arXiv subjects

Xiaoli Xu

Publications and source records attributed to Xiaoli Xu.

48 records · Page 3Linked to original sources

arXiv2019

Mobile Video Action Recognition

Video action recognition, which is topical in computer vision and video analysis, aims to allocate a short video clip to a pre-defined category such as brushing hair or climbing stairs. Recent works focus on action recognition with deep neural networks that achieve state-of-the-art results in need of high-performance platforms. Despite the fast development of mobile computing, video action recognition on mobile devices has not been fully discussed. In this paper, we focus on the novel mobile video action recognition task, where only the computational capabilities of mobile devices are accessible. Instead of raw videos with huge storage, we choose to extract multiple modalities (including I-frames, motion vectors, and residuals) directly from compressed videos. By employing MobileNetV2 as backbone, we propose a novel Temporal Trilinear Pooling (TTP) module to fuse the multiple modalities for mobile video action recognition. In addition to motion vectors, we also provide a temporal fusion method to explicitly induce the temporal context. The efficiency test on a mobile device indicates that our model can perform mobile video action recognition at about 40FPS. The comparative results on two benchmarks show that our model outperforms existing action recognition methods in model size and time consuming, but with competitive accuracy.

cs.CV↗

arXiv2019

Path Design for Cellular-Connected UAV with Reinforcement Learning

This paper studies the path design problem for cellular-connected unmanned aerial vehicle (UAV), which aims to minimize its mission completion time while maintaining good connectivity with the cellular network. We first argue that the conventional path design approach via formulating and solving optimization problems faces several practical challenges, and then propose a new reinforcement learning-based UAV path design algorithm by applying \emph{temporal-difference} method to directly learn the \emph{state-value function} of the corresponding Markov Decision Process. The proposed algorithm is further extended by using linear function approximation with tile coding to deal with large state space. The proposed algorithms only require the raw measured or simulation-generated signal strength as the input and are suitable for both online and offline implementations. Numerical results show that the proposed path designs can successfully avoid the coverage holes of cellular networks even in the complex urban environment.

cs.NI↗

arXiv2018

V2X Content Distribution Based on Batched Network Coding with Distributed Scheduling

Content distribution is an application in intelligent transportation system to assist vehicles in acquiring information such as digital maps and entertainment materials. In this paper, we consider content distribution from a single roadside infrastructure unit to a group of vehicles passing by it. To combat the short connection time and the lossy channel quality, the downloaded contents need to be further shared among vehicles after the initial broadcasting phase. To this end, we propose a joint infrastructure-to-vehicle (I2V) and vehicle-to-vehicle (V2V) communication scheme based on batched sparse (BATS) coding to minimize the traffic overhead and reduce the total transmission delay. In the I2V phase, the roadside unit (RSU) encodes the original large-size file into a number of batches in a rateless manner, each containing a fixed number of coded packets, and sequentially broadcasts them during the I2V connection time. In the V2V phase, vehicles perform the network coded cooperative sharing by re-encoding the received packets. We propose a utility-based distributed algorithm to efficiently schedule the V2V cooperative transmissions, hence reducing the transmission delay. A closed-form expression for the expected rank distribution of the proposed content distribution scheme is derived, which is used to design the optimal BATS code. The performance of the proposed content distribution scheme is evaluated by extensive simulations that consider multi-lane road and realistic vehicular traffic settings, and shown to significantly outperform the existing content distribution protocols.

cs.IT↗

arXiv2018

Overcoming Endurance Issue: UAV-Enabled Communications with Proactive Caching

Wireless communication enabled by unmanned aerial vehicles (UAVs) has emerged as an appealing technology for many application scenarios in future wireless systems. However, the limited endurance of UAVs greatly hinders the practical implementation of UAV-enabled communications. To overcome this issue, this paper proposes a novel scheme for UAV-enabled communications by utilizing the promising technique of proactive caching at the users. Specifically, we focus on content-centric communication systems, where a UAV is dispatched to serve a group of ground nodes (GNs) with random and asynchronous requests for files drawn from a given set. With the proposed scheme, at the beginning of each operation period, the UAV pro-actively transmits the files to a subset of selected GNs that cooperatively cache all the files in the set. As a result, when requested, a file can be retrieved by each GN either directly from its local cache or from its nearest neighbor that has cached the file via device-to-device (D2D) communications. It is revealed that there exists a fundamental trade-off between the file caching cost, which is the total time required for the UAV to transmit the files to their designated caching GNs, and the file retrieval cost, which is the average time required for serving one file request. To characterize this trade-off, we formulate an optimization problem to minimize the weighted sum of the two costs, via jointly designing the file caching policy, the UAV trajectory and communication scheduling. As the formulated problem is NP-hard in general, we propose efficient algorithms to find high-quality approximate solutions for it. Numerical results are provided to corroborate our study and show the great potential of proactive caching for overcoming the limited endurance issue in UAV-enabled communications.

cs.IT↗

arXiv2017

Trajectory Optimization for Completion Time Minimization in UAV-Enabled Multicasting

This paper studies an unmanned aerial vehicle (UAV)-enabled multicasting system, where a UAV is dispatched to disseminate a common file to a number of geographically distributed ground terminals (GTs). Our objective is to design the UAV trajectory to minimize its mission completion time, while ensuring that each GT is able to successfully recover the file with a high probability required. We consider the use of practical random linear network coding (RLNC) for UAV multicasting, so that each GT is able to recover the file as long as it receives a sufficiently large number of coded packets. However, the formulated UAV trajectory optimization problem is non-convex and difficult to be directly solved. To tackle this issue, we first derive an analytical lower bound for the success probability of each GT's file recovery. Based on this result, we then reformulate the problem into a more tractable form, where the UAV trajectory only needs to be designed to meet a set of constraints each on the minimum connection time with a GT, during which their distance is below a designed threshold. We show that the optimal UAV trajectory only needs to constitute connected line segments, thus it can be obtained by determining first the optimal set of waypoints and then UAV speed along the lines connecting the waypoints. We propose practical schemes for the waypoints design based on a novel concept of virtual base station (VBS) placement and by applying convex optimization techniques. Furthermore, for given set of waypoints, we obtain the optimal UAV speed over the resulting path efficiently by solving a linear programming (LP) problem. Numerical results show that the proposed UAV-enabled multicasting with optimized trajectory design achieves significant performance gains as compared to benchmark schemes.

cs.IT↗

arXiv2015

Reliable Broadcast to A User Group with Limited Source Transmissions

In order to reduce the number of retransmissions and save power for the source node, we propose a two-phase coded scheme to achieve reliable broadcast from the source to a group of users with minimal source transmissions. In the first phase, the information packets are encoded with batched sparse (BATS) code, which are then broadcasted by the source node until the file can be cooperatively decoded by the user group. In the second phase, each user broadcasts the re-encoded packets to its peers based on their respective received packets from the first phase, so that the file can be decoded by each individual user. The performance of the proposed scheme is analyzed and the rank distribution at the moment of decoding is derived, which is used as input for designing the optimal BATS code. Simulation results show that the proposed scheme can reduce the total number of retransmissions compared with the traditional single-phase broadcast with optimal erasure codes. Furthermore, since a large number of transmissions are shifted from the source node to the users, power consumptions at the source node is significantly reduced.

cs.IT↗

arXiv2015

Two-Phase Cooperative Broadcasting Based on Batched Network Code

In this paper, we consider the wireless broadcasting scenario with a source node sending some common information to a group of closely located users, where each link is subject to certain packet erasures. To ensure reliable information reception by all users, the conventional approach generally requires repeated transmission by the source until all the users are able to decode the information, which is inefficient in many practical scenarios. In this paper, by exploiting the close proximity among the users, we propose a novel two-phase wireless broadcasting protocol with user cooperations based on an efficient batched network code, known as batched sparse (BATS) code. In the first phase, the information packets are encoded into batches with BATS encoder and sequentially broadcasted by the source node until certain terminating criterion is met. In the second phase, the users cooperate with each other by exchanging the network-coded information via peer-to-peer (P2P) communications based on their respective received packets. A fully distributed and light-weight scheduling algorithm is proposed to improve the efficiency of the P2P communication in the second phase. The performance of the proposed two-phase protocol is analyzed and the channel rank distribution at the instance of decoding is derived, based on which the optimal BATS code is designed. Simulation results demonstrate that the proposed protocol significantly outperforms the existing schemes. Lastly, the performance of the proposed scheme is further verified via testbed experiments.

cs.IT↗

arXiv2014

Degrees of Freedom of the 3-User Rank-Deficient MIMO Interference Channel

We provide the degrees of freedom (DoF) characterization for the $3$-user $M_T\times M_R$ multiple-input multiple-output (MIMO) interference channel (IC) with \emph{rank-deficient} channel matrices, where each transmitter is equipped with $M_T$ antennas and each receiver with $M_R$ antennas, and the interfering channel matrices from each transmitter to the other two receivers are of ranks $D_1$ and $D_2$, respectively. One important intermediate step for both the converse and achievability arguments is to convert the fully-connected rank-deficient channel into an equivalent partially-connected full-rank MIMO-IC by invertible linear transformations. As such, existing techniques developed for full-rank MIMO-IC can be incorporated to derive the DoF outer and inner bounds for the rank-deficient case. Our result shows that when the interfering links are weak in terms of the channel ranks, i.e., $D_1+D_2\leq \min(M_T, M_R)$, zero forcing is sufficient to achieve the optimal DoF. On the other hand, when $D_1+D_2> \min(M_T, M_R)$, a combination of zero forcing and interference alignment is in general required for DoF optimality. The DoF characterization obtained in this paper unifies several existing results in the literature.

cs.IT↗

arXiv2013

A Method for Calculating Collision Probability Between Space Objects

A method is developed to calculate collision probability in this paper. Based on the encounter geometric features of space objects, it is reasonable to separate the radial orbital motions from that in the cross section for most encounter events in near circular orbit. Therefore, the collision probability caused by orbit altitude difference in the radial direction and the collision probability caused by arrival time difference in the cross section are calculated respectively. The net collision probability is expressed as an explicit expression by multiplying the above two components. Numerical cases are applied to test this method by comparing the results with the general method. The results indicate that this method is valid for most near circular orbital encounter events.

astro-ph.IM↗

arXiv2012

Detailed Steps of the Fourier-Motzkin Elimination

This file provide the detailed steps for obtaining the bounds on $R_1$, $R_2$ via the obtained results on $(R_{1c},R_{1p},R_{2c},R_{2p})$. It is subplementary material for the paper titled "On the Capacity Region of Two-User Linear Deterministic Interference Channel and Its Application to Multi-Session Network Coding"

cs.IT↗

arXiv2012

On the Capacity Region of Two-User Linear Deterministic Interference Channel and Its Application to Multi-Session Network Coding

In this paper, we study the capacity of the two-user multiple-input multiple-output (MIMO) linear deterministic interference channel (IC), with possible correlations within/between the channel matrices. The capacity region is characterized in terms of the rank of the channel matrices. It is shown that \emph{linear precoding} with Han-Kobayashi type of rate-splitting, i.e., splitting the information-bearing symbols of each user into common and private parts, is sufficient to achieve all the rate pairs in the derived capacity region. The capacity result is applied to obtain an achievable rate region for the double-unicast networks with random network coding at the intermediate nodes, which can be modeled by the two-user MIMO linear deterministic IC studied. It is shown that the newly proposed achievable region is strictly larger than the existing regions in the literature.

cs.IT↗

arXiv2012

Reduced Functional Dependence Graph and Its Applications

Functional dependence graph (FDG) is an important class of directed graph that captures the dominance relationship among a set of variables. FDG is frequently used in calculating network coding capacity bounds. However, the order of FDG is usually much larger than the original network and the computational complexity of many bounds grows exponentially with the order of FDG. In this paper, we introduce the concept of reduced FDG, which is obtained from the original FDG by keeping only those "essential" edges. It is proved that the reduced FDG gives the same capacity region/bounds with the original FDG, but requiring much less computation. The applications of reduced FDG in the algebraic formulation of scalar linear network coding is also discussed.

cs.IT↗