Kai Pan

A kernel-based l 2 norm regularized least square algorithm for vehicle logo recognition

We consider the problem of automatically recognizing the vehicle logos from the frontal views with varying illumination, as well as certain corruption. To better address the problem, a kernel-based l2 norm regularized least square (RLS) algorithm is proposed in the paper. Kernel technique is smoothly combined with the l2 norm RLS algorithm to enhance the performance of vehicle logo recognition (VLR). As an extension, the improvement of dictionary is also considered. A simple mechanism of constructing an adaptive online dictionary has been presented and experimented. Experimental results show that our proposed algorithm outperforms the original l2 norm RLS algorithm and the l1 norm based algorithms.

QoS Oriented Embedding for Network Virtualization

Network virtualization (NV) is an indispensable approach to handle the ossification of current Internet. Previous research has been conducted concerning the implementation issues like heuristic-based mapping algorithms, yet little has focused on quality of service (QoS) for end users who launch virtual network (VN) requests, and revenue both service providers and infrastructure providers can obtain. Through experiments we find that with the same mapping algorithm, varying mapping order tends to bring about different substrate resources allocation and correspondingly, different acceptance ratio and revenue. However, existing research has not yet paid enough attention to this issue. In this paper, we address the issue by proposing a new and comprehensive revenue model, which is employed as a ranking criterion for VN requests before mapping. Additionally, we define the notion of fairness ratio to quantify the QoS that our model centers. The evaluation results show that our practice outperforms exiting ones with desirable fairness and user satisfaction along with higher revenue and increased acceptance ratio (6.5%).

Prologue: Unified Polymorphic Routing Towards Flexible Architecture of Reconfigurable Infrastructure

Today’s Internet architecture was designed and proposed in the 60s and 70s with the intention to interconnect several computing resources across a geographically distributed user group. With the advent of substantially various Internet businesses, traditional Internet is increasingly powerless to satisfy the unprecedented demands. This paper probed the polymorphic routing prototype based on proposed Flexible Architecture of Reconfigurable Infrastructure (FARI) which attempts to emerge as a clean-slate revolution of future Internet and resorts to centralized control manner. Routers in FARI were reconfigurable to adapt to different businesses in terms of identifier type. Moreover, a preliminary framework of FARI is proposed in the end of the article.

Exploring Cache Coding Scheme for Information-centric Networking

The tremendous growth of the Internet application has driven wider efforts to advanced future Internet architecture especially Content-Centric Networking. In this paper we propose a Centralized Architecture for ICN (CAN), which efficiently combines Content-Centric Networking and network coding (NC). In order to achieve optimal coding performance, we design a global controller which is mainly responsible for calculating entire network routing tables and controlling optimal trade off among cache coding schemes. Specifically, we propose a coding chunk naming scheme and an efficient network coding management approach, which apply to recover the lost native content packets according to NC. Experiment results show that NC in CAN has a great performance for content packet loss recovery. In particular, it can reach nearly 100% recovery in sufferable packet loss rate. Meanwhile, the efficiency and performance of which are superior to the conventional CCN.

Flying against lossy light-load hybrid networks

Emerging as an important potential approach to the operation of lossy communication networks, Network Coding (NC) has been proposed to combine with TCP called TCP-NC by MIT. However, the key parameter R that is brought in for sending redundant packets against random loss is preseted and stays constant, therefore, chances are that it may fail to work. In this paper, we proposed a dynamic redundancy algorithm to control the number of sending packets. The algorithm is executed at the time of every arrival of feedback and makes redundancy approximate the loss rate as much as possible. All of these is done by the sender who is completely unconscious of the network situation. Simulation results under hybrid networks show that our new protocol retains all the advantages of TCP-NC over TCP-Reno and TCP-Westwood in terms of channel utilization and overall throughput, and simultaneously outperforms TCP-NC in time-varying light-load hybrid networks.

An efficient transmission scheme for data aggregation in wireless sensor networks

In this paper, we introduce a novel algorithm aiming at maximizing the transmission capacity in wireless sensor networks (WSNs). Based on the Dijkstra algorithm and the max-flow theorem, the proposed algorithm solves the contradiction between the path length and forwarding capacity of various nodes, and thus scales up the relay capacity in terms of the number of transmitted packets from the source node to the sink node. Specifically, our algorithm first transforms the network from an undirected graph to a directed one, and then trims the excess paths in order to achieve the optimal capacity. We evaluate the proposed algorithm atop a testbed, and the experimental results show the significant performance enhancement by using the proposed algorithm.

Content Consistency Strategy for data storage in Content-Centric Networking

In Content Centric Networking (CCN), registered data objects are frequently modified by users due to individual requirements, which results in downloading vulnerability to nodes. The overall consistency of data stored in multiple nodes is affected by two major factors, the changes in data content and the loss of data on each host, which significantly impacts the correctness and usability in the networks. To solve this problem, we apply optimization problem to minimize the overall transmission error of registered data delivered from the source node to the sink node in CCN. Specifically, we first describe the problems of transmission error rate and transmission time based on CCN. Then, we propose a specific strategy to unify the stored data and optimize system performance, and then achieve the minimized overall transmission fault for a certain amount of data. Finally, comprehensive simulations show that significant performance enhancement can be achieved by using our proposed scheme.

Toward a Loss-Free Packet Transmission via Network Coding

Network coding promises significant benefits in network performance, especially in lossy environment. As the Transmission Control Protocol (TCP) forms the central part of the Internet protocol, it is necessary to find out the way that makes these benefits compatible with TCP. This chapter introduces a new mechanism for TCP based on network coding which only requires minor changes to the protocol to achieve incremental deployment. The center of the scheme is transmitting linear combination of original packets in the congestion window and simultaneously generating redundant combinations to mask random losses from TCP. Original packets in the congestion window can be deleted even before it is decoded at the receiver side, since the receiver acknowledges the degree of combinations instead of packet itself. Thus, all the original packets can be obtained once enough combinations are collected. Simulation results show the scheme achieves much higher throughput than original TCP in lossy network. Though it still seems far from being deployed in the real network, it has finished the first step in taking the concept of NC into practice.

Little for more: An active discovery strategy for content routing in ICN

While receiving a certain packet with unknown destination, a router either simply drops it or sends queries around as a remedy.Dropping it sacrifices transmission stability, but sending queries may cause too much overload.Such situation can be seen more frequently in Information-Centric Networking(ICN) since it has a larger namespace for routing destinations.In this paper, we propose the neighboring chunk aware discovery(NCAD), an active discovery strategy for content routing. NCAD can reduce the cost of resolving unknown contents, which is not yet considered by the conventional routing mechanism. The idea behind NCAD is to ask the nodes to report back more possession information about the queried content in prediction of later use. To this end, we de-sign NCAD with several different considerations,and balance the tradeoff between effectiveness in conjunction with the overhead to achieve optima on the whole. Extensive experiments are conducted on a small typical topology and a large real-world Internet AS-level topology.Experimental results show an overall superiority of NCAD in terms of both effectiveness and overhead.

Towards minimizing the within-packet redundancy in wireless network coding

Network coding, by right of the concise form and remarkable effect, has been applauded around the world since its emergence especially in the field of wireless communication. However, we find the superiority of network coding may be concealed by the over redundant information in coded packets under certain circumstances. That is, the back-and-forth coding at intermediate nodes makes the old information live forever in the network, occupying buffers and sophisticating decoding at receiving nodes. In this paper, we concentrate on the topic that has not aroused enough attention and introduce a simple method to trim the redundant information resort to Information-Centric Networking (ICN) with the invertibility of network coding before encoding again. After quantifying the calculation overhead of the method, we conduct experiments to validate the analysis in conjunction with the transmission efficiency. Simulation results demonstrate an approximate linear characteristic for the calculation overhead which we believe is affordable for hardware.