QoE_U Tility Based MIMO-OFDM System Rescource Allocate Technology Research And Realization

Recently, the resource allocation algorithm has always been a research hotspot which provides QoS (Quality of Service) support. However, QoS could not reflect user experience to service due to its ignorance about user's subjective factor. Therefore, ITU defines QoE (Quality of Experice), releases its statistical parameters and establishes related evaluatation system, which is the first time to introduce the user experience as a standard. Accordingly, researches on resource allocation based on QoE starts warming up.Aimed at the maximum mean user QoE, this paper starts a reaserch on resource allocation in MIMO-OFDM system after establishing utility functions according to QoE measuring indexs under different environment. The main work can be explained as follows:1. According to different environmental requirements, this paper has proposed three different utility functions based on statistis: Capacity_QoE utility function aimed at user throughput, RT_QoE and NRT_QoE utility functions oriented to RT(Real Time) and NRT(Non-Real Time) services. Analysis of utility function characteristic simulation indicates that the utility functions are in accordance with user's subjective experience to services.2. Adapted to circumstances subjected to user throughput, this paper proposed a subcarrier and power allocation algorithm aimed at maximum mean user QoE based on Capacity_QoE utility function. The algorithm accomplishes power and subcarrier allocation by two steps: first allocate subcarriers according to the user channel quality and QoE indexs; and then get approximate optimal solution to power allocation through derivative iteration. Simulation results indicate that this algorithm could achieve higher QoE indexs than classic algorithm and meantime maintaining high system throughput.3. Taking user's different experiences to RT and NRT services, this paper proposed a cross layer resource allocation algorithm based on RT_QoE and NRT_QoE utility functions. The algorithm first integrates subcarrier and timeslot resources into time-frequency blocks and classifies users according to their delta QoE caculated by different RT and NRT utility functions. Time-frequency blocks are allocated according priority level and pseudo water flooding power allocation is implemented. The simulation result indicates that the algorithm effectively increases mean user QoE and satisfies traditional QoS indexs.4. In order to realize the cross layer resource allocation algorithm on hardware platform, this paper designes a platform with a FPGA-DSP chipset processing core. According to the characteristic of TMS320C6474 chip, algorithm realization is accomplished efficiently and quickly through a series of program optimization.