Receiver based Rate Adaptation Approach through Blind SNR Estimation of Satellite Channel Sorum Kotia, Dr. K. S. Dasgupta and Dr. M. D. Desai Abstract— A joint source and channel coding approach for image transmission over a satellite based broadcast channel is proposed. A resolution scalable source is matched to a hierarchical constellation in order to achieve an adaptive decoding capability. An estimate of signal to noise ratio (SNR) is derived blindly at the receiver and is used to facilitate reception at lower information rate. The rate is reduced by switching to a lower level constellation from hierarchical QAM. This ensures graceful degradation in performance and guarantees service with lower resolution for channel under-going high AWGN noise. The simulation results suggest that, for very noisy channels, the end to end distortion so obtained is much lesser compared to decoding at full resolution. Keywords- Hierarchical (HR) mapping, Joint source and channel coding(JSCC), resolution scalability, , moment based signal to noise ratio estimation., I. INTRODUCTION Reliable transmission of digital multimedia data over a noisy channel is a challenging task. It is especially difficult to guarantee a quality of service for band-limited satellite channels, which experience a large variation in channel noise. Primarily, it calls for a careful tradeoff between the throughput and quality of reception under the constraints of limited bandwidth and limited power. Secondly, it needs a mechanism to obtain channel state information (CSI) which could facilitate either adaptive transmission or adaptive reception under a varying link quality. A broadcast channel does not allow retransmission requests which could help mitigate data loss in poor channel conditions. Neither does it provide a dedicated link between transmitter and receiver which could help obtain information regarding link quality and hence provide adaptive throughput adjustment. Therefore, we suggest an approach which engineers resources at the transmitter to match a variety of channel conditions; derives an estimate of SNR at the receiver, uses these estimates to adapt the rate at the receiver and thus ensures service quality for satellite channels under broadcast scenarios. Manuscript received on January 6, 2010. Sorum Kotia is perusing PhD at EC Department of Nirma University Ahmedabad, India. She is presently working as a Faculty in the Electrical Engineering Department of The Maharaja Sayajirao University of Baroda, Baroda, India. (e-mail: smkotia-eed@msubaroda.ac.in). Dr. K. S. Dasgupta is the Deputy Director of SITAA, Space Application Center, Ahmedabad, India (e-mail: ksd@sac.isro.gov.in). Dr. M. D. Desai was a Faculty with EC Department, Nirma University, Ahmedabad, India. Presently he is working as a Professor with Kalol Institute of Technology, Kalol, India. (e-mail: mdd_p_ic.it@nirmauni.ac.in). A combined source and channel coding scheme is investigated here for satellite based digital image broadcast, where the rate is adapted at the receiver using blind estimates of channel SNR. It is a static rate adaptation scheme for a satellite channel which is undergoing a large change in AWGN noise. The source coding technique and the error protection mechanism are loosely coupled to provide a suboptimal rate allocation, whereas adaptive reception assisted by blind SNR estimation at the receiver ensures guaranteed QoS and yet calls for a minimalist modification in the receiver. The multimedia information, in this case an image, is described in a resolution scalable form. The hierarchy of source resolution is matched to a hierarchical (HR) modulation constellation. Such an integration of a multiresolution (MR) source coder and a HR constellation mapper facilitates a channel which virtually comprises of two embedded channels, each having unequal error protection (UEP). The information stream with stringent error sensitivities i.e. the coarser resolution is directed to the subchannel in a better condition. The less important stream i.e. finer resolution information is directed to the subchannel which is in a less perfect condition. UEP channels enable the receiver to decode the received signal with a resolution or quality which is in tune with the channel SNR. In good channel condition, the decoding is done at full resolution; whereas in poor channel condition the image is received with a reduced resolution. Similar MR-JSCC schemes for transmission over time-varying channels are discussed in [1]-[4]. However, in most of these cases, it is assumed that the CSI is available apriori and not much is said about how to obtain it. This paper suggests information rate adaptation at the receiver by employing estimates of SNR which are derived blindly at the receiver. During large fluctuations in the channel noise or under deep fade conditions, it becomes essential to reduce the rate so that a reasonable quality of reception is achieved. We reduce the rate by decreasing the number of modulation levels so as to create a larger Euclidean distance between the signaling points. This is realized by intentionally neglecting the least significant bits of a symbol, i.e. by dropping off the QAM constellation bits which are overriding a basic QPSK constellation. Thus switching to a smaller constellation ensures a coarse level of granularity under adverse channel conditions. The intention is to moderately sacrifice the quality of reception and in turn attain graceful degradation with deteriorating link conditions. Proceedings of the International MultiConference of Engineers and Computer Scientists 2010 Vol II, IMECS 2010, March 17 - 19, 2010, Hong Kong ISBN: 978-988-18210-4-1 ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online) IMECS 2010