Research On Rate-adaptive Image Transmission System For Deep-space Exploration

High reliable and efficient image transmission from the remote space vehicle back to the Earth ground station is of primary importance for deep-space explorations.However,the extremely long distance and complex deep-space environment have introduced severe design as well as implementation challenges,including the large and time-varying signal attenuation,large propagation delay and intermittent disruptions and so on.Therefore,it is of great significance to select efficient and reliable techniques and design suitable transmission strategies for deep-space environment to ensure the high efficiency and reliability of the precious image data transmission in deep-space exploration.In this work,we propose a novel high-efficiency system to address all the design and implementation challenges for deep-space image transmissions.This novel system is designed to work over the Licklider Transmission Protocol(LTP)of the Delay-Tolerant Network(DTN)protocol stack.By incorporating the Compressed Sensing(CS)and the Spinal codes under the DTN protocol,this system can perfectly satisfy the constraints of asymmetrical node resources.However,the extremely long distance and the complex communication environment have brought great challenges to the design of deep-space image transmission systems.The traditional transmission control strategy,where the transmission rate can not adaptively match the time-varying deep-space channel,will lead to a waste of transmission resources.To realize dynamic matching with the time-varying deep-space channels,three coarse-grained rate-adaptive transmission strategies are further designed for the transmission system.The three strategies employ different CS image decompression mechanisms,namely the erasure-tolerant-reconstruction based and the error-tolerant-reconstruction based and the erasure-error-tolerant-reconstruction based decompression,to exploit the robustness of CS non-linear reconstruction to erasures and errors,respectively.In addition,the rates of CS image compression and Spinal channel coding for both of the three transmission strategies are jointly optimized to maximize the throughout.Taking the Gilbert-Elliott model as the basic deep-space channel model,the optimization model is established and solved.According to the results,the CS compression ratio and the Spinal code rate are dynamically adjusted.The transmission system with low frequency feedback and channel state prediction can realize the high efficiency transmission of deep-space image.In order to verify the effectiveness of the proposed deep-space image transmission system,we carried out extensive simulations on the Earth-Mars image transmission scenario on the existing semi-physical deep-space communication simulation platform.Results demonstrate that the actual throughput of the proposed system is very close to the theoretical throughput.And the designed system with our proposed rate-adaptive transmission strategies have significant performance improvement over the existing counterparts by achieving the near-ideal image transmission efficiency.