Ultrasonic sensor with adaptive data compression

Abstract

The present invention provides ultrasonic sensors, sensor controllers, and sensing methods that improve critical aspects of measurement signal fidelity while retaining the main advantages of data compression. [Solution] An ultrasonic sensor, a sensor controller, and a sensing method may employ adaptive data compression. One exemplary sensor includes a piezoelectric transducer and a sensor controller. The sensor controller includes a receiver coupled to the ultrasonic transducer to acquire a received signal having one or more reflections of an acoustic burst within a measurement interval associated with the acoustic burst; a correlator configured to generate an output signal having a peak for each of the one or more reflections; and a compressor configured to determine a digital representation of the output signal for communication over a bus using at least one coding parameter that changes within the measurement interval.

Description

Technical Field

[0001] The present disclosure generally relates to data compression of ultrasonic sensor measurements to improve measurement signal fidelity within existing bus bandwidth constraints.

Background Art

[0002] Modern vehicles are equipped with a surprising number and variety of sensors. For example, a vehicle currently typically has an array of ultrasonic sensors to monitor the distance between the vehicle and any nearby person, pet, vehicle, or obstacle. Due to environmental noise and safety concerns, each sensor may be required to provide dozens of measurements per second, and each measurement requires analysis of hundreds of multi-bit signal samples for each channel. When the analysis of multiple such sensors is performed by a so-called fusion processor that collects various sensor data streams to enable inter-sensor measurements and combined sensor measurements to improve sensing accuracy and reliability, the required communication bandwidth becomes somewhat difficult to provide in a cost-effective manner while conforming to industry-standard limits on electromagnetic radiation.

[0003] One proposed solution is to employ irreversible data compression to reduce the number of bits carried over the communication bus. Such compression involves some loss of fidelity of the measurement signal, which can adversely affect the obstacle detection accuracy of the fusion processor. However, without such compression, existing automotive networks would need to be completely redesigned with the associated impacts related to their cost and performance.

Summary of the Invention

[0004] Accordingly, ultrasonic sensors, sensor controllers, and sensing methods are disclosed that employ adaptive data compression to improve important aspects of measured signal fidelity while retaining the main advantages of data compression. One exemplary sensor includes a piezoelectric transducer and a sensor controller. The sensor controller includes a receiver coupled to the ultrasonic transducer to acquire a received signal having one or more reflections of an acoustic burst within a measurement interval associated with the acoustic burst; a correlator configured to produce an output signal having a peak for each of the one or more reflections; and a compressor configured to determine a digital representation of the output signal for communication over a bus using at least one coding parameter that changes within the measurement interval.

[0005] One exemplary method includes acquiring a received signal having one or more reflections of an acoustic burst within a measurement interval associated with the acoustic burst, applying a correlation filter to generate an output signal having a peak for each of the one or more reflections, and determining a digital representation of the output signal using at least one coding parameter that changes within the measurement interval.

[0006] Each of the embodiments described above can be adopted individually or in combination and may include one or more of the following features in any preferred combination: 1. At least one coding parameter is at least one of the following: signal decimation rate, bit resolution, quantization threshold, and signal type. 2. The signal type can be selected from a set including at least the magnitude of the output signal and the output signal value expressed with respect to common-mode and quadrature components, or magnitude and phase. 3. At least one coding parameter changes based on elapsed time within the measurement interval. 4. At least one coding parameter is the signal decimation rate. 5. The signal decimation rate increases during the measurement interval. 6. At least one coding parameter is the bit resolution. 7. The bit resolution decreases during the measurement interval. 8. At least one coding parameter is the signal type. 9. The signal type is initially the output signal value and switches to the magnitude of the output signal during the measurement interval. 10. The compressor identifies the corresponding output signal segment for each peak. 11. At least one coding parameter changes based on the content of the output signal segment. 12. The coding parameter is a higher signal decimation rate by default and changes to a lower signal decimation rate when the output signal segment contains a phase shift or interpolation error exceeding a predetermined threshold. 13. The coding parameter is the signal type and is the magnitude of the output signal by default and changes to the output signal value when the output signal segment contains a phase shift or interpolation error exceeding a predetermined threshold. 14. The coding parameter is a lower bit resolution by default and changes to a higher bit resolution when the output signal segment contains a limiting signal below the minimum quantization threshold. 15. The coding parameter is the standard minimum quantization threshold by default and changes to the noise threshold when the output signal segment contains a limiting signal below the standard minimum quantization threshold. 16. The measurement interval includes at least a noise monitoring period and an echo detection period. 17. At least one coding parameter is a signal decimation rate that switches between 1:1 during the noise monitoring period and at least 2:1 for at least a portion of the echo detection period. [Brief explanation of the drawing]

[0007] [Figure 1] This is an overhead view of an example vehicle equipped with driver assistance sensors. [Figure 2] This is a block diagram of an example driver assistance system. [Figure 3] This is a circuit diagram of an example ultrasonic sensor. [Figure 4] This is a block diagram of an example ultrasonic sensor. [Figure 5] This graph compares signals useful for echo detection. [Figure 6] This is a block diagram for generating the output magnitude signal shown in Figure 5. [Figure 7A] This is a graph showing the magnitude of correlation for examples of overlapping echoes. [Figure 7B] This is a graph of an example curve interpolated from a decimated signal. [Figure 7C] This is a graph of an example curve interpolated from an adaptively compressed data stream. [Figure 8A] This is a graph showing the magnitude of the correlation with the limit signal. [Figure 8B] This is a graph of an example curve interpolated from a decimated signal. [Figure 8C] This is a graph of an example curve interpolated from an adaptively compressed data stream. [Figure 8D] This is a graph of an example curve interpolated from an alternative adaptively compressed data stream. [Figure 9] This is a flowchart illustrating an example of a sensing method. [Figure 10] This is a diagram illustrating an example of a digital representation of a compressed output signal segment. [Modes for carrying out the invention]

[0008] The drawings and the following description are not intended to limit this disclosure, but rather to provide a basis for those skilled in the art to understand all the variations, equivalents, and substitutions contained within the language of the claims.

[0009] As an illustrative context, Figure 1 shows a vehicle 102 equipped with sets of ultrasonic sensors 1-6 on the rear and a similar set on the front. An ultrasonic sensor is a transceiver, meaning each sensor can transmit and receive bursts of ultrasound. The emitted burst propagates outward from the vehicle until it encounters an object or some other form of acoustic impedance mismatch and is reflected from there. The reflected burst returns to the vehicle as an "echo" of the emitted burst. The time from the emission of the burst to the reception of the echo indicates the distance to the point of reflection. The number and configuration of sensors in a sensor array can vary.

[0010] In the illustrated array, sensor 1 is the left-side sensor, sensor 2 is the left-center sensor, sensor 3 is the left-center sensor, sensor 4 is the right-center sensor, sensor 5 is the right-center sensor, and sensor 6 is the right-side sensor. From the front sensor array, arrow 104 represents the acoustic bursts emitted by sensors 1, 3, 4, and 6. The bursts are intended to be emitted simultaneously, with sensors 1 and 4 transmitting on the low channel and sensors 3 and 6 transmitting on the high channel. The low channel and high channel may represent non-overlapping frequency bands that fall within the frequency band detectable by the ultrasonic transducer.

[0011] Arrow 106 represents the reflection of an acoustic burst returning to the sensor array. Sensor 1 may be configured to detect the direct reflection of its transmitted acoustic burst. Sensor 2 may be configured to detect the indirect reflection of an acoustic burst transmitted by Sensor 3. Sensor 3 may be configured to detect the direct reflection of its transmitted acoustic burst and the indirect reflection of an acoustic burst transmitted by Sensor 4. Sensor 4 may be configured to detect the direct reflection of its acoustic burst and the indirect reflection of an acoustic burst from Sensor 3. Sensor 5 may be configured to detect the indirect reflection of an acoustic burst from Sensor 4. Sensor 6 may be configured to detect the direct reflection of its acoustic burst. Overall, each sensor array monitors eight propagation paths, and a fusion processor may potentially use the sensors for individual and coordinated (e.g., triangulation, multi-receiver) measurements to detect and measure the distance to objects in various detection zones.

[0012] Figure 2 is a block diagram of an exemplary autonomous driver assistance system. The exemplary system includes a set of interconnected ECUs (Electronic Control Units) 201-203 organized in a hierarchical bus topology. Of course, other topologies, including serial, parallel, star, and ring topologies, are also suitable and intended for use according to the principles disclosed herein. Although multiple ECUs are illustrated and described here, in some cases it may be preferable for a single ECU to perform various functions.

[0013] A first ECU 201, also referred to herein as a fusion processor, is coupled to a bus controller for two arrays of ultrasonic sensors 204 and 206 to detect and monitor objects around the vehicle. An automotive serial bus 205 may be used for the sensor arrays, such as those provided by DSI3, LIN, and CAN standards. To provide automatic parking or other advanced features, a second ECU 202 may be coupled to a set of actuators, such as a door lock actuator 210, a throttle actuator 212, a brake actuator 214, a steering actuator 216, and a turn signal actuator 218. A third ECU 203 can be coupled to a user-interactive interface 220 to accept user input and provide display of various measurements and system status. Using the interfaces, sensors, and actuators, ECUs 201-203 can provide automatic parking, assisted parking, lane change assistance, obstacle and blind spot detection, and other desirable features.

[0014] Here, with reference to Figure 3, which shows a three-terminal configuration having two terminals for power and one terminal for I / O, one potential sensor configuration is described. The intended communication and power supply techniques include, and are preferred, those provided in the DSI3, LIN, and CAN standards, and are intended for use in accordance with the principles disclosed herein. In addition to the two power terminals (Vbat and GND) shown in the embodiment of Figure 3, each of the exemplary ultrasonic sensors is connected to the bus controller by only a single input / output ("I / O" or "IO") line. When the I / O line is not actively driven low ("asserted") by the bus controller or sensor controller 302, the I / O line can be biased to the supply voltage by a pull-up resistor ("deasserted"). The communication protocol is designed so that at any given time, only one of the two controllers asserts the I / O line.

[0015] When in the recessive mode, the sensor controller 302 monitors the I / O lines for assertion by the bus controller, and when in the dominant mode, includes an I / O interface 303 that drives the state of the I / O lines. The bus controller communicates commands to the sensor by asserting the I / O lines to carry a command word or a frame start pulse. The sensor controller 302 operates according to firmware and parameters stored in the non-volatile memory 305, and includes core logic 304 that analyzes the received commands and performs appropriate operations including transmission of ultrasonic bursts and reception of reflected signals. To transmit ultrasonic bursts, the core logic 304 is coupled to a transmitter 306 that drives a set of transmission terminals on the sensor controller 302 using a suitably modulated local oscillator signal. The terminals of the transmitter can be coupled to the piezoelectric element PZ via the transformer M1. The windings of the transformer have an inherent inductance LP and a resistance RLP. The transformer M1 boosts the voltage from the sensor controller (e.g., 12 volts) to a level suitable for driving the piezoelectric element (e.g., tens of volts). A parallel resistor RP may be provided to attenuate the residual vibrations of the piezoelectric element.

[0016] As used herein, the term "piezoelectric transducer" includes not only the piezoelectric element but also auxiliary circuit elements for tuning, driving, and sensing the piezoelectric element. However, the use of the term "piezoelectric transducer" does not necessarily require the presence of any support circuit elements, since the piezoelectric element can be employed alone without such support elements.

[0017] In various implementations, chirp-modulated signals, such as linear frequency-modulated ("LFM") chirps, are used. A chirp is a pulse whose frequency changes during transmission. An up-chirp is a signal pulse whose frequency increases during transmission, and a down-chirp is a signal pulse whose frequency decreases during transmission. For clarity, the examples used herein consider linear increases or decreases, but in various intended alternative implementations, the increases or decreases are not linear. The echo of a chirp may be compressed in a correlator without introducing much or any correlation noise. Thus, peak detection of the echo becomes easier without reducing the temporal resolution. Furthermore, LFM chirps tolerate Doppler frequency shifts with little to no increase in correlation noise. LFM chirps can be used as a transmission pulse to measure the distance to an obstacle or object located within the sensing range of a sensor system.

[0018] In other implementations, an AM (amplitude modulation) signal, e.g., a shaped pulse of a fixed-frequency carrier wave, is used. The AM signaling mode may allow for the use of shorter bursts (e.g., on the order of 200 - 300 microseconds), shortening the transmission time and increasing the sensitivity to nearby obstacles. Other implementations may employ pulses with a modulated carrier wave, e.g., pulses modulated with binary phase shift keying (BPSK). For clarity, the term "burst" as used herein can refer to an AM (fixed frequency), BPSK (modulated), or chirp (swept frequency) pulse, which can be one of a series of bursts generated by driving a piezoelectric element or other ultrasonic transducer. A chirp-modulated pulse may have a longer duration than a typical AM pulse, e.g., in the range of 2 - 3 milliseconds, longer than 1 millisecond. Here, the burst length can be varied, with shorter bursts used to facilitate detection of nearby obstacles and longer bursts used to increase the burst energy (and echo energy) for more distant obstacles. Note that the burst length for detecting nearby obstacles may be half or a quarter of the burst length used for more distant obstacles. The sensor may be switched between modes for different detection distances.

[0019] It is considered particularly useful to systematically vary the characteristic frequency (e.g., the start frequency, or equivalently, the center frequency or end frequency) of a series of chirp-modulated pulses, although such frequency variations can also be applied to the carrier frequency of a series of AM pulses. The frequency variation can be expressed for each pulse as a frequency displacement from a nominal characteristic frequency (e.g., a nominal start frequency or nominal carrier frequency).

[0020] To transmit an acoustic burst, the core logic configures the transmitter to drive the output pin of an ultrasonic transducer coupled to a piezoelectric element PZ. A transformer M1 and / or a resonant tuning network may be provided for voltage amplification and control of the transducer's resonant frequency. The transmitter may receive a carrier frequency signal from an oscillator having, for example, a nominal frequency of 50 kHz. The transmitter may use the carrier frequency signal to generate a series of AM (amplitude modulation) or chirp pulses, each corresponding to an acoustic burst. An example of a chirp pulse may be a low-channel up-chirp, a pulse with a frequency swept upward from 7 kHz below the carrier frequency. Another example may be a high-channel up-chirp, where the chirp frequency is swept upward from the carrier frequency to 7 kHz above the carrier frequency. Alternatively, a down-chirp may be employed, where the frequency is swept linearly downward rather than upward.

[0021] To receive an acoustic signal, the core logic configures the ADC310 to digitize the electrically received signal from the piezoelectric element PZ. The digitized signal may be provided directly to the DSP for real-time processing or buffered and stored in memory for subsequent processing by the DSP. To reduce I / O bandwidth requirements, the DSP may implement data compression to reduce the number of bits required to represent the received signal data or the magnitude of the baseband signal. To further reduce bandwidth requirements, the DSP may optionally perform on-chip processing for peak detection and distance estimation. Various suitable processing techniques for detecting reflections of acoustic bursts are well known in the art, including the jointly owned U.S. Patent Application Publication No. 2024 / 0069192, “Motion-compensated distance sensing with concurrent up-chirp down-chirp waveforms,” which is incorporated herein by reference.

[0022] Since the received electrical signal is typically in the millivolt or microvolt range, a receiving amplifier 308 may be included for buffering, storing, and amplifying the signal from the receiving terminal. For further filtering and processing by the DSP, an analog or digital mixer may be included to downconvert the received signal to the baseband. In one implementation, the mixer is a common-mode / quadrant-phase (I / Q) digital mixer, which yields zero intermediate frequency (ZIF) IQ data as its output. (Although the term "ZIF" is used herein, the downconverted signal may actually be a low intermediate frequency or "near baseband" signal.) The DSP 304 applies a programmable method to monitor the piezoelectric transducer during burst transmission, detect any echoes, and measure their parameters such as time-of-flight (ToF), duration, and peak amplitude. Such a method may use threshold comparison, minimum interval, peak detection, zero crossing detection and counting, noise level measurement, and other customizable techniques tuned to improve reliability and accuracy. Alternatively, the DSP may perform an initial processing step and then transport the processed signal to a fusion processor for further analysis.

[0023] Figure 4 is a block diagram of an exemplary signal processing path that can be embodied as a firmware module, i.e., instructions retrieved from memory for execution by a DSP. Alternatively, the module may be implemented as an application-specific integrated circuit, a field-programmable gate array (FPGA), or some other form of programmable logic device (PLD).

[0024] The transmit control module 402 combines a waveform template from the waveform control module 404 with a carrier signal from the digital carrier generation module 406 to form a digital burst signal. In at least some intended implementations, the digital burst signal is a linear frequency modulation chirp lasting approximately 2.5 milliseconds, during which the frequency is swept upward from 7 kHz below the carrier frequency to 1 kHz below the carrier frequency (lower sideband up-chirp) or swept upward from 1 kHz above the carrier frequency to 7 kHz above the carrier frequency (upper sideband up-chirp). One or both of the up-chirps can be replaced by a down-chirp, in which the frequency is swept from a higher value to a lower value. The exact duration and frequency range may be customized, and some intended implementations may divide a useful transducer frequency range into more than two channels, any one of which may be selected for adoption by the transmit control module 402.

[0025] The driver 408 converts the digital burst signal into a drive signal for the piezoelectric transducer 410. The piezoelectric transducer 410 vibrates in response to the analog burst signal, thereby generating an acoustic burst signal that propagates outward from the transducer. The reflection of the acoustic burst vibrates the piezoelectric transducer, inducing a detectable analog received signal. The receiver 412 amplifies and digitizes the analog received signal to generate a digital received signal. The sampling rate is at least twice the expected highest frequency component in the signal, preferably higher, such as an integer multiple, e.g., 8x, which can enable efficient digital down-conversion. Based on the digital received signal, the gain controller 414 adjusts the gain of the driver 408 and / or the receiver 412 to optimize performance while preventing saturation of the receiver's analog-to-digital converter (ADC). The diagnostic module 416, alone or together with the reverberation monitor 418, analyzes the digital received signal to detect and diagnose any transducer fault conditions. Some failure conditions may be indicated by, for example, an excessively short reverberation period (which may result from a disconnected or defective transducer, suppressed vibration, etc.), while other failure conditions may be indicated by an excessively long reverberation period (which may result from a defective mounting, insufficient damping resistance, etc.). The diagnostic module 416 can detect and classify several such transducer failure conditions and store the appropriate fault code in an internal register or non-volatile memory 305. The reverberation monitor 418 detects the end of the transducer reverberation period and signals it. The broadband noise detector 419 monitors the noise level and operates outside of the measurement period to detect potential interference with the operation of the sensor array.

[0026] Mixer 420 combines the digital received signal with a digital carrier signal to downconvert the digital received signal to a zero intermediate frequency ("ZIF") representation with common-mode and quadrature signal components. Low-pass filter 422 filters the ZIF signal component to eliminate modulation products that could otherwise cause aliasing. Magnitude module 424 optionally combines the common-mode and quadrature signal components to obtain a magnitude signal representing the magnitude of the received signal. Decimation or downsampling module 426 reduces the sample rate of the low-pass filtered ZIF signal component to reduce the computational load on downstream components. In at least some intended implementations, downsampling module 426 reduces the rate to approximately 20 kHz.

[0027] The correlator 428 filters the downsampled ZIF signal components using a correlation filter having an impulse response that matches each channel's waveform template, for example, an up-chirp in the upper and lower sidebands. The correlation filter generates a channel correlation signal in which burst echoes are represented as peaks. In at least some implementations, the correlation filter output signal is converted from in-phase and quadrature component representations to magnitude and phase representations to facilitate downstream processing. The correlation filter can narrow the peaks of the correlation signal using a modified waveform template (e.g., an up-chirp combined with a Gaussian window function). In some cases, the correlation filter can vary its impulse response as a function of time elapsed since the end of the reverberation period, improving detection performance at short ranges while continuing to suppress channel crosstalk at medium or long ranges. As an example, the applied window function can be varied to reduce the bandwidth of the chirp waveform as a function of elapsed time. At short ranges, a larger bandwidth allows partial echoes to pass through. A partial echo is an echo whose initial portion arrives before the measurement start time, corresponding to the end of the transducer's reverberation. This change in the impulse response allows a correlation filter to provide a peak in the correlation signal of these partial echoes. As more time passes, the partial echoes gradually become full echoes, and the bandwidth of each correlation filter is narrowed to provide better separation between frequency channels. Interpolation can be used to determine the filter coefficients between the values ​​initially used for correlation with the partial echoes and the values ​​used for correlation with the full echoes.

[0028] However, it should be noted that any correlation signal peaks associated with a partial echo are attenuated relative to the peak representing the complete echo, not only because the partial echo is shorter, but also because their frequency components differ from those of the complete echo. The chirp attenuation control module 430 can scale the channel correlation signal based on the time elapsed since the start of measurement and apply channel-dependent gain to compensate for the associated frequency dependence of the partial echo and transducer response, thereby substantially improving short-range detection performance. Module 430 may also apply time-dependent gain to compensate for echo attenuation due to propagation to and from obstacles.

[0029] As further described below, the noise detection / suppression module 432 applies a nonlinear function to the attenuated and compensated channel correlation signal to suppress noise and amplify peaks representing echoes. In some intended implementations, the module's output signal 433 is fed to a compressor 434 to reduce the number of bits required to transmit the correlation signal to the fusion processor for further processing. Other intended implementations include an echo detection module that detects peaks in the channel correlation signal, determines the magnitude of the peaks, and calculates the time of flight associated with each peak (or equivalently, determines the distance). In particular, the time of flight calculation takes into account the delay caused by the correlation filtering operation. The echo detection module can store the magnitude and time of flight information of the echoes detected in each channel in memory 305. The sensor interface module 436 transmits the desired information to be transmitted to the bus controller.

[0030] In reality, the received and digitized response includes not only arbitrary reflections from the ranging signal emitted by the acoustic transducer, but also noise. Such noise originates from various potential sources. Figure 5 shows a thin solid line representing an illustrative correlation magnitude curve ("MAGN"). Also shown is a dashed line representing the average noise magnitude ("Noise") derived from the non-peak region of the correlation magnitude curve, as will be further explained below. The dashed line represents an illustrative threshold curve ("THR") derived as described in relation to Figure 6. Finally, the thick solid line represents one of the output signals 433 ("OUT") of the noise detection / suppression module 432. In the peak region (the region where the magnitude curve exceeds the threshold curve), this output signal follows the magnitude curve. Outside these peak regions, this output signal is zero.

[0031] Figure 6 provides additional details for an exemplary implementation of the noise detection / suppression module 432. A correlator 428 convolves the down-converted signal with the waveform of the selected channel. Attenuation control 430 applies time-dependent gain to compensate for the expected attenuation during the measurement period. A magnitude element 604 combines the common-mode and quadrature components of the signal to determine the magnitude of the signal. An inverse tangent module 606 may combine the common-mode and quadrature components to determine the phase of the correlated signal.

[0032] CFAR element 608 acts on the correlation magnitude signal according to the Constant False Alarm Rate (CFAR) algorithm to provide a CFAR Threshold (CT) signal. Various CFAR algorithms are described in the literature, including the jointly owned U.S. Patent Application No. 16 / 530,654 filed on August 2, 2019, entitled "Ultrasonic Sensor Having Edge-Based Echo Detection" by inventors M. Hustava and J. Kantor (citing U.S. Patent No. 5,793,326 "Hofele"). Variations of preferred CFAR algorithms include, for example, CASH-CFAR (Cell-Averaged Statistics Hofele CFAR) and Ordered Statistics-CFAR (OS-CFAR). Briefly, a CFAR algorithm performs statistical processing within a moving window to determine a threshold representing background "clutter." This processing works to exclude any strong peaks that are likely to represent valid echoes from the threshold determination. Variations of CFAR differ in the exact nature of the statistical processing, such as whether to use minimum-maximum sum, rank ordering, or averaging operations in combination with suitable weighting or scaling to enable proper distinction between valid echoes and background noise. Various parameters of the algorithm (e.g., block size, window size) can be adjusted to optimize the adaptability of the threshold. CFAR offset values ​​can be stored in memory and added to the algorithm-based threshold to provide further tuning of the CT signal.

[0033] The CFAR element 608 can operate in a symmetric or asymmetric window around the "current" sample of the correlation magnitude signal. Thus, the delay element 610 can be used to provide a suitable time offset between the "early" correlation magnitude signal supplied to the CFAR element 608 and the "current" correlation magnitude signal 612 supplied to other elements of the processing circuit. The comparator 614 compares the current correlation magnitude signal 612 with the CFAR threshold signal CT and asserts a selection signal for the multiplexer 616 to indicate when the correlation magnitude signal is above the threshold ("peak region") and deassers the selection signal to indicate when the correlation magnitude signal is below the threshold ("non-peak region").

[0034] The noise averaging block 620 receives a selection signal at the inverting enable ( / EN) input, also known as the disable input, which disables the operation of the noise averaging block 620 while the comparator output is asserted. In this way, the averaging block 620 operates on the non-peak region of the signal and ignores the peak region of the correlation magnitude signal when determining the average noise magnitude ("noise" in Figure 5). The noise averaging block 620 calculates the average of the non-peak correlation magnitude signal over a given portion or within a moving window. The averaging block may be configured, for example, to sum the signals in the non-peak region of the signal portion and divide it by the duration of the non-peak region of the signal portion. Although this application refers to the mean, it should be understood that the resulting mean may be any type of mean known to those skilled in the art, including the median, arithmetic mean (average), mode, geometric mean, and / or weighted mean, as well as the exponential moving mean.

[0035] Current correlation magnitude: For each peak region in the signal, the peak measurement element 618 determines the signal intensity by identifying the peak value (maximum value). The signal-to-noise ratio (SNR) block 622 receives each peak value from the peak measurement element 618 and calculates the SNR value of that peak using the corresponding noise mean value from the noise averaging block. Here, block 622 calculates SNR = 20log 10It should be noted that the definition is not limited to any arbitrary formula such as (signal / noise). In fact, given the hardware complexity typically associated with logarithmic calculations, it may be preferable to use a simple ratio or other calculation that is monotonically related to the definition in the region of interest. SNR block 622 generates a peak detection threshold that is inversely related to the SNR value. The peak detection threshold may be low in the region where the SNR is relatively high, and relatively high in the region where the SNR is low. SNR block 622 can employ a lookup table to convert the SNR value to the peak detection threshold.

[0036] The summing element 624 adds the peak detection threshold to the average noise level to determine the threshold curve THR in Figure 5. The comparator 626 compares the current correlation magnitude signal 612 with the threshold THR and asserts the selection signal only when the signal magnitude exceeds the threshold. In response to the selection signal, the multiplexer 628 selects the current correlation magnitude signal 612 when the selection signal is asserted and selects a zero signal to suppress the output signal OUT when the selection signal is deasserted. Another multiplexer 630 may also respond to the selection signal, outputting the correlation signal phase when the selection signal is asserted and suppressing the phase output when the selection signal is deasserted.

[0037] In the context described above, the operation of compressor 434 is then referred to. In one intended implementation, the compressor downsamples the correlation magnitude signal before scaling and quantizing it to a lower bit resolution. For example, it drops every other sample. The compressor can treat the correlation magnitude signal as a series of signal segments, each having a predetermined number of samples. A scale factor is determined for each segment and normalizes the peak value in that segment to 1 (unless it deviates). As an example, the scale factor may have a 5-bit representation. In some cases, the scale factor uses a nonlinear scale. The compressor then quantizes the scaled sample values ​​with the reduced bit resolution. In one particular example, the compressor may use a 3-bit resolution for each sample in a segment.

[0038] Given the scale factor and quantization sample for each segment, the fusion processor can rescale and interpolate the data to reconstruct the correlation magnitude signal with sufficient fidelity to detect and monitor reflectors within the measurement zone of the ultrasonic sensor array. However, with increasingly sophisticated processing, system designers desire the ability to better distinguish densely packed peaks and / or detect limiting signal energies around major peaks that can reveal additional information useful for distinguishing different types of objects, such as a person against a curb. Thus, the compressor 434 can appropriately adapt the coding parameters associated with selected segments of the correlation magnitude signal.

[0039] Figure 7A shows an example output signal before the compression process. (For convenience, we assume that the samples in the example output signal segment are scaled to have a peak value of 1.) Samples of the correlation magnitude signal are represented by white circles. Samples outside the peak region (i.e., below the noise threshold 704) are zeroed out. Within the peak region, sample 702 shows a dip that can distinguish between two densely packed peaks. This dip is typically associated with a substantial shift in signal phase, but can also be detected as a significant interpolation error. In Figure 7B, for example, black diamonds represent downsampled and quantized samples. The solid line is the linear interpolation between these compressed samples, and it should be noted that sample 702 deviates from the line beyond two quantization steps. (Linear interpolation is used for ease of explanation, but any preferred interpolation method, including, for example, cosine interpolation, cubic interpolation, and Hermitian interpolation, may be used.) The threshold for detecting excessive interpolation error is preferably at least half the quantization steps, but may be customized to optimize system performance.

[0040] If interpolation error is detected, or alternatively, if a phase shift exceeding a predetermined threshold (e.g., 30 degrees) is detected, the compressor 434 may reduce the downsampling rate, for example, from 2:1 to 1:1. Figure 7C shows a compressed representation with a reduced sampling rate. Note that the maximum interpolation error is reduced to about half of the quantization step. Alternatively, the compressor may maintain the downsampling rate but change the signal type, transmitting not only the magnitude of the correlation signal but also the magnitude and phase (or in-phase and quadrature phase components) of the correlation signal. The additional information is expected to substantially eliminate or at least reduce the interpolation error.

[0041] Figure 8A shows another example output signal before the compression process. Note the presence of a limiting signal, i.e., a sample 802 representing a signal that is above the noise threshold 804 but has a small magnitude after peak normalization of its output signal segment. When a sample has a magnitude smaller than the minimum quantization threshold, which is usually equal to half the quantization step, the quantization process suppresses the limiting signal, as shown in Figure 8B. If such a situation is detected, the compressor 434 may adapt the bit resolution used to quantize the samples in a given signal segment. For example, increasing the bit resolution from 3 bits to 4 bits. Figure 8C shows an example of a sample compressed using the increased bit resolution.

[0042] Alternatively, the compressor 434 may maintain bit resolution and instead adapt a minimum quantization threshold for the signal segment. For example, the threshold may be lowered from half a quantization step to a noise threshold, thereby quantizing any sample with a magnitude above the noise threshold to a non-zero value. Figure 8D shows an example of compressed samples obtained with this approach.

[0043] When such adaptation of coding parameters is performed, the compressor 434 may provide a field for indicating the current values ​​of a given parameter for each segment. However, the described example is expected to tend to occur in short measurement ranges, and in longer ranges, attenuation is expected to make such an effect rare or at least undetectable. Therefore, the compressor may be configured to adapt the coding parameters as a function of the elapsed time in the measurement interval following each acoustic burst. As an example, the compressor may employ one set of parameters (signal type, downsampling rate, bit resolution, minimum quantization threshold) for the first 12 milliseconds of the measurement interval, and then employ a second set of parameters for the remainder of the measurement interval. The signal type may progress from correlation magnitude and phase to correlation magnitude only, and the bit resolution may decrease from 4 bits to 3 bits. As another example, the downsampling rate may increase from 1:1 to 2:1 while the remaining parameters remain constant, when the minimum quantization threshold is adjusted from one-quarter of a quantization step to one-half of a quantization step. Alternatively, the downsampling rate and bit resolution may be adjusted while other parameters are maintained. Of course, any individual parameter can also be adjusted independently of the others.

[0044] In at least some intended implementations, a portion of the measurement interval is dedicated to noise monitoring. For example, during the last milliseconds of a measurement period immediately preceding a subsequent acoustic burst, the correlation magnitude signal can be analyzed by the fusion processor to estimate the ambient noise level. The noise detection / suppression module is disabled during this period, allowing the compressor 434 to operate on the "raw" correlation magnitude data. The encoding parameters employed by the compressor may be selected to minimize additional distortion, for example, by minimizing the downsampling rate (set to 1:1), maximizing the bit resolution (e.g., set to 4 bits), and / or by setting the signal type to transmit complex-valued correlation signals (in-phase and quadrature components, or magnitude and phase) rather than just magnitude.

[0045] Figure 9 is a flowchart of an exemplary sensing method employing adaptive compression. Various operations can be performed by the sensor controller described above. In block 902, the sensor controller acquires a received signal from the ultrasonic transducer. The receiving operation may occur immediately after the transmission of an acoustic burst by the same transducer or another transducer in the sensor array. The receiving operation may include digitization, downmixing, and filtering to acquire a baseband digital received signal. Downsampling may be performed optionally to reduce the processing load on subsequent components in the chain. In block 904, the controller filters the digital received signal using one or more correlation filters to acquire a correlation signal for each channel. The correlation filter output may be complex numerical, i.e., it has in-phase and quadrature-phase components, and therefore the controller can calculate the magnitude of the correlation signal in block 906. Optionally, the correlation phase signal can also be determined in this block.

[0046] In block 908, the controller analyzes the correlation magnitude signal to distinguish peak regions from non-peak regions and suppress noise in the latter. In block 910, the controller divides the correlation magnitude signal into segments in preparation for compression. In block 912, the default encoding parameters are set to their initial values. The controller then proceeds through the segments sequentially, which may be repeated through blocks 914-934 as it progresses.

[0047] In block 914, the controller determines whether the current segment is associated with a time-based change in the coding parameters, and if so, adjusts the parameter values ​​to the new defaults. For example, the default signal type may initially be the output signal value (representable in terms of magnitude and phase) and then switch to the magnitude of the output signal only after 13 milliseconds. As another example, the default bit resolution may initially be 4 bits and then switch to 3 bits after 9 milliseconds. In block 918, the controller determines whether the current segment corresponds to a noise monitoring interval. If so, the controller adjusts the signal type and / or decimation rate in block 920 before proceeding to block 930. Otherwise, in block 922, the controller determines whether the current segment has a phase shift exceeding a given threshold. Alternatively, the controller determines whether the current segment has excessive interpolation errors with the default coding parameter values. If so, in block 924, the controller switches or changes the coding parameter values ​​accordingly before proceeding to block 930. Otherwise, in block 926, the controller determines whether the current segment contains limiting signals that would be suppressed by the default coding parameters. If so, in block 928, the controller adjusts the bit resolution and / or minimum quantization threshold.

[0048] In block 930, the controller uses the selected coding parameter values ​​to form a compressed digital representation of the current segment, which is then transmitted to the fusion processor via the bus controller. In block 932, the controller determines if there are any other segments in the measurement period, and if so, in block 934, it resets the coding parameters to their default values ​​before returning to block 914 to process the next segment. Once the measurement period has been fully processed, the sensor controller returns to block 902.

[0049] Although the operations shown and described above are treated as sequential for illustrative purposes, in practice, the processes can be performed by multiple integrated circuit components operating simultaneously (and in some cases even inference-wise), allowing for parallel or non-sequential operation. With a full understanding of the above disclosures, many variations, equivalents, and alternatives of these and many others will become apparent to those skilled in the art. The following claims are intended to be construed as encompassing all such variations, equivalents, and alternatives, where applicable.

[0050] Figure 10 shows an example of a digital representation of an output signal segment that may be generated by the compressor 434. The illustrated digital representation is, for example, a word having 32 bits. The coding parameter 1002 may be indicated by a first bit indicating, for example, the signal type, decimation rate, bit resolution, or whether a first or second value is used for the minimum quantization threshold. This field may be omitted if the change in coding parameter value switches automatically based only on elapsed time. The illustrated representation includes a field for a scale value 1004 indicating the scale factor used to normalize the peak value to 1. This field may have, for example, 5 bits. The illustrated representation further includes fields for a predetermined number of output signal sample values ​​1006. These fields may each have, for example, 3 bits or 4 bits.

[0051] While dependent claims are written to refer to a single claim, as is the issue of claim drafting regulations in a particular country, it is observed that any combination of a dependent claim and any of its preceding claims is foreseen by the inventors and is considered to be included in the complete disclosure of this application. Furthermore, it should be understood that a dependent claim designated for one claim category also applies to another claim category, but is simply omitted to limit the total number of claims and any resulting claim fees payable.

[0052] While the context of driver assistance systems is used as an example herein, the concepts of this disclosure may be applied to any type of obstacle monitoring or distance measuring system, and may be particularly suitable for systems where reliability and rapid response are prioritized.

Claims

1. A received signal having one or more reflections of an acoustic burst is acquired within the measurement interval associated with the acoustic burst. Applying a correlation filter to generate an output signal having a peak for each of the one or more reflections, A method comprising determining a digital representation of the output signal using at least one coding parameter that changes within the measurement interval.

2. The at least one encoding parameter is one of the following: signal decimation rate, bit resolution, quantization threshold, and signal type. The signal type can be selected from a set that includes at least the magnitude of the output signal and the output signal value expressed with respect to the common-mode and quadrature components, or with respect to magnitude and phase. The method according to claim 1, wherein the at least one coding parameter changes based on the elapsed time within the measurement interval.

3. The method according to claim 1, wherein the determination includes identifying an output signal segment for each of the peaks, and the at least one coding parameter changes based on the content of the output signal segment.

4. A receiver connected to an ultrasonic transducer, which acquires a received signal having one or more reflections of an acoustic burst within a measurement interval associated with the acoustic burst, A correlator configured to generate an output signal having a peak for each of the one or more reflections, A sensor controller comprising: a compressor configured to determine a digital representation of the output signal for communication over a bus using at least one coding parameter that changes within the measurement interval.

5. The sensor controller according to claim 4, wherein the compressor identifies an output signal segment for each of the peaks, and the at least one coding parameter changes based on the content of the output signal segment.

Images