TH8361B - Methods and devices used to evaluate the symbols sent to the receiver in digital transmission. - Google Patents

Abstract

Digital Timeshare Radio Transmission System where the radio (14) receives a signal (R(T)) whose symbol frequency (1/TS) is lower than the band width of the system's channel. Correlation and sampling circuit (15) baseband receiver (y(T)),sampling the signal(21) eight times (y(k/8), with each symbol time (TS), proceeding to have a channel correlation (23), generating a channel approximation (HF ) and sampling the sample signal one at a time (y(k/8) to an observation signal (y(k/2)) with two values for each symbol time (TS). The channel balancer (17) sorts Some of the Waiterbi Algorithm Sequences which uses delta-metric values Two values for each state transition and produces the alternating zero symbol (formula) and the approximate symbol (31) is adjusted (34) with the help of the error signal (e(k/2)- y(l/2)- y( k/2)) and the filter sends the channel approximation (H(k/2)) to the channel balancer (17). A weighted multiplier((formula)(k/2)) is created. (35) With the help of error signals (e(k/2) and delta-metric values mentioned above The two values are weighted together with the delta-metric values. with the help of your weight The ((formula)(k/2)) improves metric calculations and simplifies the channel estimation filter (31) and the channel balancer (17). The interpolation of the zero symbol (formula) simplifies the generation of the weighted multiplier ((formula)k/2) and modulation (34) of the channel estimation filter (31). Fig. 4 for publishing

Claims (1)

1. In digital transmission across radio channels (13), Method of approximation in receivers (15, 16, 17) Symbols transmitted by the transmitted radio signal (R (T) where such symbol approximation is performed according to the Vitee algorithm. Norb (17) which has a number of states (B) as defined by the above method consists of the following algorithms: - receiving and filtering (14) transmitted signals (R (T) to generate baseband signal. (Y (T)); - sampling (21, y (k / 8)) baseband signal at the time point of the sampling. At least two points for each symbol; - Setting correlation (23) to determine the approximate impulse (HF) signal of a radio channel. (13), assisted channel estimation of the sample signal value (y (k / 8)); - determination of the symbol sample time point (TO) at a sample time point; - Selection of (22) at least two sample time points Along with each symbol Where one of them is the sample time point, the symbol (TO) and the selection of the observed sampling signal value (y (k / 2)) at these time points; - Determination of metric delta values (Jij formula (k-1/2), Jij formula (k)) according to the weiterb algorithm (17) for the transmitted symbol represented (k) by This judgment process This is carried out for each observed sample signal value (y (k-1/2), y (k)) and for each state change (i to j) of the Weiterbi algorithm. And - Providing at least an initial approximate notation (SD, (k)) according to the Weiterbi algorithm (17), where the methodology consists of a method of: - Origin of the sequence of symbols (SD, (k), (formula)) from approximate symbols (SD, (k) and zero-value symbols (formulas) by that order of symbols. There is at least one story symbol (formula) between two approximately two consecutive symbols; - Generation of approximate signal values (y (k-1/2), y (k)) at the selected sampling time point with the help of channel estimation (HF, H (k / 2) and symbol order. ; - Error signal generation (e (k-1/2), e (k) at each selected sampling time point of the indicator (k), with the help of the selected and observed sample signal value (y)). (k-1/2), y (k)) and approximate signal values (y (k-1/2), y (k)); - you weighted determination. ((Formula) k-1/2, (formula) k) that depends on the error-generating signal (17A) delta-matrix value. For observed state transitions of state transitions (i to j) of indicator symbols. (k) by multiplying the delta-metric value (Jij formula (k-1/2), Jij formula (k)) by its matching weighted multiplier ((formula) k-1/2, (formula ) k) and consolidation 2. The method specified in Clause 1, where directed by alignment (34), the channel estimate. (H (k / 2)) of the radio channel (13) with the help of an error signal (e (k / 2)) according to the selected modulation algorithm (LMS) 3. The specified method. In claim 1, where each Viterbi algorithm is connected to a paired channel estimate (H0, H1, H2, H3), directed by: - Select (82) channel estimate (H0), which is connected to the state (0) with the least metric value; And - Generation of approximate signal values (y (k-1/2) with the help of the selected channel estimation (H0) 4. Any method as specified in Clause 1,2 or 3? The one directed by the method includes the following: - root-root (51,52) the value of the error signal (e (k-1/2), e (k) at the selected sampling time point; And - Low-pass filtering (53,54) Squared error signal 5. Method according to claim 4, where directed by reverse (55, 56). Filter low-pass squared. ((Formula) k-1/2, (formula) k)) to create a weighted multiplier. ((Formula) k-1/2, (Formula) k) 6. Method according to Clause 4, which is directed by: - Root and summation (H0 (k) 2 + H2 (k) ) 2; H2 (k) 2 + H3 (k) 2) coefficient in the channel approximation; And - Divide (55,56) sum by matching exponent and filtered error signal. ((Formula) k-1/2, (formula) k) 7. Placement of devices in the receiver of the digital radio transmission system for approximation of symbols from radio signals. (R (T) transmitted to radio channel (13) by such arrangement includes: - radio receiver (14) with a filter that receives radio signal (R (T) and generates baseband signal ( y (t)); - one sampling unit (21) which samples the baseband signal and conveys the signal value (y (k / 8)) at at least two sample time points (k / 8) with each symbol; - Correlation circuit (23), which provides channel estimation (HF) for radio channels (13) with the help of sample signal values (y (k / 8)); - Synchronization circuit (24) for Determination of the sample time point symbol (TO) in one of the sample time points; - The second sampling unit (22), which is connected to the first sampling unit (21) and is controlled by a synchronous (24) and least conveyed cycle. Two observed sample signal values (y (k-1/2), y (k)) with each symbol; And - channel equilibrium (17), which for the indicator symbol (k) follows a witerbi algorithm that generates delta-matrix values (Jij formula (k-1/2), Jij formula. (k)) for each observed signal value (y (k-1/2), y (k)) of the state transition (I to j) and at least the initial approximate symbol (SD ( k)), where the arrangement also includes: - Cycle (36) which generates symbolic sequences (SD, (k), (formula)) of approximate symbols (SD, (k)) and zero-value story symbols ( Formula) which is in between; - channel approximation filter (31), which generates approximate signal values (y (k-1/2), y (k)) with the help of the generated symbol sequence; - differential generator (33), which generates an error signal (e (k-1/2), e (k) for the indicator symbol (k), with the help of the observed signal value (y (k-)). 1/2), y (k)) and at approx. (Y (k-1/2), y (k)); - squared and mean (35) generating circuit, which produces a weighted multiplier. ((Formula) k-1/2, (formula) k) that depends on the error signal (e (k-1/2), e (k); And - A matrix compute circuit (17A), which produces a combined delta-matrix. By delta-metric values ((Formula) k-1/2, (formula) k) of the transition (I to j) by the corresponding weighted multiplier ((formula) k-1/2, (formula) k) and the sum of the product. 8. Arrangement according to claim 7, where the filter estimates the channel (31), the coefficient circuit (42), whose value is adjusted to approximate the channel of the radio channel (13 ) By modulation circuit (34) with the help of an error signal (e (k-1/2) according to the selected modulation algorithm (LMS) 9. Handling according to the clause in the claim Either clause 7 or 8 where such arrangement includes: - Root extractor (51,52) which takes the root of the error signal (e (k-1/2), e (k) where Selected sampling time point; And - a low-pass filter (53, 54), which filters a square error signal of 1 0. The arrangement is stated in Clause 9, where such arrangement includes a reverse converter (55, 56). ) Which receives an error signal that filters low pass And square ((Formula) k-1/2, (formula) k) and transport the weighted multiplier. ((Formula) k-1/2, (formula) k) 1. 1. Arrangement as set out in Clause 9, where such arrangement includes: - Converters (55,56) that receive a square low-pass filtered error signal. ((Formula) k-1/2, (formula) k) and convey the conversion value. Back matching; And - circuit (59, 60), which is connected to the converter and where the values (H0 (k) 2 + H2 (k) 2; H2 (k) 2 + H3 (k) 2) of the coefficient circuit (42) raise. Square and sum, multiplied by that inverse value.