Binary signal decoding apparatus and method

Abstract

An apparatus and method is provided for decoding a transmitted comma free coded signal, having a plurality of bits and two message frequencies, and suitable for application in controlling the passenger doors of a vehicle station. A sampling of the signal bits is made at a rate that is a multiple of the known transmission rate and a predetermined selection relationship is applied to make unnecessary the provisions of receiver synchronization apparatus responsive to the transmission rate control clock. This selection relationship involves summing the values of an odd plurality of sample points during a bit period and comparing this sum to a predetermined number to determine the binary value of the bit.

Description

BACKGROUND OF THE INVENTION

In the prior art it was desired to transmit and receive six bit comma free coded word commands back and forth between a vehicle and a station. These word commands had to be synchronized, since the six bits of each coded word were sent in serial form one word after another word. If the comma free code for a given word happened to be 101111, the transmitter would send 101111, 101111, and so forth repeatedly. When decoding the word on either the station or the vehicle, the decoding had to be synchronized as disclosed in U.S. Pat. Nos. 3,562,712; 3,992,698 and 4,015,082 with a recovered transmission clock in relation to a particular word. The previous practice to indicate the beginning of each bit in a word was to reverse the phase of the bit. For each bit whether it changed from a one to a zero, a zero to a one or if there were two ones or two zeros, every 18th of a second the phase was reversed, as shown by U.S. Pat. No. 3,551,889. A sinewave signal of a first frequency is transmitted for a ZERO and a sinewave signal of a second frequency is transmitted for a ONE. Each phase reversal indicates the beginning of another bit of the six bit comma free coded signal. It was the prior art practice in the transit industry in relation to command signal transfer to recover the source clock information, along with the binary data, and to synchronize the signal decoder with the source clock information. A circuit was provided to detect this phase reversal to recover the source clock information. Each bit lasted 1/18 second, with a typical frequency of 8880 hertz for the ZERO code and frequency of 5920 hertz for the ONE code. It is known in the prior art, as discussed in a published article entitled "Design Techniques For Automatic Train Control", by R. C. Hoyler in the July, 1972 issue of the Westinghouse Engineer, to send out information signals from the transmitter at a station or on a vehicle in the 5 to 10 kilohertz range and which signals are received by a receiver at the other of the vehicle or the station, including hardware circuits that demodulate the signal into the form of digital data. The BART train control apparatus performed such an operation. The input signal to the receiver is in the form of a 5 to 10 kilohertz FSK frequency shift keyed comma free code signal, but the demodulator takes that signal and converts it into digital data including ones and zeros, as disclosed in U.S. Pat. No. 3,992,698. The signal level is about five or six volts when it is a ONE and zero volts when it is a ZERO. The input signal line is going to be going high and low representing the ONE and ZERO bit signals that are received continuously, and it is desired to know what the meaningful code signal is out of that data stream.

For the example of vehicle and station door control, it is desired to recover out of this received data stream the coded word information that was originally embedded in it by the station transmission to the vehicle. A typical door control system transmits from the station to the vehicle four command word bit patterns, namely to open alighting doors, to open boarding doors, to open both doors, and to close both doors. For this purpose, four different code word commands are transmitted to the vehicle, with the first code word command being to close all doors. Since there are doors on both sides of the car, this coded word command signal closes doors on both sides of the car. Whenever a car pulls into a station lobby, one side of the car and station is designated as alighting for people setting off and one is designated as boarding for people getting on. The people come in the car from one side and go off the car from the other side. The alighting doors are opened first, which means the people on the car will start to exit toward one side of the car because those doors open first, and after everybody starts moving out then the boarding doors are opened a few seconds later. In that way people move off the car and move on the car through the vehicle and station doors at the same time to speed up the people transfer process. Typically, signals are first sent to open alighting doors, then signals are sent to open both doors, and then signals are sent to close both doors. The station initially sends these signals to the vehicle and the vehicle in turn sends these same code signals back to the station.

There is provided a decoding apparatus and method to interpret a binary code pattern of ones and zeroes that does not need an additional signal at the receiver representing the clock frequency of the source originating the received signal command.

SUMMARY OF THE INVENTION

The input command signal code is supplied to the present signal decoder apparatus and method in the form of ones and zeros, representing the actual information data stream received from the source transmitter. With the approximate clock frequency of the source known, each bit of the incoming code signal is sampled at a multiple of the approximated frequency of the source clock rate to provide at least three samples of that bit. For the example of an 18 hertz clock frequency, an illustrative sampling multiple of 4.0 or 72 hertz is used in the following description.

Three of the four samples at the 72 hertz rate are added to each other, and the fourth sample is not considered. One of the first samples or the last sample can be discarded and the other three of the four samples can be summed. It is desired that an odd number of samples be summed, such that a majority of those samples will determine the value of the sampled input signal bit. If the sum of the three sample bits equals 2 or 3, the incoming data bit is declared a ONE. Otherwise, it is declared a ZERO. In either case, the sum is stored.

To provide accurate decoding without the source clock rate signal, if the first, second or third sample indicates a ONE and the current sample sum equals ZER0, then the one through four sample counter is reset to a count of one, regardless of its previous setting. This has the effect of synchronizing the decoding sampling process sample set to the rising edge of the first ONE bit received in each group of samples. The sampling of the data stream can be repeated for each data bit of the word command received. The decoded bit is stored in a shift register device which permits storage, comparison or manipulation of all or part of the decoded data stream.

In accordance with the present invention, the voltage of each bit of the received input signal is sampled by strobing at a 72 hertz rate with a pulse having a time duration substantially less than the time duration of the sample interval. Since the coded information is transmitted at an 18 hertz rate, a sample at the receiver is made at 4 times the information data rate.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows an illustrative prior art signal waveform for a comma free coded word;

FIG. 2 shows a prior art binary signal decoding apparatus for a coded command word signal having two message frequencies;

FIG. 3 shows signal waveforms to illustrate the output signals provided by the signal decoder apparatus of FIG. 2;

FIG. 4 shows the signal decoding receiver apparatus of the present invention applied in relation to a transmit vehicle system door control apparatus;

FIG. 5 shows the functional operation of the present signal decoding apparatus;

FIGS. 6A and 6B show a program routine flow chart to functionally illustrate the operation of the present binary signal decoder apparatus and method;

FIG. 7 shows a reset of the local clock to correct an undesired shift of the sampling operation.

PREFERRED EMBODIMENT OF THE INVENTION

The preferred embodiment of the invention discloses a signal decoding apparatus and method operative with a comma free binary coded command signal. One suitable aplication is shown in relation to station door control apparatus for a transit vehicle system.

FIG. 1 shows a typical and well known frequency shift key modulation (FSK) waveform 10 for a comma free command signal word that has been utilized in the prior art for the control of transit system operations, such as illustrated in U.S. Pat. Nos. 3,551,889 and 4,015,082, the disclosures of which are incorporated herein by reference. The command signal word is shown as having six binary bit values, however, it is well known to provide more or less bits for a given signal word. The waveform 12 shows the corresponding binary signal values for each bit of the command signal word shown by curve 10.

In FIG. 2 there is shown a prior art binary signal decoding apparatus for a coded command word signal having two message frequencies, as disclosed in greater detail in U.S. Pat. Nos. 4,015,082. The binary coded message is comprised of 1 and 0 message words in which the 1's are transmitted at a first frequency and the 0's are transmitted at a second frequency. A coded command signal is sensed by an antenna 20 which is coupled through transformer 22 to a preamplifier 24. The preamplifier 24 is coupled through a transformer 26 to signal channels 28 and 30 of the multiple channel signal decoder 32. The signal channel 28 includes bandpass filter 34, threshold detector 36, synchronous clock 38, a storage shift register 40 and decoder 42. The bandpass filter 34 is tuned to the 1 frequency and the threshold detector 36 operates to detect the presence of a 1 frequency. The signal channel 30 includes bandpass filter 46 tuned to the 0 frequency, threshold detector 48, synchronous clock 50, a storage shift register 52 and decoder 54. Output signal lines 56 and 58 connect with respective signal processors 60 and 62, where the decoded command signals are compared with feedback signals. For the example of vehicle speed command signals as shown in U.S. Pat. No. 4,015,082, the actual vehicle speed feedback signal 64 is compared with the 1 frequency desired speed command signal 58 to determine a speed error 66, and the actual speed feedback signal 68 is compared with the 0 frequency desired speed command signal 56 to determine a speed error signal 70. If the error signals 66 and 70 are substantially the same, as compared in AND gate 72, the speed error signal is applied to control the vehicle traction equipment 74.

In FIG. 3 there is shown the output signal waveform 80 that is provided by the 1 frequency filter 34, and the output frequency signal waveform 82 is provided by the 0 frequency filter 46. The output binary signal waveform 84 is provided by the 1 frequency detector 36, and the output binary signal waveform 86 is provided by the 0 frequency detector 48.

In FIG. 4 there is shown the signal decoding receiver apparatus of the present invention in an application for the control of passenger doors. This could typically be used for controlling the doors of a passenger station in relation to a transit vehicle arriving at that station and then leaving that station location.

Comma free command signals can be utilized for this purpose as follows:

100001 to open alighting door; 101010 to open boarding door; 100011 to open both doors; 100111 to close both doors. In an article entitled "Atlanta Airport People Mover", by T. C. Selis that was published in the Conference Record for the 28th IEEE Vehicular Technology Group Meeting on Mar. 22, 1978, there is diagrammatically shown a typical door and dwell time control apparatus that is operative with both the vehicle passenger doors and the station lobby doors when a passenger vehicle arrives at a station location.

For the example of the six bit comma free command signal 100111 being provided to close the vehicle and the station lobby doors, the antenna 20 would sense this signal. In the station lobby door control application, the antenna 20 would receive this signal from a vehicle carried antenna. The 1 signal filter 34 would output the signal waveform 80 as shown in FIG. 3, and the 1 signal detector 36 would output the signal waveform 84. The 0 signal filter 46 would output the signal waveform 82 as shown in FIG. 3, and the 0 signal detector 48 would output the signal waveform 86. The signal waveforms 84 and 86 could have a high value of about 5 volts for each detected bit signal having respectively a 1 or 0 binary value as shown. A bistable circuit device, such as flip-flop 90, is connected to provide a high or 1 output signal in response to each high value of the signal 84 and to provide a low or 0 output signal in response to each high value of the signal 86. In this way the output signal waveform 88 shown in FIG. 3 is provided by the flip-flop 90 in response to the comma free command word 100111 that controls the closing of the station lobby doors on both sides of the passenger vehicle. The signal 88 is transmitted at a known frequency, such as 18 hertz. The signal decoding apparatus 92 operates to sample the voltage of each bit of the signal 88 at a predetermined multiple of the input signal frequency, such as four times 18 hertz or at a 72 hertz rate, to provide four samples to be taken for each transmitted signal bit. This timing relationship of the sample strobing is illustrated at the bottom of FIG. 3 for each of the six bits shown in that figure, to provide four equally spaced strobes 1, 2, 3 and 4 for every 1/18 second time interval of each bit of the transmitted signal 88. The decoded signal bits for each received command word are stored in a shift register to permit storage comparison or manipulation of all or part of the decoded data stream. The decoded output signal from the signal decoding apparatus 92 is applied to the well known door control apparatus 94, which can be similar to that shown in the latter published article.

In FIG. 5 there is shown the functional operation of the present signal decoding apparatus. The door control command signal, such as shown in FIG. 3, is sent from the station to the vehicle when a passenger vehicle arrives at a station location. The vehicle then retransmits the door control command signal from the vehicle door control apparatus 100 through the vehicle antenna 102 back to the station antenna 104 for the control of the station lobby doors. The filter and detector 106 provides the command signal waveforms, such as shown in FIG. 3. The signal decoding apparatus 92 then samples the command signal bits of waveform 88 at functional block 108, determines the respective values of sampled data bits at functional block 110 and then stores and outputs the command signal data bits at functional block 112. The station door control apparatus 94 responds to the output command signal word, such as 100111 shown in FIG. 3 to close both station lobby doors on the respective sides of the passenger vehicle in the station location.

In FIG. 6 there is shown a program routine flow chart to illustrate the functional operation of the signal decoding apparatus 92 as shown in FIG. 5. The first part of the flow chart keeps track of the four samples taken for a given data bit, and a local 1 to 4 counter is incremented at a 72 hertz rate, to indicate when a sample set of four counts has been taken. After the data bit has been sampled four times, a determination is made based on a selected portion of these four samples to establish if the transmitted bit was a one or a zero. The program routine operates at a 72 hertz rate to sample the input data bit as determined by a local clock.

The present sampling operation looks to see if a particular sample is a one and the previous samples in that sample set are zeros. For instance, if at the third of four samples a one is found, and the previous two samples were zeros, the program operates to reset the local counter to one to begin another sample set and to catch the rising edge of that one. The sampling operation will then start counting another three samples for a given signal bit, with the first sample being the last one that was just detected. This is done on a one, but only on the first one, in a particular sample set as shown in FIG. 7 in relation to the comma free input signal 101010, where the sampling operation shown by line 114 illustrates a situation where the local clock has shifted for some reason and improper determination of the signal bits shown by curve 89 might result. The curve 91 would be output by the flip-flop 90 for the comma free input signal 101010. This reset of the local counter results in the sampling operation beginning the sample set synchronized in time with the signal 91 as shown by line 116. On the other hand, if the sample operation samples four ones and then samples four zeros and then samples four ones, there is no need to shift the sample set, and there is similarly no need to shift when two ones and then two zeros are sampled. In addition, the present sampling after a set of four samples throws away the fourth sample. The determination process is to accumulate a set of four samples for a given input data bit, and on the basis of a majority of ones or a majority of zeros in the remaining three bits of that sample set, the particular input data bit at the 18 hertz rate is then declared either a one or a zero. At this point the declared bit is stored to build up a re-constructed six bit data word stream. If a set of four samples finds three zeros, this bit is declared a zero. If the three samples in a set are all ones, the bit is declared a one. In the event of two zeros and a one for the three samples of a set, that bit is declared a zero. For a sample set including two ones and a zero, that bit is declared a one.

There is no problem with using the four samples of a given set, if all four of them are ones, since that input data bit is a one. If three of the four samples in a sample set are ones, the bit is a one. A decision was made to use only three of the samples in a given set rather than all four samples, to avoid a problem that occurs when for a given sample set two samples are ones and two samples are zeros, since there is no certain way to declare the value of that bit. When only three samples are used to determine each input data bit, a majority of those samples will establish if the bit is a one or a zero. The local clock is the standard interrupt function of an Intel 8080 microprocessor, and an interrupt function is selected in this case at a 72 hertz rate.

The flow chart shown in FIG. 6A and FIG. 6B illustrates the algorithm and logic processing of the command signal decoder apparatus of the present invention. The program operates at an interrupt clock driven rate. The flow chart block 120 is related to acquiring the sample data in relation to providing a strobe, compliment and mask for the receiver data line sample of the input signal. The program is going out to input data line supplied by the filter and detector 106 shown in FIG. 5 and looking at the input signal, which is stored as a snapshot of the data line at that time. At block 122 a sample counter is incremented, which counter is an individual counter in each of the several receiver inputs that can be processed since the program is written to process several receiver inputs if desired. For this description of the present invention, only one receiver input is processed. This one to four counter can be reset independently for each input data signal receiver. A typical train of passenger vehicles communicating with a station has more than one transmitter. For example, if there are two cars, there can be two transmitters and all of the input data comes to a common station signal decoder apparatus. The algorithm used to increment the sampling counter that is assigned to each receiver is determining whether this is sample count number four of a given sample set, because when the count equals four the program is going to determine for a sampled receiver data line if the sample set sum establishes a one or a zero data bit. If this is less than the fourth count, the sample is just stored to accumulate the samples in a particular sample set.

The program is running an accumulative sum within a particular range. If this is the first one sampled, it is desired to start the one to four count over again. When the count of four is reached for a given sample set, the program in effect discards the fourth sample by not storing it in the summation of samples. However, during that time of the fourth count, the program looks at the accumulative sum of the samples and determines if the sum is equal or greater than two, and if it is then the sampled bit is declared to be a one. Decision block 124 determines whether the program is at a sample count four or less than four. Block 126 operates when the program is at a sample set count of four to decide when the cumulative sample sum is one of equal or greater than two or less than two to declare the sampled data bit at the 18 hertz rate was a one bit or a zero bit. If it is equal to or greater than two, it is declared a one and anything less than two is declared a zero. Blocks 128 and 130 store just a one or a zero respectively based on the outcome of block 126. Blocks 128 and 130 provide a temporary storage location of the decoded bit until a later part of the program incorporates the contents of block 128 and 130 into existing data to re-construct the transmitted data word. Blocks 132 and 134 store just a one or a zero respectively for optional use by other programs or hardware that will energize an operator light on a panel to indicate the decoded bit is a one or a zero. Block 136 tags the decoded bit in this example by setting the most significant bit to one to indicate that a complete sample set has been achieved for the receiver input decoded bit. The decoded bit it tagged to facilitate separating the program functions of sampling and bit decoding from the program functions of data word decoding. Block 138 sets the storage location for the sum of the sample bits to zero in anticipation of the next sample set for the associated receiver input signal. Whenever the sample counter is less than four at block 124 the strobed receiver input data signal is tested to determine whether it is a one or a zero at block 142. If the signal is a zero the signal value is accumulated into the current sample set at block 148. If the signal is a one when tested at block 142 then the accumulated sample sum is also tested to determine if this is the first instance of a one value in the partial sample set at block 144. Whenever the partial sample set contains a value equal or greater than one, the input data signal having a value of one is accumulated in the current sample set at block 148. However, when the partial sample set at block 144 is zero the program resets the sample counter to a count of one at block 146, and then adds the input signal value to the sample set at block 148. The significance of resetting the sample counter to one is that it has the effect of re-calibrating the sample counter to the transmission clock rate each time the first one value is strobed in each sample set. Each valid comma free code contains at least a single one bit, therefore this counter is re-calibrated at a minimum of once each word interval.

At this point the sampling operations have been completed for a single receiver input and either the strobed input value has been stored in a partial sample set or the decoded sample bit has been tagged and stored re-constructing the data word later in the program. Block 140 provides a loop to sample a bit decode each receiver input data line in applications containing more than one receiver input. The next step is to re-construct the transmitted data word.

At block 150 of FIG. 6B a counter is incremented each time the program is called. The decision block 152 resets the counter to zero at block 154. For each sample set of four counts this counter is included to provide address offsetting when multiple receiver input signal lines are being processed. In applications containing only one receiver input this counter would not be necessary, nor would block 156, which selects the receivers to be processed on a cyclical basis; nor would block 160 which provides the capability of processing more than one receiver each time through the program.

The actual word decoding starts at block 158 which tests whether the decoded bit location of the associated receiver input has a complete sample set and also has not been processed by this section of the program. When the decoded bit contains a tag which was inserted at block 136 the bit will be processed, otherwise the processing will be omitted for this receiver input data on this pass through the program.

If the sample was tagged it indicates that it can be built into a six bit word. This most significant bit is not tagged for example, when the counter is reset to one because a one following some zeros had occurred. Since the loop from block 156 on is occurring at an 18 hertz rate, it could be thrown out of time or out of synchronization if it is not tagged, so to avoid this it will ignore further processing of this bit and wait until it is tagged, which happens at block 158 that goes to block 160 for skipping this intermediate phase. At block 162 a valid sample has been decoded, and the tag has to be removed because from this point on only the data bit is of interest. Block 164 is a decision block which is determining that the comma free code is a six bit word as compared to perhaps a future interest in a seven bit or eight bit command signal word. The present application uses a six bit code word.

Block 166 and following are related to taking the decoded bit and using it to construct a six bit comma free code word. Block 166 takes the previous six bits of information that are stored, and sets the three most significant bits to zero. Since a six bit comma free command signal word is involved with the available storage eight bits wide, the two most significant bits remain zero throughout this program. The hardware of the computer is eight bits and the desired code word is only six bits wide, so for bookkeeping purposes all of the locations worked with are eight bits wide and one assigned convention is that all of the desired six bit codes are going to occupy the least six bits in any location. Another convention is that unused locations are to be zero to facilitate using lookup tables with the decoded word, so block 166 will zero the three most significant bits, because since the present six bits represent old information the oldest data bit is thrown away and then the remaining five bits are moved over one to make room for the new data bit. Block 168 rotates the last five bits of data to make room for the most recent data bit in the least significant bit location. At block 170 the new data bit is ORed with the previous five data bits, and the OR function then creates a six bit comma free code that represents the data stream up to that point. Block 172 uses conventional and well known program operations to determine if the decoded data word in block 170 is a valid bit pattern. This application uses a sorting algorithm, however a lookup table can also be used. The decoded word and whether it is a valid code, are stored for use by other control programs known in the prior art to effect the desired door control in this application before this program is exited.

In the Appendix there is included an instruction program listing that has been prepared to control the signal decoding. The instruction program listing is written in the assembly language of the Intel 8080 microprocessor. Many of these microprocessors have already been supplied to customers, including technical instruction manuals and descriptive documentation to explain to persons skilled in the art the operation of the microprocessor apparatus. This instruction program listing is included to provide an illustration of one suitable embodiment of the present invention that has been developed. This instruction program listing at the present time has not been extensively debugged through the course of practical operation for the real time control of a signal decoding operation. It is well known by persons skilled in this art that most real time control application programs contain some bugs or minor errors, and it usually takes varying periods of actual operation time to identify and routinely correct the more critical of these bugs. ##SPC1## ##SPC2##

Claims

1. In apparatus for decoding an input signal word having a plurality of bits and a known transmission rate, said signal word including first and second coded message frequencies, the combination of

means for filtering and detecting a first message frequency signal having one of a high and low value in accordance with each bit of the comma free code of said first frequency,

means for filtering and detecting a second message frequency signal having one of a high and low value in accordance with each bit of the comma free code of said second frequency,

means providing a bit of a third signal for each bit of the first and second message frequency signals,

means for sampling each said bit of the third signal at a sampling rate that is a selected non-unity multiple of said transmission rate to provide a sample set for each bit of the third signal consisting of an odd plurality of samples,

means for determining the binary value of each bit of the third signal in accordance with the sum of the odd plurality of samples being greater than a predetermined number, and

means for establishing the binary value for all of the plurality of bits of the input signal word.

2. The signal decoding apparatus of claim 1, with said determining means being operative to determine if the sum of said samples in each sample set is equal to or greater than said odd plurality of samples minus one.

3. The signal decoding apparatus of claim 1, with the sampling means initially providing more than one desired number of samples in the sample set for each bit of the third signal and with the determining means then being operative to not consider at least one of those samples such that a desired odd plurality of samples is thereby provided.

4. The signal decoding apparatus of claim 1, with said determining means being operative to determine the binary value of each bit of the third signal in accordance with the sum being at least a majority of said odd plurality of samples.

5. The signal decoding apparatus of claim 1, including means for resetting a given sample set in relation to said sampling rate when the first, second or third sample in said sample set indicates a ONE value and the sum of the previous samples taken for that sample set is ZERO.

6. A process for decoding an input signal word having a plurality of bits and a known transmission rate, with said signal word having first and second coded message frequencies, including the steps of:

providing a first signal having one of a ONE or a ZERO value bit in accordance with each bit of the first frequency,

providing a second signal having one of a ONE or a ZERO value bit in accordance with each bit of the second frequency,

providing a third signal having a ONE value bit for each ONE value bit of the first signal and having a ZERO value bit for each ONE value bit of the second signal,

sampling each bit of the third signal at a sampling rate that is a predetermined non-unity multiple of said known transmission rate to provide at least an odd plurality of samples for each bit of the third signal,

determining the binary value of each bit of the third signal in accordance with the sum of said odd plurality of samples being at least a majority of those samples, and

establishing the binary value for all of the plurality of bits of the input signal word to decode said input signal word.

7. The input signal word decoding process of claim 6,

with the sampling step being in accordance with more than said odd plurality of samples.

8. The input signal word decoding process of claim 6,

with the sampling step initially providing more than said odd plurality of samples, and

with the determining step being operative in accordance with the sum of said odd plurality of samples and in relation to a predetermined total number.