CNTK

The CNTK (Centre for Network and Service Convergence) project is funded by the Danish Ministry of Science, Technology and Development presently involves three industrial partners, two research partners and the Danish Technology Institute. The project is running during the period from August 1, 2002 until October 1, 2006. To develop and optimize packet switched networks with multi layer architecture independent of access technologies/methods is one of the great challenges for telco?s in the years to come. The project focus on the provision of development and promotion of new and optimized (wireless) services that will increase the flexibility and the accessibility for the end-user offering portable end-to-end wireless networks and services. The technology and the established tools will facilitate development of the market place for systems suppliers, operators, and service and content providers. The goal of the CNTK project is to create access independent adaptive and flexible network and in that context necessary information the traffic and the network will be collected via the industrial partners. The research efforts will focus on traffic-/network analysis and modelling, QoS and security. From a market perspective the goal of the CNTK project is a real extension of the wireless market ? at lower prices. CNTK will develop a system which will increase the total use of capacity by allowing concurrent transactions at securing security, adaptability, flexibility and an end-to-end QoS without any severe modification of standards. Generally the purpose of the CNTK project is to develop principles for and prototypes of tools and methods for the planning and optimizing of packet switched access and core networks by net operators. This purpose is based on the assumption that the traffic on the network will be a mixture of traffic types with varying demands for QoS. The focus of research within the SNX group is on access network-based services via speech enabled human-computer interfaces that also have multimodal functionality to assist the end-user in his/her communication often taking place in environment highly influencing the acoustic signal. In this context noise robust recognition of speech and methods to mitigate the influence of transmission errors are essential. The CNTK project is funding Ph.D. study within the SMC focusing on Robust and noise resistant human-computer speech enabled services using Distributed Speech Recognition The achievements so far involves DSR implementation, New front-end processing algorithm for noise reduction for speech recognition AND New algorithms for packet loss in DSR :