Baseband decoding method based on differential edge detection combined with adaptive decision threshold

Abstract

The application discloses a baseband decoding method based on differential edge detection combined with an adaptive decision threshold, and the implementation steps are as follows: differentiating a digital signal after ADC sampling and quantization and median filtering processing; adopting a windowing statistical method, calculating an adaptive decision threshold by comparing the size of the signal differential level at each sampling moment and the windowing statistical threshold, and performing edge detection on the signal after the differential processing by using the adaptive decision threshold; and reconstructing the signal waveform according to the edge detection result after the continuity detection, and recovering the original data information. The application avoids the influence of the variable signal-to-noise ratio and the variable direct current bias on the decoding accuracy, improves the reliability of the baseband decoding, and simultaneously realizes the efficient baseband decoding of the large dynamic variable bias signal, and has the advantages of many application scenarios in practical engineering.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of communication technology, and further relates to a baseband decoding method based on differential edge detection combined with adaptive decision threshold in the field of wireless communication technology. The present application can be used for efficient baseband decoding of signals with uncertain signal-to-noise ratio, amplitude changing in a large range, strong dynamicity, and uncertain DC bias component. BACKGROUND

[0002] In the field of wireless communication technology, the original baseband signal is encoded by the sending end after source encoding and channel encoding, and then converted into a high-frequency signal suitable for transmission in the channel through modulation technology and sent. The receiving end recovers the original baseband signal from the received high-frequency signal through demodulation. The demodulated original baseband signal is an analog signal, in order to be compatible with the subsequent digital signal processing system, it needs to be further converted into a digital signal through an analog-to-digital converter ADC (Analog-to-Digital Converter), and then recovered through baseband decoding to restore the original data information, channel decoding to detect and correct errors in transmission, and source decoding to recover the original compressed data, and finally the data is recovered to the original information available to the user. In the wireless communication scene, the large dynamic variable bias signal refers to a signal with uncertain signal-to-noise ratio, amplitude changing in a large range, strong dynamicity, and uncertain DC bias component due to factors such as multipath effect, shadow fading, interference, etc. in a complex environment. The difficulty of large dynamic variable bias signal baseband decoding lies in that the decoding method needs to be able to adapt to different noise levels and signal amplitudes, while being able to eliminate the influence of uncertain DC bias component. This requires the decoding algorithm not only to have high sensitivity and accuracy, but also to have fast response speed, so as to accurately identify and extract the effective information in the signal in the case of rapid change of signal quality.

[0003] The Space Information Co., Ltd. discloses a correlator composed of a radio frequency identification (RFID) baseband decoding system in its applied patent document "A signal processing device and reader" (patent application number CN 201810889628.X, application publication number CN 110830086 A). The baseband decoding system is composed of a register group, taps set for different rate offsets, a correlator array, and a decoding correlator array. The method steps for solving the receiver error code problem caused by large rate offset in the RFID system using the baseband decoding system are as follows: the correlator in the correlator array selects the corresponding tap according to the pre-set rate offset, and performs correlation operation on the baseband signal and the local signal output by the register group, to obtain the first correlation value and the preamble in the baseband signal. The decoding correlator in the decoding correlator array decodes the baseband signal according to the received synchronization flag bit. Although this method has high precision and high flexibility, it still has two shortcomings: first, the correlation operation adopted by the correlator depends on the absolute level of the signal and the known signal template, and the variable direct current bias will change the absolute level of the signal, resulting in mismatch between the signal characteristics and the preset signal template, thereby reducing the decoding accuracy. Second, in the environment of large dynamic variable bias signal, the calculation complexity and processing time of the correlation operation will increase significantly, and the real-time requirement cannot be met. Therefore, this method is not suitable for large dynamic variable bias signal.

[0004] The baseband decoding based on edge detection judges data by detecting the signal edge, mainly focusing on the relative change of the signal level, rather than the absolute value. This method has the ability to resist direct current bias, and does not depend on the constancy of signal-to-noise ratio, and is sensitive to strong dynamic signal, and is suitable for large dynamic variable bias signal. The role of decision threshold in edge detection is crucial, and the fixed threshold detection technology is difficult to accurately determine the edge position when applying strong dynamic signal, while the adaptive decision threshold technology can automatically adjust the decision threshold according to the received signal strength change, has a wide dynamic range and high detection accuracy.

[0005] In the published paper "Optimization of adaptive decision threshold and DC extraction technology for OOK / FSO system performance" ([J] Journal of Quantum Electronics, 2018, 35 (01)) by Zhu Baqin et al., an adaptive decision threshold method based on the quantized voltage of the decision signal to obtain the current symbol decision threshold is proposed. The method obtains the decision threshold of the current symbol through the quantized voltage of the decision signal, that is, in each symbol period, the DC component in the signal is extracted and analyzed, and the decision threshold is dynamically adjusted according to the change. The method has important reference value for baseband decoding of large dynamic bias signal. However, the method still has two shortcomings, one is that it depends on the decision signal and the extracted DC component to adjust the threshold, which leads to insufficient response when the channel condition changes rapidly. The second is that the method needs to use complex DC extraction technology to accurately extract the DC component, which will increase the complexity of the actual application. SUMMARY

[0006] The purpose of the present application is to overcome the shortcomings of the prior art, and to provide a baseband decoding method based on differential edge detection combined with adaptive decision threshold, which aims to solve the problem that the existing related technology cannot avoid the influence of variable signal-to-noise ratio and variable DC bias on decoding accuracy for large dynamic bias signal baseband decoding, and cannot meet the high real-time decoding requirements of large dynamic signal.

[0007] The idea of realizing the purpose of the present application is that for large dynamic bias signal baseband decoding, for the digital signal sampled and quantized by ADC, first, the median filter is used for processing, which can filter out pulse noise while retaining signal edge information. Secondly, the differential-based edge detection method is used, which uses the differential algorithm to eliminate the uncertain DC bias component in the original signal, thereby solving the problem of uncertain change of signal absolute level caused by variable DC bias, which leads to edge misjudgment. At the same time, the adaptive decision algorithm based on windowing statistics is used, which can adaptively update the decision threshold within a few symbol periods according to the change of signal amplitude, ensuring accurate and rapid detection of signal edges under the conditions of large range of signal amplitude and variable signal-to-noise ratio, and realizing the reconstruction of baseband signal at the receiving end. It solves the edge misjudgment caused by the inflexibility of fixed threshold when the signal level is lower than the noise limit at some sampling moments, while meeting the real-time requirements, thereby improving the decoding accuracy and efficiency, and realizing the efficient and rapid baseband decoding of large dynamic bias signal.

[0008] The implementation steps of the present application are as follows:

[0009] Step 1: Use the median filtering algorithm based on sliding window to perform median filtering processing on the digital signal sampled and quantized by ADC;

[0010] Step 2, the filtered digital signal is differentiated by using a first-order forward difference formula;

[0011] Step 3, a windowed statistical method is used to calculate an adaptive decision threshold by comparing the size of the signal differential level at each sampling time and the windowed statistical threshold value;

[0012] Step 4, the adaptive decision threshold is used to perform edge detection on the differentiated signal;

[0013] Step 5, a continuous detection window is used to perform continuity detection on the edge detection result;

[0014] Step 6, the signal waveform is reconstructed according to the continuity detection result, and the original data information is recovered.

[0015] Compared with the prior art, the present application has the following advantages:

[0016] First, the present application adopts a differential-based edge detection method, which completely eliminates the DC bias component and avoids the problem of edge misjudgment caused by the uncertain change of the absolute level of the signal due to the variable DC bias in the prior art, so that the present application can accurately detect the signal edge under any DC bias condition, thereby improving the reliability of the baseband decoding.

[0017] Second, the present application adopts an adaptive decision algorithm, which adaptively updates the decision threshold according to the change of the signal amplitude, ensuring accurate detection of the signal edge under the condition of large range change of the signal amplitude and uncertain change of the signal-to-noise ratio, overcoming the problem of edge misjudgment caused by the inflexibility of the fixed threshold value in the prior art, which causes the signal level to be lower than the noise limit at some sampling times.

[0018] Third, compared with the baseband decoding system composed of a correlator, the present application does not require complex correlation array circuits in hardware implementation; compared with the DC component extraction technology combined with an adaptive decision threshold, the present application also avoids complex edge detection algorithms in software implementation, so that the present application can meet the real-time requirements in a large dynamic environment, thereby realizing efficient baseband decoding of large dynamic variable bias signals. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 is a flowchart of the present application;

[0020] Figure 2 is a structure block diagram of a full-duplex visible light communication system based on FPGA according to an embodiment of the present application;

[0021] Figure 3 is an implementation result diagram of a full-duplex visible light communication system based on FPGA according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] The application will be further described in detail below in combination with the drawings and specific embodiments.

[0023] Figure 2 The application is applied to a full-duplex visible light communication system based on FPGA (Field Programmable Gate Array), and 10 Mbps duplex communication is realized when the distance between the sending end and the receiving end changes in the range of 0.5-5 cm. The J30J interface module is used to input the RS-422 signal, which is converted into a logic level (TTL, Transistor-Transistor Logic) signal through the isolation RS-422 conversion chip ADM2682. Then, the signal is modulated, converted into an analog signal through the digital-to-analog conversion module DAC (Digital-to-Analog Converter), and then driven by the LED driving module. One of the two LED lamp beads emits the TX+ signal, and the other emits the TX- signal. The receiving circuit uses two identical photodetectors, one of which receives the RX+ signal (corresponding to the TX+ signal) and the other receives the RX- signal (corresponding to the TX- signal). The RX+ and RX- signals are transmitted to the signal amplification module for amplification, and then the amplified signal is demodulated. The demodulated signal is sampled and quantized by the 12-bit precision ADC and output to the adaptive decoding module. Finally, the TTL level output by the adaptive decoding module is converted into an RS-422 signal through the level conversion module. The adaptive decoding module is realized by FPGA, and the internal clock is 200M. Through the method of the application, the baseband decoding in the input range of 6mV-1V can be realized (the input voltage change is allowed to be as high as 166 times).

[0024] Reference Figure 1 and Figure 2 The implementation steps of the embodiments of the application are further described in detail.

[0025] Step 1: The median filtering algorithm based on the sliding window is used to perform median filtering processing on the digital signal sampled by the ADC.

[0026] The width of the sliding window is W, and W is an odd number selected in the range of [3, 11] according to the ADC sampling frequency. If W is less than 3, the median filtering effect will not be obvious; if W is greater than 11, the signal edge will be damaged; in the embodiments of the application, the sliding window with W of 5 is selected to perform median filtering on the digital signal sampled and quantized by the ADC.

[0027] The median filtering algorithm is a nonlinear filtering method, which can effectively remove impulse noise while preserving the edge information of the signal, and the steps of median filtering processing of the digital signal after ADC sampling and quantization are as follows:

[0028] Firstly, all signal levels in the set sliding window are initialized to 0, and the signal levels after ADC sampling and quantization are moved into the sliding window one by one; when the sliding window receives a new signal level, all signal levels in the sliding window are moved to the left by one numerical unit, and the signal level at the left end is removed;

[0029] Secondly, at each sampling time, the signal levels in the sliding window are sorted, and only when the number of signal levels other than 0 in the sliding window is greater than or equal to (W1+1) / 2, W1 and W are equal in value, the middle value of the sorted signal levels in the sliding window is taken as the signal level of the median filtering output at the current sampling time, so as to ensure that the sliding window contains a sufficient number of valid signal levels, so that the signal level of the median filtering output is not 0.

[0030] Step 2, the filtered digital signal is differentiated by using the following first-order forward difference formula:

[0031]

[0032] Wherein, represents the differential level obtained by differentiating the signal level of the median filtering output at the nth sampling time, represents the signal level of the median filtering output at the (n+1)th sampling time, represents the signal level of the median filtering output at the nth sampling time. When n is greater than 0, the output is calculated by using the first-order forward difference formula .

[0033] Step 3, a windowed statistical method is adopted to calculate an adaptive decision threshold by comparing the size of the signal differential level at each sampling time and the windowed statistical threshold.

[0034] The step of calculating the adaptive decision threshold by comparing the size of the signal differential level at each sampling time and the windowed statistical threshold by using the windowed statistical method is as follows:

[0035] Firstly, a section of environmental noise without signal is collected, the amplitude range is analyzed, the numerical value of D level values higher than the maximum amplitude of the collected environmental noise is set as the windowed statistical threshold, the value of D is a rational number selected in the range of [1, 20], D is less than 1, which will result in that the adaptive decision threshold is too low, and D is greater than 20, which will result in that the adaptive decision threshold is too high, in the embodiment of the present application, the maximum amplitude of the environmental noise is 30, and the windowed statistical threshold is set to 35 when D is set to 5.

[0036] Second step, set a statistical window with a width of N, the value of N is an integer selected from the range of [10, 1000] according to the sampling frequency, N is less than 10, the statistical value in the statistical window does not have enough statistical characteristics, N is greater than 1000, the adaptive decision threshold maintains the initial decision threshold for too long, initialize all values in the set statistical window to 0, in the embodiment of the present application, set N to 10;

[0037] Third step, select a value greater than the windowed statistical threshold and less than the minimum value of the absolute value of the differential level at the first N1+1 sampling time as the initial positive decision threshold, the value of N1 corresponds to N, the initial positive decision threshold is used to provide an initial reference value before the statistical window accumulates enough valid data, to ensure that edge detection can be directly performed, in the embodiment of the present application, the initial positive decision threshold is set to 62;

[0038] Fourth step, when the absolute value of the differential level is greater than the windowed statistical threshold, the cumulative sum of the absolute values of the differential levels corresponding to the current sampling time, the previous sampling time and the next sampling time is moved into the statistical window, and the values in the statistical window are shifted left by one value unit, and the leftmost value is removed; At each sampling time when N2 non-zero differential level absolute values are not accumulated in the statistical window, the positive decision threshold is always maintained as the initial positive decision threshold, the value of N2 corresponds to N; When N2 non-zero differential level absolute values are continuously accumulated in the statistical window, from the next sampling time, at each sampling time in the statistical window, the statistical value in the statistical window is averaged and then divided by a fixed statistical parameter D as the positive decision threshold of the current sampling time, the value of D is a rational number selected from the range of [2, 5], in the embodiment of the present application, D is set to 2.8;

[0039] Fifth step, at each sampling time, the positive decision threshold is taken as the negative decision threshold of the current sampling time, since the sampling signal as a whole has symmetry, setting the negative decision threshold can effectively detect the negative edge, thereby perfecting the overall edge detection method.

[0040] Step 4, using the adaptive decision threshold to detect the edges of the differential processed signal.

[0041] The edge detection of the signal after the differential processing by using the decision threshold refers to judging the differential level of the signal at each sampling time, when the differential level of the signal is greater than 0 and greater than the positive decision threshold value at the current sampling time, it is determined that the signal rises at the current sampling time, the signal jumps from 0 to 1, and the rising edge flag of the signal is 1. When the differential level of the signal is less than 0 and less than the negative decision threshold value, it is determined that the signal falls at the current time, the signal jumps from 1 to 0, and the falling edge flag of the signal is 1; otherwise, it is determined that the signal remains unchanged, and the rising edge flag and the falling edge flag of the signal are both 0.

[0042] Step 5, the results of the edge detection are detected continuously by using a continuous detection window;

[0043] The continuous detection window refers to setting two continuous detection windows C1 and C2 with the same width K, K is an integer selected from the range of [2, 10], K is less than 2, which will cause that the false edges caused by transient noise or temporary interference cannot be effectively filtered out, and K is greater than 10, which will cause that the real edges of the signal cannot be tracked, in the embodiment of the application, C1 and C2 are set to 3.

[0044] The continuous detection of the results of the edge detection refers to filtering out the continuous edge mutations caused by single-point noise or temporary interference by continuously detecting the results of the edge detection after detecting the edges of the signal, so as to ensure that only one valid edge is detected in the multiple continuous edges, and the steps are as follows:

[0045] Firstly, all the values in the continuous detection windows C1 and C2 are initialized to 0, and the count values of the continuous rising edge counter and the continuous falling edge counter are initialized to 0;

[0046] Secondly, if the rising edge flag of the signal at the current sampling time is 1, the rising edge flag of the signal is moved into the continuous detection window C1, and the count value of the continuous rising edge counter is increased by 1, a new signal rising edge flag is received each time, and the values in C1 are shifted left by one value unit at the same time, and the leftmost value in C1 is removed; similarly, if the falling edge flag of the signal at the current sampling time is 1, the falling edge flag of the signal is moved into the continuous detection window C2, and the count value of the continuous falling edge counter is increased by 1, a new signal falling edge flag is received each time, and the values in C2 are shifted left by one value unit at the same time, and the leftmost value in C2 is removed.

[0047] Third, when the count value of the continuous rising edge counter is equal to K, the rising edge flag corresponding to the count value of the rising edge counter in C1 is determined to be a valid rising edge flag, and the valid rising edge flag is output; when the count value of the continuous falling edge counter is equal to K, the falling edge flag corresponding to the count value of the falling edge counter in C2 is determined to be a valid falling edge flag, and the valid falling edge flag is output.

[0048] Step 6: Reconstruct the signal waveform based on the continuity detection results to recover the original data information.

[0049] When no signal transmission is set according to the baseband signal characteristics, the receiving end level is always 0 or 1. When a valid rising edge flag is detected, the receiving end signal level is set to 1; when a valid falling edge flag is detected, the receiving end signal level is set to 0; otherwise, the receiving end signal level remains unchanged. In the embodiment of the present invention, when no signal transmission is set, the receiving end level is always 1.

[0050] The following is combined with Figure 3 The implementation result diagram further describes the effects of the present invention.

[0051] The FPGA-based full-duplex visible light communication baseband decoding effect of the invention embodiment, as observed by an oscilloscope, is as follows: Figure 3 As shown in the diagram, the yellow waveform is the digital baseband signal output after ADC sampling and quantization, while the purple waveform is the signal containing the original data information obtained after baseband decoding using the algorithm of this invention. By directly comparing the two waveforms, it can be clearly seen that this invention reconstructs the digital baseband signal waveform within several symbol periods, achieving efficient decoding of the digital baseband signal. This overcomes the shortcomings of existing technologies for baseband decoding of large dynamic bias signals, demonstrating accuracy and reliability.

Claims

1. A baseband decoding method based on differential edge detection combined with adaptive decision threshold, characterized in that, The baseband decoding method involves differentiating the digital signal after ADC sampling, quantization, and median filtering; determining an adaptive decision threshold based on the windowing statistical threshold and the magnitude of the signal differential level at each sampling time; and performing edge detection on the digital signal before differential processing. Step 1: A median filtering algorithm based on a sliding window is used to perform median filtering on the digital signal after ADC sampling and quantization. Step 2: Differentiate the filtered digital signal using the first-order forward difference formula; Step 3: Using the windowed statistical method, the adaptive decision threshold is calculated by comparing the signal differential level at each sampling time with the windowed statistical threshold. The steps of the windowed statistical method are as follows: The first step is to collect a segment of ambient noise without signal, and set the values ​​of D level values ​​that are higher than the maximum amplitude of the collected ambient noise as the windowing statistical threshold. The value of D is a rational number selected in the range [1,20]. The second step is to set a statistical window with a width of N, where N is an integer selected from the range [10, 1000] based on the sampling frequency; and initialize all differential levels within the statistical window to 0. The third step is to select a value that is greater than the windowed statistical threshold and less than the minimum absolute value of the differential level in the first N1+1 sampling times as the initial positive decision threshold, and the value of N1 is equal to that of N. Fourth step: If the absolute value of the differential level at the current sampling time is greater than the windowed statistical threshold, the sum of the absolute values ​​of the differential levels at the current sampling time, the previous sampling time, and the next sampling time is moved into the statistical window, and the differential levels in the statistical window are shifted to the left by one numerical unit, and the leftmost differential level is removed; at each sampling time when N2 non-zero absolute values ​​of differential levels have not been accumulated in the statistical window, the positive decision threshold is always kept at the initial positive decision threshold, and the value of N2 is equal to that of N; when N2 non-zero absolute values ​​of differential levels have been accumulated in the statistical window, starting from the next sampling time, at each sampling time in the statistical window, the statistical values ​​in the statistical window are averaged and then divided by a fixed statistical parameter D as the positive decision threshold for the current sampling time, and the value of D is a rational number selected in the range of [2,5]. Fifth step: At each sampling time, the calculated positive decision threshold is inverted and used as the negative decision threshold for the current sampling time; Step 4: Use an adaptive decision threshold to perform edge detection on the differentiated signal: judge the differential level of the signal at each sampling time. When the differential level of the signal is greater than 0 and greater than the positive decision threshold value at the current sampling time, it is determined that the signal is rising at the current sampling time, the signal jumps from 0 to 1, and the rising edge flag of the signal is set to 1. When the differential level of the signal is less than 0 and less than the negative decision threshold, it is determined that the signal is decreasing at the current moment, the signal jumps from 1 to 0, and the falling edge flag is 1; otherwise, it is determined that the signal remains unchanged, and both the rising edge flag and the falling edge flag are 0. Step 5: Use a continuous detection window to perform continuity detection on the edge detection results; Step 6: Reconstruct the signal waveform based on the continuity detection results to recover the original data information.

2. The baseband decoding method based on differential edge detection combined with adaptive decision threshold as described in claim 1, characterized in that, The width of the sliding window mentioned in step 1 is W, which is an odd number selected based on the ADC sampling frequency within the range of [3, 11].

3. The baseband decoding method based on differential edge detection combined with adaptive decision threshold as described in claim 1, characterized in that, The median filtering algorithm described in step 1 performs median filtering on the digital signal sampled and quantized by the ADC. The steps are as follows: The first step is to initialize all signal levels within the set sliding window to 0, and then move the ADC-sampled and quantized signal levels into the sliding window one by one; each time the sliding window receives a new signal level, it shifts all signal levels in the sliding window to the left by 1 unit and removes the leftmost signal level. The second step is to sort the signal levels within the sliding window at each sampling time. Only when the number of non-zero signal levels within the sliding window is greater than or equal to (W1+1) / 2, the values ​​of W1 and W are equal. The median value of the sorted values ​​within the sliding window is then used as the median filtered signal level at the current sampling time.

4. The baseband decoding method based on differential edge detection combined with adaptive decision threshold as described in claim 1, characterized in that, The first-order forward difference formula mentioned in step 2 is as follows: ； in, This represents the differential level obtained after differentiating the signal level of the median filter output at the nth sampling time. This represents the signal level of the median filter output at the (n+1)th sampling time. This represents the signal level of the median filter output at the nth sampling time, where n is an integer greater than or equal to 0.

5. The baseband decoding method based on differential edge detection combined with adaptive decision threshold according to claim 1, characterized in that, The differential processing mentioned in step 2 refers to the output calculated using the first-order forward difference formula when n is greater than 0. .

6. The baseband decoding method based on differential edge detection combined with adaptive decision threshold according to claim 1, characterized in that, The continuous detection window mentioned in step 5 refers to setting two continuous detection windows with the same width and both being K, namely C1 and C2, where K is an integer selected from the range [2, 10].

7. The baseband decoding method based on differential edge detection combined with adaptive decision threshold as described in claim 1, characterized in that, The steps for performing continuity testing on the edge detection results in step 5 are as follows: The first step is to initialize the values ​​of the continuous detection windows C1 and C2 to 0, and to initialize the count values ​​of the continuous rising edge counter and the continuous falling edge counter to 0. The second step is as follows: If the rising edge flag of the signal at the current sampling time is 1, the rising edge flag of the signal is shifted into the continuous detection window C1, and the count value of the continuous rising edge counter is incremented by 1. Each time a new rising edge flag is received, the values ​​in C1 are shifted to the left by one numerical unit, and the leftmost value in C1 is removed. Similarly, if the falling edge flag of the signal at the current sampling time is 1, the falling edge flag of the signal is shifted into the continuous detection window C2, and the count value of the continuous falling edge counter is incremented by 1. Each time a new falling edge flag is received, the values ​​in C2 are shifted to the left by one numerical unit, and the leftmost value in C2 is removed. The third step is to determine that the signal rising edge flag corresponding to the rising edge counter value of 1 in C1 is a valid signal rising edge flag and output the valid signal rising edge flag. When the count value of the continuous falling edge counter is equal to K, the signal falling edge flag corresponding to the count value of the falling edge counter in C2 is determined to be a valid signal falling edge flag, and the valid signal falling edge flag is output.

8. The baseband decoding method based on differential edge detection combined with adaptive decision threshold according to claim 1, characterized in that, The step 6, which describes reconstructing the signal waveform based on the continuity detection results and recovering the original data information, refers to setting the receiver level to 0 or 1 when there is no signal transmission, based on the characteristics of the baseband signal. When a valid rising edge flag is detected, the receiver signal level is set to 1. When a valid falling edge of the signal is detected, the signal level at the receiving end is set to 0; In all other cases, the signal level at the receiving end remains unchanged.

Images