ISSN 
2003 Vol. 32, No. 5
2003, 32(5): 473-476,529.
Abstract:
Self-encoded spread spectrum(SESS) system employs random spreading sequences extracted from the information date to assure dynamically random varying. Performance analysis of self-encoded spread spectrum in additive white Gaussian noise channels is presented. An equivalent spread spectrum gain is used to describe the characteristic of self-encoded direct sequence spread spectrum system. Improved approaches generating self-encoded spread spectrum sequences are introduced. Theoretical analysis and computer simulation show that the spread spectrum gain of self-encoded direct sequence spread spectum system is almost the same to conventional spread spectrum system, but private of self-encoded direct sequence spread spectum system increases with variability of spread spectrum codes.
Self-encoded spread spectrum(SESS) system employs random spreading sequences extracted from the information date to assure dynamically random varying. Performance analysis of self-encoded spread spectrum in additive white Gaussian noise channels is presented. An equivalent spread spectrum gain is used to describe the characteristic of self-encoded direct sequence spread spectrum system. Improved approaches generating self-encoded spread spectrum sequences are introduced. Theoretical analysis and computer simulation show that the spread spectrum gain of self-encoded direct sequence spread spectum system is almost the same to conventional spread spectrum system, but private of self-encoded direct sequence spread spectum system increases with variability of spread spectrum codes.
2003, 32(5): 477-480.
Abstract:
This paper introduces time-hopping pulse-position-modulation, binary time-hopping pulse-amplitude-modulation and direct sequence ultra wideband first. Then we analyze the bit error rate performances of the three Ultra Wideband radio models in additive white Gaussian noise and in the presence of single-tone jamming and the results obtained are compared by theoretic analysis and computer simulation. The simulation results show that the Ultra Wideband radio is of good performance in low SNR indeed, and the bit error rate performances of binary time-hopping pulse-amplitude-modulation and direct sequence ultra wideband schemes are better than that of time-hopping pulse-position-modulation scheme.
This paper introduces time-hopping pulse-position-modulation, binary time-hopping pulse-amplitude-modulation and direct sequence ultra wideband first. Then we analyze the bit error rate performances of the three Ultra Wideband radio models in additive white Gaussian noise and in the presence of single-tone jamming and the results obtained are compared by theoretic analysis and computer simulation. The simulation results show that the Ultra Wideband radio is of good performance in low SNR indeed, and the bit error rate performances of binary time-hopping pulse-amplitude-modulation and direct sequence ultra wideband schemes are better than that of time-hopping pulse-position-modulation scheme.
2003, 32(5): 481-484.
Abstract:
Due to the dependence of UWB signal receiver on A/D apparatus of high sampling rate at present, a structure of a low sampling rate receiver for UWB signals is proposed. The principle and the performance of the receiver in AWGN channel have also been analyzed. In the end we take four parallel filters as example and get the approximate results between the computer simulation and the theory analysis.
Due to the dependence of UWB signal receiver on A/D apparatus of high sampling rate at present, a structure of a low sampling rate receiver for UWB signals is proposed. The principle and the performance of the receiver in AWGN channel have also been analyzed. In the end we take four parallel filters as example and get the approximate results between the computer simulation and the theory analysis.
2003, 32(5): 485-489.
Abstract:
Union upper bound of pairwise-error-probability (PEP) for space-time/frequency block codes in multiple-antenna based orthogonal frequency-division multiplexing (OFDM) systems is derived. And new design considerations were revealed. Then performance comparison between space-frequency block (SFBC) codes, space-time block codes (STBC) and space-time-frequency block codes (STFBC)is performed at different throughput by link-level simulations. The results show that SFBC is superior to STBC in fast fading channel. And power-delay profile of fading channel should be considered for designs of space-time-frequency codes as well as channel orders. SFBC with 4 transmit antenna is robust to both vehicular A and B channel environment of ITU-R M.1225.
Union upper bound of pairwise-error-probability (PEP) for space-time/frequency block codes in multiple-antenna based orthogonal frequency-division multiplexing (OFDM) systems is derived. And new design considerations were revealed. Then performance comparison between space-frequency block (SFBC) codes, space-time block codes (STBC) and space-time-frequency block codes (STFBC)is performed at different throughput by link-level simulations. The results show that SFBC is superior to STBC in fast fading channel. And power-delay profile of fading channel should be considered for designs of space-time-frequency codes as well as channel orders. SFBC with 4 transmit antenna is robust to both vehicular A and B channel environment of ITU-R M.1225.
2003, 32(5): 490-494.
Abstract:
The current analysis focuses on the influence of demodulation performance in multipath fading propagation environments, which results from the pilot power variation of pilot symbol assisted modulation M-PSK in orthogonal frequency division multiplexing systems. The pilot-to-data power ratio is optimized analytically, and the loss due to imperfect channel estimation is calculated. Theoretical and simulation results show that the optimum pilot-to-data power ratio is affected by the signal-to-noise ratio, Doppler frequency and interpolation size etc. Once the optimization parameter of pilot-to-data power ratio is used, system transmission power is minimum while still meeting the fixed bit error rate target.
The current analysis focuses on the influence of demodulation performance in multipath fading propagation environments, which results from the pilot power variation of pilot symbol assisted modulation M-PSK in orthogonal frequency division multiplexing systems. The pilot-to-data power ratio is optimized analytically, and the loss due to imperfect channel estimation is calculated. Theoretical and simulation results show that the optimum pilot-to-data power ratio is affected by the signal-to-noise ratio, Doppler frequency and interpolation size etc. Once the optimization parameter of pilot-to-data power ratio is used, system transmission power is minimum while still meeting the fixed bit error rate target.
2003, 32(5): 495-498.
Abstract:
In this paper we proposed a new OFDM frequency synchronization method by using PN sequence. In multi-path channel, the frequency estimation precision of this method is better than conventional method for it using multi-path signals. The simulation indicates the proposed method can satisfy the demand of OFDM system at moderate moving rate in Rayleigh channel. One of the advantages of this method is that its frequency estimation range can be dozens of subcarrier spacing when the parameters are properly designed. And the frequency estimation range and precision can be adjusted.
In this paper we proposed a new OFDM frequency synchronization method by using PN sequence. In multi-path channel, the frequency estimation precision of this method is better than conventional method for it using multi-path signals. The simulation indicates the proposed method can satisfy the demand of OFDM system at moderate moving rate in Rayleigh channel. One of the advantages of this method is that its frequency estimation range can be dozens of subcarrier spacing when the parameters are properly designed. And the frequency estimation range and precision can be adjusted.
2003, 32(5): 499-502.
Abstract:
Peak-to-average power ratio of orthogonal frequency division multiplexing signals is analyzed when the sampling rate is different. The samples according to Nyquist sample rate in one period of OFDM symbol are independent identically distributed random variables. When oversampling factor is 2, the covariance matrix composed of the coefficients of correlation between those samples can be calculated by numerical method. As a result, all samples in one symbol are approximately independent when oversampling factor is 2. So the probability distribution function of peak-to-average power ratio can be achieved. The results are proved by simulation.
Peak-to-average power ratio of orthogonal frequency division multiplexing signals is analyzed when the sampling rate is different. The samples according to Nyquist sample rate in one period of OFDM symbol are independent identically distributed random variables. When oversampling factor is 2, the covariance matrix composed of the coefficients of correlation between those samples can be calculated by numerical method. As a result, all samples in one symbol are approximately independent when oversampling factor is 2. So the probability distribution function of peak-to-average power ratio can be achieved. The results are proved by simulation.
2003, 32(5): 503-507.
Abstract:
For orthogonal frequency-division multiplexing (OFDM) communication systems, a loss of subcarrier orthogonality due to time-variant multipath channel leads to intercarrier interference (ICI).This paper studies an efficient ICI cancellation method termed segment equalization schemes.The signal-to-interference power rate can be increase 6 dB and 12 dB when the segment size is two or four, and the system has the same spectrum efficiency with standard equalization. Simulations and performance analysis show that OFDM system using the proposed segment equalization can effectively reduce the ICI cause by frequency offset and large doppler frequency shifts in mobile radio channels.
For orthogonal frequency-division multiplexing (OFDM) communication systems, a loss of subcarrier orthogonality due to time-variant multipath channel leads to intercarrier interference (ICI).This paper studies an efficient ICI cancellation method termed segment equalization schemes.The signal-to-interference power rate can be increase 6 dB and 12 dB when the segment size is two or four, and the system has the same spectrum efficiency with standard equalization. Simulations and performance analysis show that OFDM system using the proposed segment equalization can effectively reduce the ICI cause by frequency offset and large doppler frequency shifts in mobile radio channels.
2003, 32(5): 508-511.
Abstract:
A method of resource scheduling in OFDMA communication system is proposed in this paper, with this scheme, according user data rates and channel parameter, transmitting slots and subcarriers of OFDM systems are allocated dynamically. In this paper, the scheduling method and algorithm are presented and simulation results are given. The throughput of the scheduling system could be improved almost three times than moral system.
A method of resource scheduling in OFDMA communication system is proposed in this paper, with this scheme, according user data rates and channel parameter, transmitting slots and subcarriers of OFDM systems are allocated dynamically. In this paper, the scheduling method and algorithm are presented and simulation results are given. The throughput of the scheduling system could be improved almost three times than moral system.
2003, 32(5): 512-515.
Abstract:
In this paper, the comparison of bit error ratio(BER) performance of symbol-by-symbol frequency domain differential detection(SS-FDDD) and multiple-symbol frequency domain differential detection(MS-FDDD) in orthogonal frequency division multiplexing system is obtained by simulation, the results indicate that when the whole bandwidth of the symbols used to MS-FDDD is much less than the coherent bandwidth of the channel, the BER performance of MS-FDDD is better than that of SS-FDDD; when the whole bandwidth of the symbols used to MS-FDDD is closed to the coherent bandwidth of the channel, the BER performance of SS-FDDD is better than that of Multiple-symbol frequency domain differential detection.
In this paper, the comparison of bit error ratio(BER) performance of symbol-by-symbol frequency domain differential detection(SS-FDDD) and multiple-symbol frequency domain differential detection(MS-FDDD) in orthogonal frequency division multiplexing system is obtained by simulation, the results indicate that when the whole bandwidth of the symbols used to MS-FDDD is much less than the coherent bandwidth of the channel, the BER performance of MS-FDDD is better than that of SS-FDDD; when the whole bandwidth of the symbols used to MS-FDDD is closed to the coherent bandwidth of the channel, the BER performance of SS-FDDD is better than that of Multiple-symbol frequency domain differential detection.
2003, 32(5): 516-520.
Abstract:
The model of pilot symbols assisted modulation adaptive modulation orthogonal frequehcy division multiplexing system is given in Rayleigh fading。Based on this model, spectral efficiency definition and calculation are analyzed. According to bit error rate of PSAM OFDM, the numerical result is presented under specific channel model. In order to be reference for practical system design, the result shows influences of receiver non-ideal channel estimation and system overhead must be considered.
The model of pilot symbols assisted modulation adaptive modulation orthogonal frequehcy division multiplexing system is given in Rayleigh fading。Based on this model, spectral efficiency definition and calculation are analyzed. According to bit error rate of PSAM OFDM, the numerical result is presented under specific channel model. In order to be reference for practical system design, the result shows influences of receiver non-ideal channel estimation and system overhead must be considered.
2003, 32(5): 521-524.
Abstract:
This paper investigated the impact of the timing error on the receiver BER performance in the pre-equalization OFDM systems through simulations. The results showed that the timing error would bring on the phase rotation of demodulation data, and degrade the system performance critically. So it must be revised by assistant methods. The timing error and the number of sub-carriers together decide the degradation. Simulations showed that the performance would fall down 1 dB every' 1/16 normalized timing error.
This paper investigated the impact of the timing error on the receiver BER performance in the pre-equalization OFDM systems through simulations. The results showed that the timing error would bring on the phase rotation of demodulation data, and degrade the system performance critically. So it must be revised by assistant methods. The timing error and the number of sub-carriers together decide the degradation. Simulations showed that the performance would fall down 1 dB every' 1/16 normalized timing error.
2003, 32(5): 525-529.
Abstract:
This paper deduces and sums up the principle and optimizing object of constructing the frequency transition function according to the characteristic of differential frequency hopping system. Also, a kind of effective constructing technique is put forward, and the merits in improving the performance of BER and randomicity of frequencies sequences are proved in theory. The technique is based on multi-dimension space transform, the realizing of this technique is convenience, and it is easy to control the secret key of frequency hopping.
This paper deduces and sums up the principle and optimizing object of constructing the frequency transition function according to the characteristic of differential frequency hopping system. Also, a kind of effective constructing technique is put forward, and the merits in improving the performance of BER and randomicity of frequencies sequences are proved in theory. The technique is based on multi-dimension space transform, the realizing of this technique is convenience, and it is easy to control the secret key of frequency hopping.
2003, 32(5): 530-534.
Abstract:
According to the characteristic of differential frequency hopping, this paper presents a nonbinary convolutional code equivalent model of G function and the design of optium receiver for differential frequency hopping signal. The paper also studies two maxium-likelihood detect algorithms (Viterbi hard-decision detection algorithm and Viterbi soft-decision algorithm) and computer simulations indicate that the maxium-likelihood method can greatly improves performance of differential frequency hopping system for AWGN channel.Finally,the realization of Viterbi hard-decision detection algorithm is described.
According to the characteristic of differential frequency hopping, this paper presents a nonbinary convolutional code equivalent model of G function and the design of optium receiver for differential frequency hopping signal. The paper also studies two maxium-likelihood detect algorithms (Viterbi hard-decision detection algorithm and Viterbi soft-decision algorithm) and computer simulations indicate that the maxium-likelihood method can greatly improves performance of differential frequency hopping system for AWGN channel.Finally,the realization of Viterbi hard-decision detection algorithm is described.
2003, 32(5): 535-540.
Abstract:
A new FFT-based detection algorithm for the existence of synchronization signal in fast frequency-hopping (FFH) system is presented in this paper. Four methods (clipping, envelop-subtraction of two connective hops, iterative double-clipping, clipping-then-subtraction-then-clipping) of evaluating normalized signal to noise ratio(Eb/N0) in frequency domain with jamming are developed and compared. Analysis formulas of these methods are established as well. Our simulations show that numerical results of normalized signal to noise ratio are close to its theoretical analysis for both AWGN and multitone jamming environment, where discrepancy are within the range of ±3 dB. The overall performance of the latter two methods are 5 dB better than the first two methods for partial-band noise jamming. Iterative double-clipping is our recommended approach because of it's low complexity of implementation.
A new FFT-based detection algorithm for the existence of synchronization signal in fast frequency-hopping (FFH) system is presented in this paper. Four methods (clipping, envelop-subtraction of two connective hops, iterative double-clipping, clipping-then-subtraction-then-clipping) of evaluating normalized signal to noise ratio(Eb/N0) in frequency domain with jamming are developed and compared. Analysis formulas of these methods are established as well. Our simulations show that numerical results of normalized signal to noise ratio are close to its theoretical analysis for both AWGN and multitone jamming environment, where discrepancy are within the range of ±3 dB. The overall performance of the latter two methods are 5 dB better than the first two methods for partial-band noise jamming. Iterative double-clipping is our recommended approach because of it's low complexity of implementation.
2003, 32(5): 541-545.
Abstract:
Base on the discussion of three kinds of basic technique of frequency synthesis such as direct frequency synthesis, phase-locked loop and direct digital synthesis, and several implementation technique of frequency synthesizer, a design of frequency synthesizer combining direct digital synthesis with a chain of frequency multiplier is given. This type of frequency synthesizer is especially suitable for ultra-high-speed frequency-hop system, and find a solution that the output frequency spectrum purity of frequency synthesizer is guaranteed on the condition of ultra high speed frequency hopping. Some experimental results are also presented in the end.
Base on the discussion of three kinds of basic technique of frequency synthesis such as direct frequency synthesis, phase-locked loop and direct digital synthesis, and several implementation technique of frequency synthesizer, a design of frequency synthesizer combining direct digital synthesis with a chain of frequency multiplier is given. This type of frequency synthesizer is especially suitable for ultra-high-speed frequency-hop system, and find a solution that the output frequency spectrum purity of frequency synthesizer is guaranteed on the condition of ultra high speed frequency hopping. Some experimental results are also presented in the end.
2003, 32(5): 546-550.
Abstract:
This paper describes the design and implement of train's internal alarm and communication system based on digital voice compressing techniques by software and hardware designing methods, using the broadcast line in the civil old type train running currently as communication channel. The system has high stability and variability, can process warning signals and notice workers in the train on time, can construct an internal voice communication system for every carriage between each other at any time.
This paper describes the design and implement of train's internal alarm and communication system based on digital voice compressing techniques by software and hardware designing methods, using the broadcast line in the civil old type train running currently as communication channel. The system has high stability and variability, can process warning signals and notice workers in the train on time, can construct an internal voice communication system for every carriage between each other at any time.
2003, 32(5): 551-554.
Abstract:
This paper presents the results of HF channel time delay measurements on 580 km, 940 km and 1 380 km paths using oblique ionospheric sounder equipment, and gives a curve fitting data processing methods to deal with measurements data of single mode. The variety characteristics of single mode FRTD is discussed on the basis of analysis of calculating results, it indicates that the FRTD values decrease with frequency interval getting small and paths distance increasing, the maximum FRTD value in a day appears at noon.
This paper presents the results of HF channel time delay measurements on 580 km, 940 km and 1 380 km paths using oblique ionospheric sounder equipment, and gives a curve fitting data processing methods to deal with measurements data of single mode. The variety characteristics of single mode FRTD is discussed on the basis of analysis of calculating results, it indicates that the FRTD values decrease with frequency interval getting small and paths distance increasing, the maximum FRTD value in a day appears at noon.
2003, 32(5): 555-559.
Abstract:
The key of developing high frequency wideband communication system is how to model the HF wideband channel. Some HF wideband models are discussed and compared in this paper, including Vogler model, parallel subband model, two improved Watterson models and Pseudo-Deterministic model. As a deterministic model, Volger model basically describes the characteristics of HF wideband channel, but it also needs to be improved. The improvement of Vogler model is given in this paper.
The key of developing high frequency wideband communication system is how to model the HF wideband channel. Some HF wideband models are discussed and compared in this paper, including Vogler model, parallel subband model, two improved Watterson models and Pseudo-Deterministic model. As a deterministic model, Volger model basically describes the characteristics of HF wideband channel, but it also needs to be improved. The improvement of Vogler model is given in this paper.
2003, 32(5): 560-563.
Abstract:
According to the software radio model, we analyse the characteristic of multiple channels in Software Radios, that is, each channel can work independently, at the same time all channels can cooperate together. Then some key networking technologies, such as network control structure, routing, wireless medium access control and resource management etc, are discussed. We propose an idea and scheme of networking based on software radios. The software radios network using the scheme will make the networking process flexible, have self-organized and dynamic load-balancing functions.
According to the software radio model, we analyse the characteristic of multiple channels in Software Radios, that is, each channel can work independently, at the same time all channels can cooperate together. Then some key networking technologies, such as network control structure, routing, wireless medium access control and resource management etc, are discussed. We propose an idea and scheme of networking based on software radios. The software radios network using the scheme will make the networking process flexible, have self-organized and dynamic load-balancing functions.
2003, 32(5): 564-568.
Abstract:
Based on framework of Wireless Multi-Protocol Label Switching protocol and characteristic of Mobile Ad hoc Network, we propose a kind of QoS routing protocol-AMPLS(Ad-hoc MPLS) for MANET using extended CR-LDP signaling protocol to establish Label Switched Path. Then, we build a MANET networking simulation model with AMPLS protocol, analyse the performance of AMPLS protocol, validate its effectivity regarding differentiated services and traffic engineering networking features, and provides one kind of method for mobile ad hoc networking with QoS guarantee to some degree.
Based on framework of Wireless Multi-Protocol Label Switching protocol and characteristic of Mobile Ad hoc Network, we propose a kind of QoS routing protocol-AMPLS(Ad-hoc MPLS) for MANET using extended CR-LDP signaling protocol to establish Label Switched Path. Then, we build a MANET networking simulation model with AMPLS protocol, analyse the performance of AMPLS protocol, validate its effectivity regarding differentiated services and traffic engineering networking features, and provides one kind of method for mobile ad hoc networking with QoS guarantee to some degree.
2003, 32(5): 569-573.
Abstract:
Because of the flexibility and convenience it can provide us, the mobile ad-hoc network becomes a very hot research topic presently. There are many related articles presented on various journals in this area, especially concentrated on the ad-hoc routing protocols. But, the solutions provided still have more or less defects. Firstly this article introduces the special characters of ad-hoc network and the existing routing protocols. Then analysis and simulation of the link break of ad-hoc on-demand distance vector in ad hoc networks are presented and a new mechanism to improve the performance of link repair is provided. The simulation results indicate that the improved method results in less delay and routing overhead while reconstructing route compared to the traditional method to some extent.
Because of the flexibility and convenience it can provide us, the mobile ad-hoc network becomes a very hot research topic presently. There are many related articles presented on various journals in this area, especially concentrated on the ad-hoc routing protocols. But, the solutions provided still have more or less defects. Firstly this article introduces the special characters of ad-hoc network and the existing routing protocols. Then analysis and simulation of the link break of ad-hoc on-demand distance vector in ad hoc networks are presented and a new mechanism to improve the performance of link repair is provided. The simulation results indicate that the improved method results in less delay and routing overhead while reconstructing route compared to the traditional method to some extent.
2003, 32(5): 574-577.
Abstract:
This paper analyzes the parallel mechanism of logarithmic maximum a posteriori algorithm and presents a modified iterative procedure by the use of the possibility of receiving the full N-symbols frame and the symmtry being inherent in the forward-backward iteration. The procedure achieves a double decoding speed faster than that of the conventional one without the increase of RAM cost because the number of iterations is limited to a half of the code-length N. Also according to the modified algorithm this paper proposes a hardware scheme characterized by a two-busses architecture suitable for implementing with FPGA.
This paper analyzes the parallel mechanism of logarithmic maximum a posteriori algorithm and presents a modified iterative procedure by the use of the possibility of receiving the full N-symbols frame and the symmtry being inherent in the forward-backward iteration. The procedure achieves a double decoding speed faster than that of the conventional one without the increase of RAM cost because the number of iterations is limited to a half of the code-length N. Also according to the modified algorithm this paper proposes a hardware scheme characterized by a two-busses architecture suitable for implementing with FPGA.
2003, 32(5): 578-582.
Abstract:
In this thesis, the performance of complementary code keying is compared between optimal decoding and majority logic decoding algorithm, and analyzed in fading channels. A structure concatenating low-density parity-check code and complementary code keying and the algorithm of soft-decision output for complementary code keying are presented, the simulation result show more better performance than complementary code keying in rayleigh fading channels.
In this thesis, the performance of complementary code keying is compared between optimal decoding and majority logic decoding algorithm, and analyzed in fading channels. A structure concatenating low-density parity-check code and complementary code keying and the algorithm of soft-decision output for complementary code keying are presented, the simulation result show more better performance than complementary code keying in rayleigh fading channels.
2003, 32(5): 583-587.
Abstract:
Multiple access interference is one kind of main interference in CDMA system. It can be restrained effectively by interference cancellation, and the realization of this technology needs the estimation of fading channels. This paper introduces the technology of channel estimation and multiple access interference, and states how user amount and deviation of channel estimation can affect the performance of multiple access interference cancellation technology basing on the data analysis result.
Multiple access interference is one kind of main interference in CDMA system. It can be restrained effectively by interference cancellation, and the realization of this technology needs the estimation of fading channels. This paper introduces the technology of channel estimation and multiple access interference, and states how user amount and deviation of channel estimation can affect the performance of multiple access interference cancellation technology basing on the data analysis result.
2003, 32(5): 588-591.
Abstract:
This paper studies multistage partial parallel interference cancellation for CDMA systems over multipath fading channels. By analysis of the variance of residual interference, we derive an expression determining the optimal interference cancellation weight in the event of channel estimation error and propose a channel-estimator-aided partial parallel interference cancellation applicable to multi-rate CDMA systems. The simulation results verify the reasonable assumption that the proposed partial parallel interference cancellation has performance better than partial parallel interference cancellation with freewill constant weight. It is indicated that, at a BER of 0.01, the performance loss of the proposed PPIC stage 2 is only 1.5 dB, compared to partial parallel interference cancellation with channel-known interference of one stage.
This paper studies multistage partial parallel interference cancellation for CDMA systems over multipath fading channels. By analysis of the variance of residual interference, we derive an expression determining the optimal interference cancellation weight in the event of channel estimation error and propose a channel-estimator-aided partial parallel interference cancellation applicable to multi-rate CDMA systems. The simulation results verify the reasonable assumption that the proposed partial parallel interference cancellation has performance better than partial parallel interference cancellation with freewill constant weight. It is indicated that, at a BER of 0.01, the performance loss of the proposed PPIC stage 2 is only 1.5 dB, compared to partial parallel interference cancellation with channel-known interference of one stage.
