong University), Li Han (Tianjin University of Technology),
Hide Authors & Abstract

The deployment of small cells of dynamic time division duplexing (TDD) using orthogonal frequency division multiple access (OFDMA) is one of the key technologies of 5th-generation (5G). The main problem limiting the performance of dynamic TDD is the existence of cross-link interference in these mobile networks. In order to adapt to the new scenario of 5G, on the basis of the traditional clustering scheme, this paper proposes a new clustering criterion considering both the user equipment to user equipment (UE-UE) interference and base station to base station (BS-BS) interference, and realizes dynamic clustering of the cells. The proposed method can simultaneously solve the uplink and downlink interference problems, and improve the system performance. Compared with the traditional clustering algorithm, the scheme proposed in this paper has certain improvement in the throughput of the system, especially in the downlink.
Authors: Junping Liu (Southeast University), Zekai Liu (Southeast University), Nan Liu (Southeast University), Zhiwen Pan (Southeast University), Xiaohu You (Southeast University),
Hide Authors & Abstract

In future non-terrestrial networks, satellites with sufficient computing resources are expected to serve as base stations or access points, which changes the structure of traditional satellite networks into a more flexible environment. For such satellite nodes, denial of service (DoS) attacks may become a potential secure threat that should be prevented. Existing researches mainly focus on the defense against DoS attacks on ground nodes and have no consideration of the attacks on future satellites. Moreover, the problem of reducing the access delay when a satellite is under DoS attack has not been addressed. In this paper, we study the DoS attack defense strategy in non-terrestrial networks. By adopting client puzzle protocol(CPP) and load balancing, we propose a cooperative defense strategy where multiple auxiliary nodes are used to help the attacked node to process the intensive attack requests. An access delay minimization problem to optimize the selection of auxiliary nodes, puzzle difficulty as well as traffic offload ratios is then formulated based on queuing theory and solved. Simulation results show that the proposed solution not only improve the network's anti-attack capability, but also achieves desirable performance in average access delay.
Authors: Zhaori Cong (Beijing University of Posts and Telecommunications), Zhilong Zhang (Beijing Laboratory of Advanced Information Network, Beijing Key Laboratory of Network System Architecture and Convergence, Beijing University of Posts and Telecommunications.), Danpu Liu (Beijing University of Posts and Telecommunications),
Hide Authors & Abstract

Millimeter wave (mmWave) system with massive multiple input multiple output (mMIMO) meets increasing data traffic requirements. However, fast beam tracking for vehicles with high mobility causes enormous overhead, especially in an ultra-dense network (UDN) with frequently base station (BS) handover. In this paper, we proposed a multi-site coordination beam tracking scheme utilizing the spatial correlation of channel state information (CSI) among different sites to reduce the signaling overhead for beam training and handover. The scenario is a hyper-cellular network (HCN) with one control-BS (CBS) and multiple traffic-BSs (TBSs). The proposed scheme consists of two stages. In the first stage, more accurate position measurement of the moving user equipment (UE) can be achieved by using uniform planar array (UPA), and Extended Kalman Filter (EKF) is exploited in CBS to predict the UE’s location in the next slot. In the second stage, the relationship between multi-site and UE’s location is used by the CBS to remotely infer the candidate beam between each TBS and the UE, and make a TBS handover decision when necessary. Given that it is the CBS in charge of beam tracking between all the TBSs and the UE centrally, the overhead for beam training and handover are both efficiently reduced. Simulation results based on realistic 3D scenario show that the proposed scheme can achieve 99% of the optimal spectral efficiency with fewer overhead for beam sweeping and handover signaling.
Authors: Xingwen He (Beijing University of Posts and Telecommunications), Danpu Liu (Beijing University of Posts and Telecommunications), Zhilong Zhang (Beijing Laboratory of Advanced Information Network, Beijing Key Laboratory of Network System Architecture and Convergence, Beijing University of Posts and Telecommunications.),
Hide Authors & Abstract

This paper considers consider clustered small cell networks (SCNs)with combined design of cooperative caching and energy-efficient policy in theCoordinated Multi-Point (CoMP)-enabled cellular network. Small base stations(SBSs) with cache storage are grouped into associative clusters which can com-municate with each other. This paper focus on movie on-demand streaming fromInternet-based servers and proposed combined caching mode, where every SBSutilizes parts of cache space to cache the most popular contents (MPC), whilethe remaining is used for cooperatively caching different partitions of the lesspopular contents (LPC). Instead of the known content popularity, we constructs acontent-aware weighted feature matrix (CWFM) in terms of spatiotemporal varia-tion. Based on estimated content popularity and transmission design, we proposea caching scheme that makes a caching decision to maximize the energy effi-ciency (EE). To tackle this problem, A two-step stepwise optimization methodis adopted. First, EE conditioning is optimized with a approach of linear pro-gramming and variable recovery. Then, the optimal proportion of cache spacefor MPC is analyzed by comparing the energy-efficient gain from the MPC withthe energy-efficient loss from the discarded contents. Extensive simulation re-sults confirm that our algorithm outperforms state-of-the-art algorithms based onMovieLens data set.
Authors: Xiang Yu (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications, CN 400065), Huiting Luo (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications, CN 400065), Long Teng (Shanghai Jiao Tong University - Minhang Campus Shanghai JiaoTong University-Minhang Campus Shanghai, CN 200240), Ting Liu (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications, CN 400065),
Hide Authors & Abstract

The urbanization process has changed urban land, which has affected the environmental quality of urban residents. It is very important to obtain urban land cover information. In this paper, Yangshuo, a small country of Guilin City , is used as the research area, and the object-oriented spatial feature extraction module (Feature Extraction, hereinafter referred to as FX) is used to carry out experiments and accuracy evaluation of land cover classification in the research area. Extracting land cover information from the GF-2 remote sensing image, establishing a classification system sample based on the characteristic information of six land cover classification objects such as urban land, waterbody, woodland, farmland, road and other lands, and finally execute Supervising the classification and verify its accuracy. The results show that this method can recognize the land cover accurately and the total accuracy verified is as high as 97.41%.
Authors: xiaomao chen (Guangxi Key Laboratory of Wireless Wideband Communication and Signal Processing, Guilin University of Electronic Technology), Jia Kun Li (Guangxi Key Laboratory of Wireless Wideband Communication and Signal Processing, Guilin University of Electronic Technology), Yuanfa Ji (National & Local Joint Engineering Research Center of Satellite Navigation Positioning and Location Service，Guilin, 541004, China),
Hide Authors & Abstract

Faced with the growth of massive data in the 5G network and the low query efficiency in the existing signaling synthesis algorithm, the traditional LTE/LTE-A signaling monitor scheme has been unable to satisfy with the new 5G network architecture. The Radio Resource Control (RRC) protocol is the core of the control plane, which manages and controls the wireless resources of the network. Based on this, a signaling monitor scheme of RRC protocol suitable for 5G road tester was proposed and the specific functions of its main submodule were introduced in detail. Additionally, this paper discussed an improved dynamic hash signaling synthesis algorithm based on AVL tree under a new network architecture. The tree structure is used to reduce the query time of traditional algorithm in the hash table, so as to quickly deal with hash collisions and improve the real time on the signaling synthesis. At present, the proposed scheme was applied to the real network testing of air interface data in the 5G road tester. The experimental results show that the improved algorithm can efficiently solve the issue of low efficiency of Call Detail Recording (CDR) synthesis and the average query time can be reduced by 49.3% and 33.1% compared with the two traditional algorithms. The signaling synthesis scheme described above achieves the expected effect and signaling messages in RRC protocol can be accurately decoded and monitored in real time.
Authors: Bingying Zhang (Chongqing University of Posts and Telecommunications), Fang Cheng (Chongqing University of Posts and Telecommunications), Bingguang Deng (Chongqing University of Posts and Telecommunications),
Hide Authors & Abstract

The new Segment Routing paradigm provides network operator the possibility of highly increasing network performance exploiting advanced Traffic Engineering features and novel network programmability functions. Anyway, as any new solutions, SRv6 has a side effect: the introduction of unknown service disruption events. In this work we focus on packet lost events due to the incorrect computation of the Maximum Transmission Unit (MTU) value of an end-to-end path in an SRv6 network. This event, referred to as MTU dependent SR Black Hole, cannot be detected by known monitoring solutions based on active probing: the reason is that in SRv6 probe packets and user data can experience different network behaviors. In this work we propose a passive monitoring solution able to exploit the SRv6 Traffic Counters to detect links where packets are lost due to MTU issues. The performance evaluation shows that the algorithm proposed is able to identify the link affected by the blackhole with a precision equal to 100\%; moreover, the flow causing the blackhole cannot be detected with the same precision, but it is possible to identify a restricted set of flows, referred to as suspected flows, containing the target one.
Authors: Marco Polverini (Sapienza), Antonio Cianfrani (Sapienza), Marco Listanti (Sapienza),
Hide Authors & Abstract

Heterogeneous network convergence, as the current development trend of wireless communication network systems, has attracted the attention and research of many experts. In order to solve the problem of incomplete handover decision parameters and single decision algorithm in 5G heterogeneous network handover system, an M-F-U hybrid algorithm based on the multiple attribute decision making (MADM), fuzzy logic, and utility function is proposed. First, the decision parameters are divided into two parts, which are calculated by the MADM and fuzzy logic methods, the results obtained as the input of the utility function, secondly, the risk attitude coefficient is introduced into the utility function to describe the user's tolerance for switching risk, then, Then calculate the value of the comprehensive utility function, and finally, choose the optimal network scheme according to the comprehensive utility value. The simulation results show that compared with the traditional algorithm, the M-F-U algorithm can improve the handover accuracy, reduce the number of handovers, and complete the switching decision in a short time.
Authors: Haodong Liu (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications), Fang Cheng (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications), Bingguang Deng (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications),
Hide Authors & Abstract

This paper investigates the security issues and performance optimization of the blockchain. Security has been a hot topic in blockchain technology. Stealing cryptocurrency and disclosing the privacy of transaction process have exposed the vulnerability of blockchain in different degrees. These vulnerabilities not only caused significant losses to the project team and users, but also raised doubts about the security of the blockchain. As a formalized contract in the code, smart contracts provide a better security method than traditional ones, while they increase the risk of blockchain. Moreover, secure transactions should resist external attacks and protect user privacy. In addition, the performance analysis of blockchain has also aroused great interest. Therefore, this paper summarizes the related work of blockchain performance analysis from the following three aspects to promote the further research of blockchain: simulation systems of blockchain, evaluation of blockchain network and optimization of blockchain application.
Authors: Dongqing Li (Hangzhou Dianzi University), Congfeng Jiang (Hangzhou dianzi University), Yin Liu (Information&Telecommunications Company, State Grid Shandong Electric Power Company), Linlin Tang (Information&Telecommunications Company, State Grid Shandong Electric Power Company), Li Yan (Information&Telecommunications Company, State Grid Shandong Electric Power Company),
Hide Authors & Abstract

Blockchain technology has shown great potential in the Internet of Vehicles(IoV) in solving the problems of data sharing and information traceability. Understanding the relationship between the IoV communication architecture and the blockchain can facilitate designing dedicated blockchain enabled IoV systems. In this paper, a two-layer wireless blockchain network architecture based on Dedicated Short Range Communication (DSRC) is proposed. Then the M/G/1 queuing model is used to analyze the delivery process of the transaction under the unsaturated condition, and the Markov model is established to analyze the unicast service process. Finally, the verification process of the block in the Tangle consensus network is deduced. The simulation results show that the network load, channel conditions, queuing and backoff service processes in the wireless environment have a significant impact on the delay and throughput of the blockchain network, which further proves that the wireless environment is the main reason for limiting the performance of the Tangle blockchain network.
Authors: QiLie Liu (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications), Liang Lin (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications), Yun Li (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications), Yongxiang Liu (North China Electric Power University),
Hide Authors & Abstract

We consider the problem of multi-message private information retrieval (MPIR) from $N$ non-colluding and replicated servers when the user is equipped with a cache that holds an uncoded fraction $r$ from each of the $K$ stored messages in the servers. We assume that the servers are unaware of the cache content. We investigate $D_{P}^*(r)$, which is the optimal download cost normalized by the message size, as a function of $K$, $N$, $r$, $P$. For a fixed $K$, $N$, we develop an inner bound (converse bound) for the $D_{P}^*(r)$ curve. The inner bound is a piece-wise linear function in $r$. For the achievability, we propose specific schemes that exploit the cached as private side information to achieve some corner points. We obtain an outer bound (achievability) for any caching ratio by memory-sharing between these corner points. Thus, the outer bound is also a piece-wise linear function in $r$. The inner and the outer bounds match for the cases where the number of desired messages $P$ is at least half of the number of the overall stored messages $K$. Furthermore, the bounds match in two specific regimes for the case $\frac{K}{P} > 2$ and $\frac{K}{P} \in \mathbb{N}$: the very high ratio regime and the very low ratio regime. Finally, the bounds meet in one specific regime for arbitrarily fixed $K$, $P$, $N$: the very high ratio regime,i.e., $r \geq \frac{1}{N+1}$.
Authors: Yang Li (Southeast University), Nan Liu (Southeast University), Wei Kang (Southeast University),
Hide Authors & Abstract

This paper proposed a dynamic data transmission scheduling algorithm based on the way a group of firefly communicating with one another. The proposed algorithm will be compared again a random transmission approach. To evaluate the performance of the algorithms, different numbers of nodes in the networks will be evaluated. The results from the hardware experiments have shown that the firefly algorithm can schedule the transmission of data packet with high delivery rate for a small network. However, as the number of nodes increases, the packet delivery rate decreases. The proposed algorithm can also increase the lifespan of the battery as the nodes will be operating in the sleep mode and will only be awake during the sychronization period for data transmission.
Authors: Norhafizah Muhammad (Universiti Teknologi Brunei), Tiong Hoo Lim (Universiti Teknologi Brunei),
Hide Authors & Abstract

This paper targets on improving the efficiency of the concurrent transmissions for the hyper-dense device-to-device (D2D) networks using the millimeter wave (mmWave) transmission technology. To this end, the increment of the multiple access interference is well controlled by the proposed energy-efficient (EE) power adjustment scheme. Therein, the nonlinear fractional programming technique is firstly applied to transform the nonlinear optimization problem into the linear from. Then, the transmission power of the D2D pairs is formulated as the noncooperative Game. Via the well-known Karush-Kuhn-Tucker condition, the optimal transmission power can then be decided. With the aid of the EE power adjustment, we modify the conventional vertex multi-coloring concurrent transmission scheme to accommodate more D2D pairs. The main concept is to judge the feasibility of the concurrent transmission based the increase of sum data rate rather than the individual rate. The superiority of the proposed scheme is verified by simulations in terms of the EE and data rate.
Authors: Wenson Chang (National Cheng Kung University), Bo-Jun Yang (ASUSTeK Computer Inc., Taiwan R.O.C),
Hide Authors & Abstract

This paper targets on solving the high collision problem in the massive machine-to-machine communications. The main idea is to restrict the numbers of allowable contentions to the low-energy devices (LEDs) by using the proposed make-or-break (MOB) backoff scheme such that the unnecessary energy consumption can be reduced. And, consequently, the high-energy devices (HEDs) can have higher probability to attain time slots for data transmissions. However, the restriction mechanism may detain the data forwarding process. To solve this dilemma, the adaptive frame structure is developed to compensate the loss of throughput. The analytical as well as simulation results demonstrate that with a huge amount of machine type devices (MTDs), the proposed scheme can outperform the conventional counterpart in the aspects of the head-of-line delay, energy efficiency and accommodation of the MTDs.
Authors: Wenson Chang (National Cheng Kung University), Chun-Wei Huang (Pegatron Corporation, Taiwan R.O.C),
Hide Authors & Abstract

In order to track 5G downlink shared channel in real time and reduce computational complexity, a linear minimum mean square error algorithm based on demodulation reference signal adaptive parameter estimation is proposed. Firstly, the SNR nonlinear centralized optimization problem is transformed into a multivariable linear programming problem due to the restriction of non-uniform energy distribution in time-domain channel. Secondly, considering the uncertainty of multipath delay channel, the combination of negative exponential distribution model and generalized correlation algorithm is taken advantage of so that the original problem is turned into a specific parameter optimization problem. At the same time, according to the obtained delay parameters and SNR, the most appropriate interpolation coefficient is selected for the LMMSE channel estimation by combining with the sliding window, which avoids the matrix inversion process, realizes the real-time matching of parameters, and reduces the computational complexity. The simulation results show that the proposed algorithm has better system performance compared with the classical channel estimation algorithm.
Authors: Bingguang Deng (Chongqing University of Posts and Telecommunications), Xiaofang Min (Chongqing University of Posts and Telecommunications), Siyi Yu (Chongqing University of Posts and Telecommunications), Qianqian Ye (Chongqing University of Posts and Telecommunications),
Hide Authors & Abstract

Due to dynamic changes in network topology and constant changes in links between satellites, the routing paths calculated by the traditional shortest multipath routing algorithm are not updated in time, resulting in that the problematic paths are still transmitting data, so a large number of problems such as packet loss and service failure exist in the network. This paper designs a multipath routing algorithm for the satellite network topology which changes frequently, constrained multipath routing algorithm (CMRA).By calculating all feasible path of the entire network topology, CMRA will choose the multiple paths which satisfy the constraint conditions. After that, it will choose the lowest cost value of these paths. Repeatedly performing these operations, select multiple high_quality paths which satisfy the constraint conditions for traffic. Simulation results show that compared with the traditional shortest multipath routing algorithm and single path routing algorithm, the proposed routing algorithm is better in packet loss rate and average time delay.
Authors: Pan Liu (School of Electronic and Information Engineering, Beihang University), Tao Zhang (School of Electronic and Information Engineering, Beihang University),
Hide Authors & Abstract

With popularity of cloud computing services, more and more tasks and services are deployed on large-scale clusters. As an emerging technology in cloud computing field, containers make virtualization extremely lightweight. However, lack of prediction causes scheduling decisions lag behind the dynamics of clouds. Thus, how to carry out performance prediction before container scaling has become an urgent problem to be resolved. Here we emphasized the necessity of container performance prediction and summarized the current research progress and effort of container performance modeling. Finally, we compared pros and cons of numerical analysis and machine learning in terms of practice.
Authors: Jiwei Wang (Hangzhou Dianzi University), Yuegang Li (Hangzhou Dianzi University), Congfeng Jiang (Hangzhou Dianzi University), Chao Ma (Information&Telecommunications Company, State Grid Shandong Electric Power Company), Linlin Tang (Information&Telecommunications Company, State Grid Shandong Electric Power Company), Shuangshuang Guo (Information&Telecommunications Company, State Grid Shandong Electric Power Company),
Hide Authors & Abstract

LoRa is one of the most prominent Low-Power Wide-Area Network (LPWAN) technologies, to accommodate pervasive Internet-of-Things (IoT) connectivities. However, its service capacity and scalability are limited due to the scarce channel resources and the Aloha-like random access mechanism specified by LoRaWAN. We propose a NOMA-enabled LoRa gateway, which permits multiple end-devices to transmit their data at the same time over a shared channel. The whole random access process are provided in detail, including collision resolution and transmission scheduling based on a Distributed Queuing (DQ) method. In addition to that, Spreading Factor (SF) allocation in the transmission scheduling phase is also considered and an optimal problem is formulated to achieve maximum data transmission rate. In order to solve the problem efficiently, an SF allocation algorithm is developed based on the matching theory. Numerical results show that our proposed scheme significantly enhances the sum achievable user rate when the number of users increases.
Authors: Wei Wu (Nanjing University of Posts and Telecommunications), Wennai Wang (Nanjing University of Posts and Telecommunications), Jihai Yang (Nanjing University of Posts and Telecommunications), Bin Wang (Nanjing University of Posts and Telecommunications),
Hide Authors & Abstract

The Compressed Sensing technology in wideband spectrum sensing (WSS) has greatly improved the utilization of spectrum resources. Based on this, we combining sparse learning and fast power spectrum estimation to achieve WSS in this paper. Sparsity adaptive matching pursuit (SAMP) algorithm is exploited to obtain the sparse sample representation for WSS. Then the limi-tations of power spectrum estimation in WSS are considered. To ease the limitations, the computational tasks are decomposed by multiple fast Fourier transforms. Theoretical performance analysis is made to further explain the proposed method. By improving the process of sample collection and power spectrum estimation, the proposed method can effectively achieve the pur-pose of fastly and exactly sensing. The final simulation results are utilized to verify the applicability of the proposed method and its advantages over other methods.
Authors: Shuai Liu (Xi'an Jiaotong University), Wen Xiao (Xi'an Jiaotong University), Yao Zhang (Xi'an Jiaotong University), Jing He (Xi'an Jiaotong University), Jixin Wu (Xi'an Jiaotong University),
Hide Authors & Abstract

IEEE 802.11 wireless local area network (WLAN) is a popular connectivity method because of its convenient deployment, low cost and flexibility. Due to the limited coverage of single access point (AP), multiple APs are often arranged in current WLANs to meet the coverage requirement. In such dense WLANs, the actual situation is that every wireless station (WS) has different quality of service (QoS) requirements for the actual acquired throughput. Thus, this paper proposed a QoS-Guaranteed AP selection algorithm to increase the overall QoS of WSs for the actual acquired throughput. The proposed algorithm relies on a centralized framework and considers the diverse QoS requirements. According to these concrete requirements, the proposed algorithm can distribute WSs to APs fairly and achieve the near-optimal access of AP based on the game theory. Finally, the numerical simulation results verify that the proposed algorithm can effectively improve the overall QoS of WSs.
Authors: Zhihui Weng (Jiangsu University of Science and Technology), Zhibin Xie (Jiangsu University of Science and Technology), Haoran Qin (Jiangsu University of Science and Technology),
Hide Authors & Abstract

The positioning techniques in wireless cellular networks detect and receive signals transmitted by the user equipment (UE) through base stations (BS), then process the time difference of arrival (TDOA) measurements carried in the received signals by the network side. However, if the network side doesn’t report the positioning information to the UE, UE will never know its positioning result. Besides, a BS can accurately locate the UE within its coverage area, the security issue of UE needs to be considered. In order to improve the autonomy and security of the UE, in this paper, we propose an autonomous positioning algorithm for the UE in cellular networks. The UE uses TDOA measurements and multiple UE positioning results sent by the network side to inversely calculate the positions of the BS participating in the positioning. After that, the UE can use calculated BS position coordinates and TDOA measurements to calculate its own position independently. The simulation results show that the method is effective, and the error of the calculated BSs' coordinates is within the acceptable range.
Authors: Yifan Xi (Beijing University of Posts and Telecommunications), Hang Long (Beijing University of Posts & Telecommunications), Tong Li (Beijing University of Posts and Telecommunications),
Hide Authors & Abstract

With the development of cloud computing, more and more video service provid-ers use services from cloud providers. A video service provider can construct a scalable video streaming platform with high availability by the cloud services. Typically, a video service provider uploads its video data to a cloud data center. Then, the cloud data center distributes the video data to its edge cloud CDN nodes. Usually, the cloud data center links with its edge cloud CDN nodes by high-capacity links, spanning different geographical regions. Video traffic across the cloud data center and the edge cloud CDN nodes of a cloud provider, brings on large operational cost to the cloud provider. How to reduce the video traffic cost is important for a cloud provider. Therefore, to reduce the video traffic cost, we propose a set of algorithms based on network maximum flow and minimum cut, called Netcut-way. The proposed Netcut-way, charged by the peak-bandwidth billing model, consists of three parts. The first is peak bandwidth calculation. The second is video segment segmentation. The third is video distribution route. Through extensive simulations, we demonstrate that Netcut-way can effectively reduce the operational cost of cloud providers in video traffic across data centers.
Authors: Pingshan Liu (Guangxi Key Laboratory of Trusted Software, Guilin University of Electronic Technology, Gui-lin, China), Kai Huang (Guangxi Key Laboratory of Trusted Software, Guilin University of Electronic Technology, Guilin, China), Guimin Huang (Guilin University of Electronic University),
Hide Authors & Abstract

With the continuous increase of types of services and data volume in data center, the traffic loads of some links are excessive, and how to balance the link load and ensure the quality of network service have become research hotpots. However, the traditional link load balancing mechanisms ignore the complexity of network and the Quality of Service (QoS) requirement of the flow when calculating the forwarding paths. Therefore, we propose a link load balancing algorithm based on Ant Colony Optimization (LLBA) in data center network. The algorithm redefines the heuristic function according to the number of elephant flows on the link and the real-time load of links, and updates the pheromones according to the path length. Then, the algorithm customizes the optimal path determination rule according to the path transmission delay and the real-time loads, so as to find a best forwarding path for the current flow under the multiple constraints including the path length, link load, and transmission delay. The experiment results show that, the proposed algorithm improves the link utilization and network throughput effectively, and also reduces the delay and delay jitter to some extent, as compared with the traditional mechanisms.
Authors: Shuqing Ma (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications), Hong Tang (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications), Xinxin Wang (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications),
Hide Authors & Abstract

Aiming at the problem that the existing local stereo matching algorithm has low matching accuracy in weak texture, disparity discontinuity and occlusion regions, an improved algorithm based on matching cost calculation and weighted guided filtering is proposed. The algorithm first improves the traditional gradient cost (GRAD) and Census transform, and normalizes and fuses these two matching costs to form a new matching cost, then proposes a weighted guided filter based on the Kirsch operator and aggregates the matching cost, finally, the method of the winner-takes-all (WTA) is used to complete the disparity calculation, and we use the method of left and right disparity consistency and the quadratic curve interpolation to complete the disparity optimization and obtain the final disparity map. A large number of experiments prove that the proposed stereo matching algorithm has an average mismatch rate of about 5.45% relative to the standard disparity map on the test platform of Middlebury. Compared with most algorithms, proposed algorithm achieves a good matching effect.
Authors: Junxing Xu (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications), Wei He (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications), Zengshan Tian (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications),
Hide Authors & Abstract

Edge computing proposes to offload the computing tasks of the vehicle to the vehicle fog node, which can achieve efficient vehicle services and higher utilization of computing resources. However, encouraging vehicles to share resources or execute applications for others remains a sensitive issue due to the selfishness of users. In this paper, we propose an incentive mechanism based on Stackelberg game and the problem of computation offloading of tasks on vehicles is modeled. In order to utilize the idle computing resources of nearby vehicles, for the computation offloading in vehicular networks to utilize the idle computation resources. Specifically, we introduce multi-hop offloading into vehicle-to-vehicle (V2V) offloading, and model the probability of forwarding packets by assisting vehicles from multiple perspectives in order to ensure the stability and reliability of communication transmission. A distributed iterative algorithm is designed to solve the final Stackelberg equilibrium, so as to minimize the cost of requesting vehicles to execute tasks and maximize the service benefits of assisting vehicles. The simulation results show that the proposed computation offloading strategy has superiorities in improving utilization efficiency, reducing tasks execution latency and enhancing service quality.
Authors: Yingdi Dai (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications), Zhanjun Liu (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications), Ya Kang (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications),
Hide Authors & Abstract

Wireless channels that are widely adopted between autonomous vehicles are vulnerable to eavesdropping or interferences, so that attacks on cybersecurity may lead to serious consequences, such as losing control of vehicles. In particular, the cryptographic methods used for information security rely on the strict privacy of keys, which is often difficult to guarantee in a wireless environment. This paper proposes a context-bound cybersecurity framework, which protects communication from eavesdroppers by encrypting critical data with a dynamic context among vehicles. The context is synchronized among vehicles through a progressive encoding method, which makes it difficult for third parties to learn the entire context by eavesdropping through the channels, especially in the case of mobility. The normal vehicles may extract a security key from the context to encrypt and decrypt key data, but it is impossible or overwhelmingly expensive for the third parties to decode the data transmitted due to the lack of the context. Besides, the proposed framework also provides a promising way to resist the upcoming quantum computers, because it will become more and more difficult for third parties to collect the complete context as the context continues to update.
Authors: Longjiang Li (Department of Network Engineering, SICE, University of Electronic Science and Technology of China, Chengdu, 611731, China), Bingchuan Ma (Department of Network Engineering, SICE, University of Electronic Science and Technology of China, Chengdu, 611731, China), Yonggang Li (Chongqing University of Posts and Telecommunications), Yuming Mao (Department of Network Engineering, SICE, University of Electronic Science and Technology of China, Chengdu, 611731, China),
Hide Authors & Abstract

Mobile edge computing has emerged as a new paradigm to enhance computing capabilities by offloading complicated tasks to nearby cloud server. To conserve energy as well as maintain quality of service, algorithms with low time complexity for task offloading is required. In this paper, a multi-user with multiple tasks scenario is considered, taking full account of factors including data size, bandwidth, channel state information, we propose a distributed unsupervised learning-based offloading algorithm for task offloading, where distributed parallel networks are employed to guarantee the robustness of algorithm. Additionally, we exploit a memory pool to store input data and corresponding decisions as key-value pairs. Based on the experience mechanism, the proposed algorithm can omit the step of data calibration compared with supervised learning method. To further reduce the communication cost, we analyze four bandwidth allocation schemes. Results reveal that the channel state-based strategy cost 3 percent less than data size-based, mixed and user size-based. Besides, the proposed algorithm can save 14, 18 percent cost than local and edge schemes respectively. Numerous results show that the proposed algorithm can achieve near-optimal decisions timely as well as having high reliability.
Authors: Jianming Wei (jiangxi university of science and technology), Qiuming Liu (Jiangxi University of Science and Technology), Shumin Liu (Jiangxi University of Science and Technology), Yiping Zeng (Jiangxi University of Science and Technology), Xin Xiong (Jiangxi University of Science and Technology),
Hide Authors & Abstract

Edge computing is to generate faster network service response and meet the basic needs of the industry in real-time business, application intelligence, se-curity and privacy protection. This paper studies the mobile edge computing network, where the computing power of the edge server (ES) is limited, and multiple user equipment (UE) can offload the thinking to the ES in order to save energy consumption and computing delay. The ES needs to determine which UEs can upload its tasks and need to allocate computing resources for these UEs, so this problem is highly coupled and difficult to calculate. This paper proposes an algorithm based on deep reinforcement learning and Se-quential Least SQuares Programming (SLSQP), which decouples and solves the problem. Experiments show that the algorithm works well and can be dy-namically adjusted according to environmental changes. The comparison with other algorithms also proves that the algorithm has better results and less time-consuming.
Authors: Zhihao Xuan (Zhejiang Gongshang University), Guiyi Wei (Zhejiang Gongshang University), Zhengwei Ni (Zhejiang Gongshang University), Jifa Zhang (Zhejiang University),
Hide Authors & Abstract

Versatile Video Coding (VVC) achieves better performance with about 40% bitrate reduction compared to H.265/HEVC under the same video quality. The improvement of VVC coding performance is at the cost of increased computational complexity of the VVC encoder. To address this issue, a fast algorithm for VVC intra prediction based on gradient information of coding unit (CU) is proposed in this paper. Experimental results illustrate that the proposed algorithm can save 47.78% coding time on average as compared with VTM7.0 with 2.25% increase in BDBR and 0.09dB loss in PSNR.
Authors: wang shiyu (Chongqing University of Posts and Telecommunications), qiang li (Chongqing University of Posts and Telecommunications),
Hide Authors & Abstract

The abstract should summarize the contents of the paper in short terms, i.e. 150-250 words. Multipath is a major source of positioning error in high precision navigation applications. Narrow correlator method and Gating signal correlator method are two effective methods for BPSK modulation signals multipath mitiga-tion. In this paper, the mathematic model of multipath error are first established, and then multipath mitigation performance of the BPSK signal and BOC modula-tion signal are analyzed and simulated based on the narrow correlator method in a comparative way. The simulation results show that BOC signal is better than BPSK signal as for multipath mitigation performance. The consistency between narrow correlator and gating signal correlator method is deduced and proven the-oretically. And it is concluded that the BOC code tracking loop phase discrimina-tor function has ambiguity using the gating signal multipath mitigation method, so the gating signal multipath mitigation method does not work for BOC signal tracking code loop phase discriminator. Finally, the correctness of theoretical der-ivation is verified through simulation.
Authors: Xi yan Sun (Guilin University of Electronic Technology), Shao jie Song (Guilin University of Electronic Technology), Yuan fa Ji (Guilin University of Electronic Technology),
Hide Authors & Abstract

There are many uncertain factors in the evaluation process of integrated equipment system of systems (IES), owing to lacking the effective evaluation method. Considering the expert evaluation process is often subjective, so take the combination of entropy weight method, the Gini coefficient weighting method and AHP method are used to calculate the weight of the combat capability index; the expert evaluation information is also vague, and the vague set theory can well describe the support, neutral and opposition information. Therefore, the combination of vague set and TOPSIS method is used to calculate the degree of closeness to measure the importance of IES; Given that the combat process, equipment may be failed. The fault function is introduced to evaluate the contribution of IES dynamically by defining the new fault function and the recurrent fault function. Finally, through the case analysis, it is proved that the proposed algorithm can more accurately evaluate the contribution of IES.
Authors: Shang wei Luo (School of Communication and Information Engineering，Chongqing University of Posts and Telecommunications), Yonggang Li (Chongqing University of Posts and Telecommunications), Yanyan Chen (School of Communication and Information Engineering，Chongqing University of Posts and Telecommunications),
Hide Authors & Abstract

We investigate machine-learning-based cross-layer energy-efficient transmission design for vehicular communication systems. A typical vehicle-to-vehicle (V2V) communication scenario is considered, in which the source intends to deliver two types of messages to the destination to support different safety-related applications. The first are periodically-generated heartbeat messages, and should be transmitted immediately with sufficient reliability. The second type are randomly-appeared sensing messages, and are expected to be transmitted with limited latency. Due to node mobility, accurate instantaneous channel knowledge at the transmitter side is hard to attain in practice. The transmit channel state information (CSIT) often exhibits certain delay. We propose a transmission strategy based on the deep reinforcement learning technique such that the unknown channel variation dynamics can be learned and transmission power and rate can be adaptive chosen according to the message delay status to achieve high energy efficiency. The advantages of our method over several conventional and heuristic approaches are demonstrated through computer simulations.
Authors: Siyuan Yu (Tongji University), nong qu (Tongji University), Yizhong Zhang (Tongji University), Chao Wang (Tongji University, China), Fuqiang Liu (Tongji University),
Hide Authors & Abstract

Compressive sensing (CS) is a new theory of data acquisition and reconstruction. It permits the data of interest being sampled at a sub-Nyquist rate, meanwhile still allowing perfect reconstruction of data from highly incomplete measurements. During this process, the construction of measurement matrix is undoubtedly the key point. However, the traditional random measurement matrices, though having good performance, are difficult to implement in hardware and lack the ability of dealing with large signals. In this paper, we construct a series of novel measurement matrices(HWKM and HWCM) based on Welch bound, by sifting the basis matrix based on Hadamard matrix. Therefore, the proposed matrices are deterministic measurement, which can be easily designed in hardware. Specially, it is proved to have low coherence, which can even approach to Welch bound. Experimental results show that the proposed matrices, compared with traditional measurement matrices, not only have considerable reconstruction performance in terms of reconstruction error and the signal-to-noise ratio, but also accelerate recovery time.
Authors: Han Zhang (Xidian University, China), Song Xiao (Xidian University, China), Ping Gan (Xidian University, China),
Hide Authors & Abstract

Since the measured phase is usually a wrapped phase during ranging, the phase-based ranging method needs to solve the ambiguity so as to obtain the true phase. HMFCW (heuristic multi-frequency continuous wave) algorithm provides the same tolerance of error for the observed phase of different frequencies. When the error of phase is within the tolerance and the range of ranging is less than the period, the phase error tolerance method can get the correct cycle number, and achieve a ranging accuracy of centimeter. However, when phase errors of some frequencies are large, HMFCW algorithm may have difficulty in solving the integer ambiguity, which leads to the decrease of ranging accuracy. In this paper, an improved HMFCW algorithm based on HMFCW algorithm is proposed. The improved HMFCW algorithm calculates the average value of the clustering phase results to eliminate the phases with large errors, and performs the cycle calculation to obtain the ranging value. Simulation results show that improved HMFCW algorithm can solve the problem of error in the cycle ambiguity solution caused by the point with large phase error effectively, and improve the ranging accuracy.
Authors: Zengshan Tian (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China), Shuwen Wu (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China), Liangbo Xie (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China), Xixi Liu (School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China),
Hide Authors & Abstract

Nowadays, mounting a base-station (BS) onto an unmanned aerial vehicle (UAV) has become a new dimension for constructing the new generation of wireless communication networks. For example, dynamically constructing an UAV-BS network can provide some instantaneous and emergent communication services when the infrastructure of the cellular network is destroyed owing to some devastating disasters. In this paper, we aim to deploy a scalable and self-organized UAV-assisted device-to-device (D2D) communication network using a minimum number of UAV-BSs (UBSs) under some link quality constraints. Specifically, a sequential UBS deploying algorithm is designed to guarantee the signal quality for the links between the UBSs and ground terminals, and those between UBSs and central controller. Via the simulation results, it is interesting to find that how to deploy a proper number of UBSs rather than constructing a highly connected UBS network is the key to guarantee higher spectrum efficiency for the UBS-assisted D2D networks.
Authors: Wenson Chang (National Cheng Kung University), Kuang-Chieh Liu (Realtek Semiconductor Co. Limited, Taiwan R.O.C), Zhao-Ting Meng (SiliTai Electronics Co. Limited, Taiwan R.O.C), Li-Chun Wang (National Chiao Tung University),
Hide Authors & Abstract

A new early stopping criterion is proposed for the Noisy Gradient Descent Bit Flipping (NGDBF) Decoding of Low-Density Parity-Check (LDPC) codes to reduce the number of decoding iterations. The new criterion is based on the number of flipped bits at certain iterations, and has extremely low complexity. It is shown in the simulation results that the proposed early stopping criterion can significantly reduce the number of decoding iterations at low signal-to-noise ratios (SNRs), and only a slight bit error rate (BER) performance decrease is experienced at high SNRs.
Authors: Li Zhang (Southeast University), Nan Liu (Southeast University), Zhiwen Pan (Southeast University), Xiaohu You (Southeast University),
Hide Authors & Abstract

Pilot pollution is an urgent problem to be solved in a massive MIMO system. In this paper, a pilot allocation scheme based on machine learning algorithm and users’ angle of arrival is proposed for the pilot pollution problem of massive MIMO. The scheme firstly classifies all users according to whether the angle of arrival overlaps. It randomly assigns pilot sequences to users whose angle of arrival do not overlap each other and continue to use machine learning algorithm to divide users whose angle of arrival overlap with each other into interference groups and non-interference groups based on user location information. We assign orthogonal pilots to users in the interference groups and randomly assign pilot sequences to users in the non-interference groups. Simulation results show that the pilot allocation scheme proposed in this paper can effectively suppress the impact of pilot pollution on the performance of massive MIMO systems and improve pilot efficiency and reduce pilot overhead.
Authors: Min YU (Chongqing University of Posts and Telecommunications), Siyuan LI (Chongqing University of Posts and Telecommunications), Dongfeng CHEN (Chongqing University of Posts and Telecommunications),
Hide Authors & Abstract

The measurement of the importance for the nodes is of great significance to the test and simulation for Heterogeneous Combat Networks (HCN), combat situation assessment and other topics. Due to the complexity of equipment types and styles in such system, traditional algorithms (degrees, betweenness, closeness, eigenvectors) are difficult to achieve both speed and accuracy in identifying the important nodes of Heterogeneous Combat Networks. This paper fully considers the heterogeneity of combat system nodes, and proposes an evaluation model based on community structure, IEBC (importance evaluation based on community), which can measure the importance of each node. We form functional modules (FM) by distinguishing the function of nodes. Then divide the network into communities according to the concentration of FM. Finally, we compare IEBC with traditional ranking models (e.g., degree centrality). After simulation calculation, compared with other algorithms, IEBC takes into account the balance of efficiency and accuracy at the same time.
Authors: zhaofeng yang (Chongqing University of Posts and Telecommunications), jinyu liu (Chongqing University of Posts and Telecommunications), Yonggang Li (Chongqing University of Posts and Telecommunications),
Hide Authors & Abstract

For the 5G NR systems in the Third Generation Partnership Project(3GPP), the computation of synchronous signal detection algorithms is very complex. Based on the characteristics of the domain synchronous signal, this paper proposed a differential and superimposed interdependent cross correlation detection algorithm, which only performed a differential interdependent processing of the received signal with the sum of the local PSS sequence to obtain the position of the relevant peak and detects the position of the coarse synchronous point, and further performed local correlation of the coarse synchronous point, and detected the precise synchronous point according to the maximum peak, which reduced the computational complexity. A theoretical derivation and a comparative complexity analysis between the traditional cross correlation and the improved algorithms shows that the differential and superposition cross correlation joint detection algorithm has high efficiency, low complexity and strong resistance to frequency deviation, which meets the synchronization requirements of 5G NR systems.
Authors: Jiang Hang (School of Communication and Information Engineering Chongqing University of Posts and Telecommunications), Zhang Zhizhong (School of Communication and Information Engineering Chongqing University of Posts and Telecommunications), Deng Bingguang (School of Communication and Information Engineering Chongqing University of Posts and Telecommunications), Cao Longhan (People's Liberation Army Chongqing Communication Institute),
Hide Authors & Abstract

Convolutional neural network (CNN) has been widely used in deep learning. However, the hardware consumption of the convolutional neural network is very large. Traditional Central Processing Units (CPUs) and Graphic Processing Units (GPUs) are inefficient and expensive for neural network, so an efficient hardware design is required. The proposed design based on Digital Signal Processor (DSP) has rapid operating speed and strong computation ability for training and inference of CNN. In this paper, the hardware architecture of convolution and softmax function is specially optimized. Winograd algorithm can reduce multiplications of convolution, thus decreases hardware complexity, since multiplication is much more complex in hardware implementation than addition. The softmax function is also simplified by replacing divider by subtractor and logarithmic function which cost fewer resources. The proposed hardware architecture dramatically decreases the complexity and hardware resources.
Authors: Zhenyu Jiang (Tongji University/College of Electronic and Information Engineering/Shanghai, China), Zhifeng Zhang (Tongji University/College of Electronic and Information Engineering/Shanghai, China), Ren Haoqi (Tongji University/College of Electronic and Information Engineering/Shanghai, China), Jun Wu (School of Computer Science, Fudan University, Shanghai, China),
Hide Authors & Abstract

The high computational complexity and the mismatch between the space-time distribution of tasks and resources raise great challenges on task planning in space information networks (SINs). This paper studies the task planning problem in SINs by exploiting resource interchange to handle the bottleneck resources. First of all, we develop a time-varying resource graph model to capture the dynamic coordination relationship among resources in SIN. Then, we explore resource interchange and derive its quantitative condition. On this basis, an optimization model for task planning based on resource interchange is formulated. Furthermore, we decompose the task planning problem into two stages for global optimization and local adjustment, and develop the algorithms respectively. Finally, simulation results show that compared with existing works the proposed algorithm strikes a better balance between the number of completed tasks and computational complexity.
Authors: Runzi Liu (Xi'an Unverisity of Architecture and Technology), Jing Li (Xidian University), Xiang Ji (Xi'an University of Architecture and Technology), Weihua Wu (Xidian University), Di Zhou (Xidian University), Yan Zhang (Xidian University),
Hide Authors & Abstract

Low latency communication has attracted much interest in recent years due to the emergence of new types of delay-sensitive applications. Ultra-Reliable and Low-Latency Communication (URLLC) is also considered as one of the important use-cases in 5G cellular system. Traditional wireless communication system is usually optimized for high data throughput, but cannot satisfy the requirement of strict latency threshold. Semi-persistent scheduling and TTI shortening are two methods to address this problem. We implemented these methods by making modification based on OpenAirInterface framework, and build up a realistic cellular network system using software-defined radio technology. Experimental re-sults show a dramatic latency reduction comparing with the baseline LTE scheme. By integrating these effective methods, we implemented a practical wireless communication system that can provide low latency transmission service.
Authors: Yujie Liu (College of Electronic and Information Engineering, Tongji University, Shanghai, China), Jun Yu (College of Electronic and Information Engineering, Tongji University, Shanghai, China), Fusheng Zhu (Guangdong Communications & Network Institute), Wenru Zhang (Guangdong Communications & Network Institute), Jun Wu (School of Computer Science, Fudan University, Shanghai, China; Pengcheng Laboratory, Shenzhen, China),
Hide Authors & Abstract

The technical route of machine learning compiler generally refers to the application of automatic or semi-automatic code generation in the optimization process instead of hand-optimization. This paper presents a deep learning compiler (DLCS) for target vector processor based on LLVM framework, which lowers deep learning (DL) models to an intermediate representation (IR) of two levels. The high-level IR realizes target-independent optimizations including kernel fusion, data replacement and data simplification, while the low-level IR allows the compiler to perform target-dependent optimizations, such as Eight-Slots VLIW and special intrinsic function. The proposed compiler customizes the architecture description of target vector processor to achieve a high-quality automatic code generation. We evaluate the performance comparison between DLCS and hand-optimization when deploying ResNet-18 model and MobileNet model to the target vector processor. Experimental results show that DLCS offers Multi-slot parallel performance for target vector processor and achieves speedups ranging from 1.5× to 3.0× over existing frameworks backed by hand-optimized libraries.
Authors: Pingping Pan (Department of Computer Science, Tongji University, Shanghai, China), Jun Wu (School of Computer Science, Fudan University, Shanghai, China), Songyuan Zhao (Department of Computer Science, Tongji University, Shanghai, China), Haoqi Ren (Department of Computer Science, Tongji University, Shanghai, China), zhifeng zhang (Department of Computer Science, Tongji University, Shanghai, China),
Hide Authors & Abstract

Sinusoidal signal frequency estimation is one of the fundamental problems in signal processing, and it is widely used in wireless communication, signal pro-cessing, navigation, radar and so on. In this paper, an interpolation frequency es-timation algorithm based on Discrete Fourier Transform (DFT) and cosine win-dows is proposed. Firstly, the sampling sequence of the signal is multiplied by a cosine window. Then, N-point DFT is used to search the position of the maxi-mum spectral line and get the coarse estimation of frequency. Finally, the accurate frequency estimation is obtained by DFT interpolation of the maximum spectral line and the two Discrete-Time Fourier Transform (DTFT) samples on the left and right of the maximum spectral line. According to the simulation results, the performance of the proposed algorithm is better than that of MV-IpDTFT(3) al-gorithm, MV-IpDTFT(2) algorithm and Candan algorithm. The effect of harmon-ic interference on the frequency estimation results can be effectively suppressed.
Authors: Jinyu Liu (Dalian Polytechnic University), Lei Fan (Dalian Polytechnic University), Renqing Li (Dalian Polytechnic University), Wenbo He (Dalian Polytechnic University), Nian Liu (Dalian Polytechnic University), Zhanhong Liu (Dalian Polytechnic University),
Hide Authors & Abstract