SENSOR AND METHOD FOR DETERMINING A RUNTIME

Description

[0001] The invention relates to a sensor and a method for determining a transit time, in particular a distance, according to a pulse-based transit time method, according to the preamble of claim 1 and 10, respectively.

[0002] Time-of-flight distance measurement is known for various signals, such as microwaves, light, or ultrasound. Many sensor types are available specifically for optical distance measurement, such as single-beam rangefinders, laser scanners, or 3D cameras. Such sensors can be used, for example, in vehicle safety, logistics, or factory automation. A particular application is safety technology, where the sensor monitors the area surrounding a hazard and initiates a safety measure, such as an emergency stop, upon detection of an unauthorized object.

[0003] There are various variants of time-of-flight measurement. In a phase method, the transmitted signal is modulated in its amplitude, and a phase shift of the received signal is evaluated. In a pulse method, a pulse is emitted, and the time until the pulse reflected from the object to be detected is received is evaluated. This method is also referred to as direct time-of-flight measurement (dToF). A pulse averaging method, known, for example, from EP 1 972 961 B1, repeatedly transmits and receives pulses and statistically evaluates them.

[0004] The time elapsed between the transmitted and received pulse must be measured with great precision. This is a challenging task and places very high demands on the hardware. If the distance measurement resolution is to achieve an accuracy of even a few tens of millimeters, the signal propagation time must be determined with an accuracy of a few hundred picoseconds. Determining the reception time with sufficient accuracy is a major challenge. The sampling rate for digitizing the received pulse would have to be several tens or even hundreds of GHz or more for the aforementioned time resolutions.

[0005] Various approaches are being pursued in the state of the art to circumvent this high and fine sampling. For example, attempts are made to reconstruct the pulse shape, and in particular the pulse center, after a mere 1-bit conversion from the amplitude, width, or energy of the pulse. However, this requires additional hardware, and the pulse reconstruction must be laboriously trained. In a so-called constant fraction discriminator, the signal is delayed and inverted in a second path, and then the zero crossing of the superposition of both paths is interpolated. For this to happen, however, the original sampling must be at least fast enough to capture the pulse at all. Even then, the interpolation is only as accurate as the sampling was able to capture the pulse amplitudes with just a few or even a single sampling point.

[0006] EP 2 762 920 B1 discloses a method for determining signal propagation time, in which a sample is computationally reconstructed with an effective resolution higher than that of the actual sample. In addition, it is proposed to pre-expand the input signal with a filter and incorporate its transfer function into the reconstruction. This method is quite complex and requires modeling of the input signal.

[0007] US Patent No. 7,834,795 B1 deals with a sensor array with compressed sensing to save hardware while still meeting the Nyquist criterion. However, this is based on multi-channel acquisition of a multitude of sensor signals, which are often not available in a time-of-flight measurement.

[0008] In EP 2 189 805 B1, the pulse averaging method of the aforementioned EP 1 972 961 B1 is further improved by individually delaying the repeated transmitted pulses according to a predetermined distribution. This effectively samples a displaced point of the pulse at the receiving end, thus refining the resolution. This would not be feasible in a single measurement, and it requires a slowly changing measurement environment, ideally with the same object being detected at the same location throughout the entire measurement.

[0009] EP 2 199 761 A1 discloses repeating the time-of-flight measurement multiple times and sampling the received signal with a comparatively coarse resolution and a mutual temporal shift. Superimposition then results in a finer resolution. Here, too, a longer measurement under the same measurement conditions is required, which is present in a level measurement according to EP 2 199 761 A1, but not in other applications of time-of-flight measurement.

[0010] Another well-known solution involves multiplying the received signal across multiple delay channels, each adding a time offset of a different fraction of the sampling frequency. Subsequent combination improves the combined effective time resolution in proportion to the number of channels. However, this comes at the cost of parallel sampling of the channels. Therefore, an analog-to-digital converter slower than the effective sampling frequency is sufficient, but it must be multi-channel capable. Examples of this approach can be found in DE 10 2004 022 911 B4, EP 1 59 6221 B1, or EP 2 088 453 B1.

[0011] US 2002 / 0126270 A1 discloses a laser range detector that measures the propagation time of a pulse. To increase resolution, various delay paths are passed through before scanning. In one embodiment, the delay lines are combined via an N:1 element.

[0012] It is therefore an object of the invention to further improve the determination of a running time.

[0013] This problem is solved by a sensor and a method for determining a time of flight according to claim 1 and 10 respectively. A pulse-based time of flight method (dToF, direct time of flight) is used in which a signal pulse is emitted and received again. Measurement is usually carried out using a scanning principle in which the transmitter and receiver are at the same location, i.e. a remitted pulse is received. However, a time of flight measurement with spatially separated transmitter and receiver is also conceivable, for example for material determination. The signal pulse is the pulse on the measuring path, while the transmitted pulse and received pulse are associated electronic, internal pulses that trigger the signal pulse or are generated from it. A time offset between the respective electronic pulse and the signal pulse can be compensated for by reference measurement or calibration.

[0014] The received pulse is digitized in an analog-to-digital converter before the actual evaluation. The received pulse is fed to the analog-to-digital converter multiple times, each time with a different delay. This is achieved via at least one delay path with a delay element that causes a well-defined signal delay (delay), for example with a delay line. In this way, the analog-to-digital converter delivers a first sample and at least one second sample with a time offset corresponding to the delay of one delay path. With the help of additional delay paths or by a different number of multiple passes through a delay path, third, fourth, and further samples with different time offsets are possible. The samples differ from one another in amplitude and / or temporal position.

[0015] A control and evaluation unit uses at least two samples to determine the travel time of the signal pulse, which in turn can be converted into a distance using the speed of light—half the speed of light in a scanning principle due to the signal pulse's forward and return path. If the signal pulse passes through a medium other than air, the speed of light changes accordingly, and this effect can be used conversely for material detection with a known signal path.

[0016] The invention is based on the basic idea of ​​serially using the analog-to-digital converter multiple times to repeatedly sample the received pulse at different amplitudes and / or timings, thus obtaining additional information about the received pulse. Depending on its state or situation, the switching element forwards the received pulse directly, i.e., via no delay path, or with a specific number of loops, repeatedly over the same delay path, so that its delay is multiplied accordingly, or optionally via a specific delay path with its respective delay. It is irrelevant whether a zero delay in direct forwarding is considered a delay or not, especially since every line inevitably implies a certain delay. It is always possible to define an arbitrary zero point as a reference through calibration or a common time offset.

[0017] According to the invention, the analog-to-digital converter is therefore operated in serial multiple sampling. A single-channel analog-to-digital converter is sufficient for this purpose; a multi-channel analog-to-digital converter for parallel sampling known from the prior art is not required. It is conceivable to operate multiple channels of an analog-to-digital converter according to the invention; in this case, the effects of parallel sampling and the invention are combined. According to the invention, the received signal is only sampled intermittently, namely in the range of a received pulse. In between, measurement gaps arise that are required for serial multiple sampling. However, in a pulse-based time-of-flight measurement, nothing of interest occurs in these measurement gaps until a subsequent received pulse is to be evaluated after the next measurement repetition or, for example, in the next angular step of a laser scanner.

[0018] The invention has the advantage that high temporal and / or amplitude resolution is achieved with simple hardware. A moderately powerful analog-to-digital converter, which itself only offers a comparatively slow original sampling frequency and / or bit resolution, is sufficient. This has a positive impact on costs, energy consumption, and waste heat. Measurement times during which no relevant measurement event occurs anyway—namely, no received pulse is registered—are advantageously utilized to increase measurement accuracy. This also results in less data being generated that is not required for the actual analysis, thus simplifying further analysis.

[0019] The control and evaluation unit is preferably designed to sample the received pulse several times in succession using the analog-to-digital converter over a sampling period that is at least one pulse duration of the signal pulse. The pulse duration of the signal pulse corresponds to the period of the relevant measurement information. The sampling period or a respective sampling time window should preferably comprise at least this pulse duration or, more preferably, be equal to the pulse duration. If it is to be expected that the received pulse will differ in width from that of the signal pulse, for example due to filters or other effects, the sampling period can be adjusted accordingly. To enable serial sampling, the respective delays of two consecutive repeated samples are preferably at least the sampling period.Other useful time definitions concern the measurement duration between two transmitted pulses, which is in particular the inverse of a measurement repetition or pulse repetition frequency, as well as the sampling period of the analog-to-digital converter, ie the inverse of the sampling frequency, for example in the order of a few hundred MHz or a few GHz.

[0020] At least one delay path preferably delays the received pulse by a multiple of a sampling period of the analog-to-digital converter. Depending on the embodiment, this applies to the single delay path, one or some of several delay paths, or all delay paths. Preferably, the condition of the preceding paragraph is additionally set that the delay is greater than or equal to the acquisition duration. When delayed by a multiple of the sampling period, the delayed received pulse to be sampled lies, so to speak, on the time or sampling grid of the analog-to-digital converter. As a result, in particular, the same point or points of the received pulse are sampled multiple times. Preferably, the serial multiple sampling takes place at different amplitudes.This is achieved, for example, by an attenuation or amplification element in the delay path, particularly with a dual function attenuator, and / or by changing the sampling threshold(s) of the analog-to-digital converter during serial repetitions of the sampling. This type of serial multiple sampling effectively increases the amplitude resolution or bit depth of the analog-to-digital converter.

[0021] At least one delay path preferably delays the received pulse by a fraction of a sampling period of the analog-to-digital converter. Again, depending on the embodiment, this applies to the single delay path, one or some of several delay paths, or all delay paths. Preferably, the additional condition is that the delay is greater than or equal to the acquisition duration, with a corresponding number of whole sampling periods being added to the fraction for the respective delay. When delayed by a fraction of the sampling period, the delayed received pulse to be sampled lies, so to speak, between the time or sampling grid of the analog-to-digital converter. Consequently, with the repetition of the sampling, other points of the received pulse are sampled, so that the time resolution or effective sampling rate increases during combined evaluation of the samples. The embodiments of the preceding paragraph and this paragraph can be combined.Then at least one sample is taken with a delay of a multiple of the sampling period at a changed amplitude and at least one sample is taken with a delay of a fraction of the sampling period, and both the amplitude and time resolution are improved.

[0022] The switching element is preferably designed as a multiplex unit for switching between different delay paths. Depending on the state of the multiplex unit, the received pulse from one of the delay paths is digitized in the analog-to-digital converter, so that the delay paths can be processed sequentially in a multiplexing process. One state of the multiplex unit can expressly be the transmission of the received pulse for direct sampling without delay, which in this context can be understood as a delay path with zero delay. Preferably, a splitter is arranged upstream of the multiplex unit and the at least one delay path, which multiplies the received signal from the receiver onto the at least one delay path or a direct transmission.

[0023] The delay path forms a loop that feeds the delayed received pulse back into the delay path via the switching element. In this embodiment, a single delay path is preferably sufficient. The switching element initially forwards the received pulse directly and then, after passing through the loop once or multiple times, forwards it to the analog-to-digital converter. The different delays arise from the number of loops passed through. A delay corresponding to a multiple of the sampling period or one corresponding to a fraction of the sampling period is conceivable, furthermore with the preferred additional condition already mentioned above that the delay is at least as long as the acquisition duration. An improvement in the amplitude resolution or the time resolution is therefore achievable.In the first case, an attenuation or amplification in the loop is preferably provided, which accumulates with the number of loops in the same way as the delay.

[0024] A filter element for broadening the received pulse, in particular a low-pass filter, is preferably arranged between the switching element and the analog-to-digital converter. The filter element is particularly preferably arranged in a direct connection or a first delay path, or can be connected there, so that the first sample samples a received pulse that has been broadened or time-extended by the filter element. This guarantees, in particular, that the received pulse is detected at all or with a higher number of sampling points, even with a low time resolution of the analog-to-digital converter. However, the shape of the original received pulse is lost unless it is calculated back from the filter's transfer function. Serial multiple sampling still allows such information about the received pulse to be detected.

[0025] The control and evaluation unit is preferably designed to sample a receiver output signal without delay using the analog-to-digital converter until a received pulse is detected, and to switch the switching element after a received pulse. This creates a kind of idle state in which the receiver output signal is primarily monitored to determine whether a received pulse is present at all. Serial multiple sampling only occurs if this is the case. In other words, a detected received pulse can be considered the trigger for serial multiple sampling.

[0026] The received pulse is detected, for example, in the digitized output signal based on a noise or pulse threshold. The digitized output signal without a received pulse can be discarded because it contains no relevant measurement information, except for additional functions such as measuring ambient light. Preferably, the digitized output signal is stored in a rolling memory for a number of sampling points corresponding to the acquisition duration. In the case of a received pulse, its first sample is then available for further evaluation after completion by serial multiple evaluation, provided the further samples are stored in a different memory area. In a particularly preferred embodiment, the output signal sampled in the idle state passes through the filter element of the preceding paragraph to broaden any existing received pulse.

[0027] The control and evaluation unit is preferably designed to continue the serial multiple sampling until the next received pulse is expected or detected. This ensures that the measurement duration is used as fully as possible. It is conceivable to specify a maximum duration because it is not necessarily sensible to fill a longer measurement duration entirely with multiple repeated samplings. The next received pulse can be theoretically expected because it is known when it could arrive at the earliest, for example, for a predetermined measurement duration up to a maximum propagation time or range or a predetermined pulse repetition frequency. In particular, the propagation time can be quickly and roughly determined; more time is then available for a nearby object than for a distant object. Instead of a theoretical expectation, a new received pulse can be registered.Then, alternatively, the earlier received pulse is further evaluated and the new received pulse is ignored or vice versa, or an attempt can be made to determine a propagation time using the samples already recorded.

[0028] The sensor is preferably designed as an optoelectronic sensor, with the transmitter being a light transmitter, the receiver being a light receiver, and the signal pulse being a light pulse. The distance is thus measured optically using a time-of-flight method. Such an optoelectronic sensor can also be understood as a distance-measuring module from which various sensors such as light sensors, light grids, or laser scanners can be constructed. As an alternative to an optoelectronic sensor, the sensor can be designed as a radar, TDR, and especially a fill-level or ultrasonic sensor.

[0029] The method according to the invention can be further developed in a similar manner and thereby exhibits similar advantages. Such advantageous features are described by way of example, but not exhaustively, in the subclaims following the independent claims.

[0030] The invention will be explained in more detail below with regard to further features and advantages, using exemplary embodiments and with reference to the accompanying drawings. The figures of the drawing show: Fig. 1 shows a schematic representation of a sensor for distance measurement using a time-of-flight method; Fig. 2 shows an exemplary received pulse with the result of serial multiple sampling according to an embodiment of the invention; Fig. 3 shows a schematic representation of an embodiment of serial multiple sampling in several paths using multiplexing; Fig. 4 shows a schematic representation similar to Figure 3now with an additional filter for a first channel; Fig. 5 shows exemplary signals and samples in different channels of a serial multiple sampling; Fig. 6 shows a schematic representation of an embodiment of a serial multiple sampling in a loop; and Fig. 7 shows a schematic representation similar to Figure 6 now with additional filter for a first channel.

[0031] Figure 1shows a schematic representation of a sensor 10 measuring a light travel time in one embodiment as a one-dimensional optoelectronic distance sensor. A light transmitter 12, for example, an LED or a laser light source, emits a light pulse 14 into a monitoring area 16. The trigger for the light pulse 14 is an electronic transmission pulse (not shown). If the light pulse 14 strikes an object 18, a portion of the light pulse 14 is remitted or reflected and returns as a remitted light pulse 20 to a light receiver 22. There, an electronic reception pulse (also not shown) is generated. In a practical embodiment, the sensor 10 has further elements, in particular transmission and reception optics and connections, which are not discussed here for the sake of simplicity.

[0032] The reception pulse generated by the light receiver 22 is first digitized in a serial scanning unit 24, which is shown here only schematically as a block and which will be described later with reference to the Figures 3 to 7will be explained in more detail. The multiple serial samples of the received pulse are fed to a control and evaluation unit 26, for example, a microprocessor or at least one other digital processing module, which also controls the light transmitter 12 with the transmitted pulse and which may be responsible for other sensor functions. From the information about the received pulse, the control and evaluation unit 26 determines the time of flight (ToF) between the transmitted pulse and the received pulse. Constant components of the measured time of flight due to internal signal propagation times and the like can be compensated for by a reference measurement or calibration, also taking into account the current temperature. The possibly adjusted time of flight corresponds to the distance of the object 18, whereby the scaling factor is half the speed of light. The halving takes into account the outward and return paths of the light pulse 14, 20.

[0033] The propagation time can be determined using a single-pulse method or a pulse averaging method, which is known per se, for example, from the aforementioned EP 1 972 961 B1. A plurality of light pulses 14 are then emitted successively at intervals of a measurement duration of, for example, 1 µs, the measurement information from the corresponding received pulses is accumulated, and the propagation time is determined by statistical evaluation.

[0034] The Figure 1The one-dimensional optoelectronic distance sensor shown is only one embodiment of a sensor 10 according to the invention. For example, the sensor 10 can be coaxially constructed, be a multi-dimensional system, such as a laser scanner or a 3D time-of-flight camera, or operate with entirely different electromagnetic signals, such as a TDR level sensor or radar. In principle, non-electromagnetic signals such as ultrasound are also conceivable, in which case the speed of sound would be used instead of the speed of light, thus relativizing the problem of extremely fast time measurement.

[0035] The propagation time can be measured more accurately the higher the resolution of the received pulse is sampled. However, a high-performance analog-to-digital converter should not be used, i.e., preferably, a sampling rate in the higher GHz range, a high amplitude resolution of four, eight, sixteen, or even more bits, and parallel capability for simultaneous sampling of multiple (partial) signals are not required. Rather, according to the invention, the received pulse is sampled several times in succession at different amplitudes and / or temporal positions using the serial sampling unit 24, and the information from these multiple samples is combined in the control and evaluation unit 26.

[0036] Figure 2illustrates this in advance for the case of improved time resolution. The precise origin of the simulation results shown will be discussed again at the end of this description. The motivation for serial multiple sampling is as follows: in a pulse-based time-of-flight measurement, the light receiver 22 only detects ambient light for most of the measurement period between two transmitted light pulses 14. The receiving system effectively remains unused during this time. In contrast, in the short period of time in which a received pulse is generated by the light receiver 22, as much information as possible should be detected. Therefore, according to the invention, the received pulse is duplicated and repeatedly sampled serially during the unused times, depending on the embodiment at different time offsets and / or amplitudes. The serial samples are then combined.

[0037] The Figure 2illustrates this for a different time offset. In the upper part of the Figure 2 The received pulse is shown with a subsequent portion of the measurement period without another received pulse, superimposed with a square wave signal that illustrates the sampling period. Without serial multiple sampling, only a single sampling point would be obtained for the received pulse, and even then only if the received pulse is positioned at a favorable time relative to the sampling period. In the lower part of the Figure 2 Delayed repetitions of the received pulse are shown in gray, and samples obtained from them by serial multiple sampling are shown in black. The received pulse is thus effectively captured with a large number of sampling points; the influence of the stretching on the reception time or the propagation time can be corrected by the control and evaluation unit 26 with knowledge of the delays.

[0038] Figure 3shows a schematic representation of an embodiment of the serial scanning unit 24. The serial scanning unit 24 can be implemented entirely or partially as an integrated circuit, in particular at least partially together with the control and evaluation unit 26. On the input side, the analog received signal 28 of the light receiver 22 is fed to a splitter 30, whereby pre-amplification is conceivable. A simple splitter 30 divides the signal into several lines of the same resistance, for example, a 50-ohm line that branches into two 100-ohm lines.

[0039] A matching network allows for equal resistances at the input and output, particularly 50 ohms. An alternative is an active splitter 30 ohms.

[0040] The splitter 30 branches to at least two channels. One of these channels is preferably a direct connection 32, and at least one further channel is provided in the form of a delay path 34 1..n. In the example of Figure 3there are n delay paths 34 1..n with n different delays, in particular 1...n times a base delay, wherein the delays are generated by a delay element (not shown). The required delay can be adjusted, for example, by the length of a delay element designed as a delay line. Such a delay line is preferably implemented in a printed circuit board of the serial sampling unit 24. However, dedicated delay lines made of ceramic or cable are also possible. The terms channel and delay path are sometimes used interchangeably below. The direct connection 32 can also be understood in this sense as a delay path with zero delay. In embodiments according to the invention, however, different channels are fundamentally created by repeatedly passing through a delay path designed as a loop.

[0041] The channels are combined in a multiplex unit 36, which selectively switches the signal from one of the channels to an analog-to-digital converter 38. The bit resolution of the analog-to-digital converter 38 is preferably small, for example, sampling with n=1, 2, 4, or 8 bits. The sampling frequency is also preferably moderate, for example, a few tens or hundreds of MHz or a few GHz, although the latter alone would still be far too low for fine-resolution distance measurement. Thanks to serial sampling, the analog-to-digital converter 38 only has to process one signal at a time and can therefore be a single-channel design. The samples are output to the control and evaluation unit 26, which can preferably control the multiplex unit 36.

[0042] During measuring operation of sensor 10, the analog received signal 28 is split by splitter 30 into a number of sub-signals corresponding to the number of channels 32, 34 1..n. At the start of a measurement, the first channel 32 is active, i.e., the multiplex unit 36 ​​switches the direct connection 32 through to the analog-to-digital converter 38, which thus samples the undelayed received signal 28. If the control and evaluation unit 26 detects a received pulse during this sampling, it initially continues to sample for a sampling period. This can also have an effect back into the past by storing the previous samples in a rolling memory with a capacity for samples of the sampling period. This allows samples of the received pulse to be recorded in which the received pulse may not have yet been recognized as such. The sampling period is preferably as long as the expected duration of the received pulse or slightly longer.Since the width of the received pulse depends on the remission and the distance of the object 18, a selection tailored to the application is preferably made. Incidentally, these dependencies also apply to the pulse height; the control and evaluation unit 26 can take this into account in its evaluation.

[0043] After the acquisition period, the control and evaluation unit 26 switches the multiplexing unit 36 ​​to the next channel, in this case the first delay path 34 1 . The received pulse is thus sampled once more for the acquisition period with the delay of the first delay path 34 1 . Through further multiplexing, this process is continued analogously for the other delay paths 34 2 ..n. The received pulse is thus sampled n+1 times in succession. The system then returns to its idle state, in which the undelayed received signal 28 from the first channel of the direct connection 32 is sampled until another received pulse is detected, for example, during a measurement repetition or in the next angular step of a laser scanner.

[0044] The additional measurement information provided by serial multiple sampling can improve the amplitude and / or time resolution of the received pulse sampling.

[0045] If the amplitude resolution or bit count of the analog-to-digital converter 38 is to be increased, the delays in the delay paths 34 1..n are selected as a multiple of the sampling period and greater than or equal to the acquisition duration. The delays are located, so to speak, on the temporal grid of the sampling of the analog-to-digital converter 38. This means that the received pulse is captured multiple times at the same sampling points in the serial sampling. The repeated samplings occur at different amplitudes, so that additional measurement information about the amplitude of the received pulse is captured at the respective sampling points, thus effectively increasing the bit count through the superposition of the samples.

[0046] There are several ways to achieve these different amplitudes. In one embodiment, the control and evaluation unit 26 simultaneously changes a sampling threshold of the analog-to-digital converter 38 when the multiplex unit 36 ​​switches to a new channel. This has the advantage that the different amplitudes can be adjusted. Alternatively, different attenuation or amplification is provided in the channels. For this purpose, additional components can be arranged in the delay paths 34 1..n, for example resistors or voltage dividers, or the delay element itself has a targeted or already existing attenuation effect, which preferably correlates with the resulting delay. Another possibility is for the splitter 30 to split the signal into sub-signals with different amplitudes.

[0047] On the other hand, if the time resolution is to be increased, thus effectively achieving a higher sampling frequency, as is the case with the Figure 2 As explained, the delays in the delay paths 34 1..n represent a fraction of the sampling period. The fraction or non-integer multiple of the sampling period shifts the sampling points, so to speak, into the intervals of the temporal grid of the sampling of the analog-to-digital converter 38. The delay is made up of a fraction less than one and whole sampling periods, so that the overall delay is greater than or equal to the acquisition time. With each serial sampling, the sampling points are thus offset differently with respect to the received pulse. The fraction can preferably be the quotient of the sampling period and the number of channels; in this case, the channels sample the sampling period evenly one after the other. Alternatively, the fraction can be subtracted from a large number of whole sampling periods; in this case, the sampling across the channels shifts backwards rather than forwards relative to the received pulse.

[0048] The two options can be combined. Some of the delay paths 34 1..n are then designed for a higher amplitude resolution, as described, while others are designed for a higher time resolution.

[0049] Due to serial sampling, the receiving system is not receptive to further received pulses in the received signal 28 for n acquisition periods after a received pulse. This is not a problem as long as no relevant information is expected during this period. The control and evaluation unit 26 knows the time at which the next transmitted pulse will be generated and does not expect another received pulse beforehand. Alternatively, it is conceivable for the control and evaluation unit 26 to switch to the first channel at regular intervals via the multiplex unit 36 ​​and to sample the current received signal 28 that is now present via the direct connection 32 in order to reduce the dead time. As a further alternative, an additional measuring channel can be created for this monitoring. This allows unexpected further received pulses to be detected, for example in the case of multiple echoes, such as on a nearby pane of glass with a solid object behind it.Depending on the application, configuration, and situation, the control and evaluation unit 26 can then abort the fine quantization of the previous received pulse and focus on the new received pulse, or continue the original evaluation and ignore the new received pulse. Splitting the old and new received pulses with evaluation of the respective partial information is also conceivable.

[0050] It is also possible to combine the serial sampling according to the invention with parallel sampling, as described in the introduction. For example, a two-channel analog-to-digital converter 38 can double the effective sampling frequency again, whereby the additional measurement information can be distributed over increased amplitude or time resolution.

[0051] Figure 4 shows a further embodiment of the serial scanning unit 24. In contrast to the embodiment according to Figure 3In the first channel, the direct connection 32, a filter 40 for temporally stretching the received pulse, in particular a low-pass filter, is arranged. This makes it possible, in particular, to detect received pulses that are narrower than a sampling period and that only become longer than the sampling period due to the time stretching of the filter 40. The filter 40 changes the pulse shape. This could be computationally compensated in the control and evaluation unit 26 using known techniques, as discussed in the introduction. However, by switching to the following channels, the effective sampling rate is increased anyway, and the pulse shape of the non-stretched received pulse can be detected.

[0052] Figure 5 illustrates the sampling of a receive pulse in different channels for the example of a receive pulse that is narrower than a sampling period and for which the embodiment according to Figure 4is particularly suitable. In other words, the sampling frequency is so low that the receive pulse is not recorded at all in some channels, and at best only with a single sample point. With a higher sampling frequency or longer receive pulses, several sample points would be obtained in a single channel. However, a longer receive pulse is undesirable, as this impairs the resolution of the time-of-flight measurement, and the higher sampling frequency requires a higher-quality and therefore more expensive analog-to-digital converter 38.

[0053] The Figure 5shows from top to bottom a) the sampling times, b) the received signal 28, c) the signal in the first channel with time stretching by the filter 40, d) the signal in the second channel and e) the signal in the third channel. In the representation, delays of entire sampling periods have been omitted, so the serial samples are already appropriately superimposed. As can be seen in b), the sampling period in this example is twice as long as the pulse duration. Numerous received pulses would otherwise go unnoticed. However, the signal in the first channel stretched by the filter 40 according to c) is reliably detected. The serial sampling only provides a sampling point of the non-stretched received pulse in the second channel according to d). Nevertheless, the combined information from the second channel according to d) and the third channel according to e) correctly corresponds to a sample with twice the time resolution compared to the original sampling frequency.For more sampling points per receive pulse, additional channels and thus serial samples can be added.

[0054] In a numerical example, it is assumed that the receive pulse is 100 ps wide. As an illustrative example, the application only needs to detect whether a receive pulse is present. For this purpose, a 1-bit converter or comparator suffices as the analog-to-digital converter 38. This converter simply needs to sample at 5 GHz instead of 10 GHz, corresponding to the 100 ps. Filter 40 broadens the receive pulse to at least 200 ps. Suitable delays in the second and third channels are therefore 300 ps and 600 ps, ​​respectively.

[0055] If two more channels are added, the effective sampling period is halved again, so that a sampling frequency of 2.5 GHz is now sufficient. Filter 40 should therefore be widened to 400 ps, ​​and the delays become multiples of 450 ps. This means a delay of 1.8 ns for the fourth or last channel. Using FR4 circuit board material at ε r = 4, this corresponds to a delay line length of 27 cm and could still be easily implemented, for example, as a meander. With another six more channels, the sampling frequency can be reduced to 1 GHz. Filter 40 then widens to 1 ns, and the individual delay is 1.1 ns or 16 cm in FR4. For multiples of this delay in the later channels, long delay lines would be required, especially concentric delay lines. Therefore, an excessive number of channels, for example more than four, is no longer necessarily advantageous because the delay elements would become large.When combined with parallel sampling, this limit is higher. For example, a two-channel comparator can halve the required sampling frequency again.

[0056] Figure 6 shows an embodiment of the serial sampling unit 24 with multiple use of a delay path 34 in a loop. This, among other things, avoids the problem that longer delays are only possible with long delay lines. In this embodiment, no splitter 30 and no multiplexing unit 36 ​​are required. The received signal 28 is fed via a switch 42, which passes it in an initial or idle state to the analog-to-digital converter 38. At the same time, the received signal 28 is fed back to the switch 42 in a loop via the delay path 34.

[0057] Now, switch 42 is designed such that, in the idle state, while the received signal 28 is being passed through, it connects the delay path 34 with its internal resistance to ground. If a received pulse is detected, switch 42 is switched, so that the received pulse delayed in the loop with the delay path 34 is fed once again to both the analog-to-digital converter 38 and the loop with the delay path 34. The channels of the serial multiple sampling are formed by the number of times the loop with the delay path 34 is traversed.

[0058] Theoretically, any number of passes and thus channels is conceivable, at least as long as there is time until the next receive pulse. In practice, however, losses will occur, so that the receive pulse becomes weaker with each pass, limiting the number. This attenuation effect could be prevented by a loss-compensating amplification element in the loop. An analog-to-digital converter 38 with a high-impedance tap reduces the losses to those of the delay path 34 and the switch 42. High impedance can also be achieved by placing a suitable preamplifier upstream of the analog-to-digital converter 38.

[0059] Conversely, if an improvement in amplitude resolution is desired, the attenuation can be used to advantage to sample at different amplitudes. For example, to achieve a 2-bit conversion with four stages using a simple comparator threshold, three loop passes plus the first channel of the direct passthrough could be used. If the receive pulse is attenuated to 90%, for example, the first channel (direct connection) is sampled at an amplitude of 100%, the second channel (first pass through the loop) at 90%, the third channel (second pass through the loop) at 0.9 2< =81%, and the fourth channel at 0.9 3< =73%.

[0060] Figure 7 shows a further embodiment of the serial scanning unit 24, which is similar to the embodiment according to Figure 6 a filter 40 for time expansion in the first channel. A splitter 44, which is used to distinguish it from the one according to Figures 3 and 4receives a new reference symbol, but can be constructed in the same way, divides the received signal 28 into two partial signals. The first partial signal reaches the analog-to-digital converter 38 via the filter 40 and a switch 42a in a first channel. The second partial signal feeds the loop with the delay path 34, which is also connected to the switch 42a, via a further switch 42b. In the idle state, the switch 42a passes the time-expanded received signal 28 of the first channel to the analog-to-digital converter 38. If a time-expanded received pulse is then detected in the first channel, the switches 42a-b are switched over. Analogous to the explanations for the embodiment according to Figure 6 The other channels are now scanned by the loop with the delay path 34 repeatedly feeding the receive pulse to the analog-to-digital converter 38 with a selectable number of passes.

[0061] In all embodiments, the delay paths can be implemented in such a way that the delays are adjustable in order to cover a wider range of applications. Thus, depending on the time in which no further received pulse is expected, serial sampling can take place in more or fewer channels, in particular the loop in the embodiments according to Figure 6 or 7 more or less frequently. In a distance measurement system with regular measurement cycles, for example, more time is available for detecting a nearby object 18 than for detecting a distant object 18 before the next receive pulse arrives. Consequently, the echo signal for a nearby object 18 can be sampled with higher resolution, and the distance can thus be determined more precisely. This also makes it possible to separate objects located behind it more precisely, such as a leg in front of a reflective object or other multi-echo analysis.

[0062] After the explanations regarding the design of the serial scanning unit 24, Figure 2 is now even easier to understand. Here, a system according to Figure 6 simulated. The receive pulse is a 1 ns wide sine pulse sampled at 500 MHz. After 4 ns, switch 42 is toggled. The total delay of the loop is 4.1 ns. Thus, the input signal is delayed and appended by 100 ps with each pass relative to the sampling rate, thus effectively sampling the receive pulse backward in 100 ps steps.

Claims

1. A sensor (10) for determining a time of flight, in particular of a distance in accordance with a pulse-based time of flight process, that has a transmitter (12) for transmitting a signal pulse (14) triggered by a transmitted pulse (14), a receiver (22) for generating a received pulse from the signal pulse (20) received again, an analog to digital converter (38) for digitizing the received pulse, at least one delay path (34) having a delay element to supply the received pulse to the analog to digital converter (38) at least twice with a respectively different delay time and to generate a first scan and at least one second scan of the received pulse and thus a multiple scan of the received pulse at different amplitudes and / or temporal positions, and having a control and evaluation unit (26) that is configured to determine a time of flight by means of the first scan and the second scan, characterized in that a switchover element (36, 42) is arranged between the receiver (22) and the analog to digital converter (38) that selectively does not conduct the received pulse over a delay path (32) and differently frequently over the at least one delay path (34), and optionally over different delay paths (34 1..n) and thus supplies the received pulse to the analog to digital converter (38) for a serial multiple scan sequentially with different delays, wherein at least one delay path (34) forms a loop that feeds the delayed received pulse into the at least one delay path (34) again via the switchover element (42).

2. A sensor (10) in accordance with claim 1, wherein the control and evaluation unit (26) is configured to scan the received pulse by means of the analog to digital converter (38) multiple times sequentially over a detection period that amounts to at least a pulse duration of the signal pulse.

3. A sensor (10) in accordance with claim 1 or claim 2, wherein at least one delay path (34) delays the received pulse by a multiple of a scan period of the analog to digital converter (38).

4. A sensor (10) in accordance with any one of the preceding claims, wherein at least one delay path (34) delays the received pulse by a fraction of a scan period of the analog to digital converter (38).

5. A sensor (10) in accordance with any one of the preceding claims, wherein the switchover element is configured as a multiplex unit (36) for switching between different delay paths (341..n).

6. A sensor (10) in accordance with any one of the preceding claims, wherein a filter element (40) is arranged between the switchover element (36, 42) and the analog to digital converter (38) to widen the received pulse, in particular a low pass filter.

7. A sensor (10) in accordance with any one of the preceding claims, wherein the control and evaluation unit (26) is configured to scan an output signal of the receiver (22) with the analog to digital converter (38) for so long without delay (32) until a received pulse has been recognized and to switch over the switchover element (36, 42) after a received pulse.

8. A sensor (10) in accordance with any one of the preceding claims, wherein the control and evaluation unit (26) is configured to continue the serial multiple scan until a next received pulse is expected or is detected.

9. A sensor (10) in accordance with any one of the preceding claims, that is configured as an optoelectronic sensor, wherein the transmitter (12) is a light transmitter, the receiver (22) is a light receiver, and the signal pulse (14, 20) is a light pulse.

10. A method of determining a time of flight, in particular a distance, in accordance with a pulse-based time of flight process, in which a signal pulse (14) triggered by a transmitted pulse is transmitted and a received pulse is generated form the signal pulse (20) received again, wherein the received pulse is supplied for digitizing to an analog to digital converter (38) by means of at least one delay path (34) having a delay element having a respectively different delay path to generate a first scan and at least one second scan of the received pulse and thus a multiple scan of the received pulse at different amplitudes and / or temporal locations; and wherein a time of flight is determined by means of the first scan and the second scan, characterized in that the received pulse is selectively not conducted over a delay path (32) and differently frequently over the at least one delay path (34) and optionally over different delay paths ((341..n) by means of a switchover element (36, 42) and the analog to digital converter (38) thus scans the respectively differently delayed received pulse multiple times sequentially in a serial multiple scan, wherein at least one delay path (34) forms a loop that feeds the delayed received pulse into the at least one delay path (34) again via the switchover element (42).