Ultrasonic sensor with echo amplification

Abstract

The present invention is titled "Ultrasonic Sensor with Echo Amplification". A sensor, a sensor controller, and a sensor control method can employ echo amplification techniques to improve threshold-based echo detection. In one illustrative embodiment, the sensor controller includes: a transmitter, a receiver, and processing circuitry coupled to the transmitter and the receiver. The transmitter drives a piezoelectric element to generate a burst of acoustic pulses. The receiver senses the response of the piezoelectric element to the echo of each burst of acoustic pulses. The processing circuitry is operable to apply an echo detection process to the response by: identifying intervals of the response that represent at least a portion of a potential echo; deriving a modified response from the response by selectively amplifying the response during the intervals; and using the modified response to detect the echo.

Description

Technical Field

[0001] The present disclosure relates to echo detection techniques for ultrasonic sensors, and more particularly to techniques for providing selective echo amplification. Background Art

[0002] Modern vehicles are typically equipped with an array of ultrasonic sensors to monitor the distance between the vehicle and any nearby people, pets, vehicles, or obstacles. Due to environmental "noise" and safety concerns, each of the sensors may be required to provide dozens of measurements per second while the vehicle is in motion. Existing echo detection techniques typically employ a variable threshold in an effort to account for the gradual attenuation of echo intensity over travel time. However, the noise level and attenuation can vary significantly in different environments, and even seemingly small differences (such as the presence or absence of a curb, or even the difference between a paved surface and a gravel surface) can significantly alter the characteristic reflection of a pole, post, or other thin-walled obstacle, such that such detection techniques may be prone to failure unless enhanced by computationally intensive adaptation of the detection template to the environment. Despite this shortcoming, such detection techniques may still be specified as product requirements. Summary of the Invention

[0003] Accordingly, various sensors, sensor controllers, and sensor control methods are disclosed herein that employ echo amplification techniques to improve threshold-based echo detection.

[0004] In one aspect of the present disclosure, a controller for a piezoelectric transducer is provided, the controller characterized by comprising: a transmitter for driving a piezoelectric element to generate acoustic bursts; a receiver for sensing the response of the piezoelectric element to the echo of each acoustic burst; and a processing circuit coupled to the transmitter and the receiver, the processing circuit operable to apply an echo detection process to the response, the process comprising: identifying intervals of the response that represent at least a portion of a potential echo; deriving a modified response from the response by selectively amplifying the response during the intervals; and using the modified response to detect an echo.

[0005] In one embodiment, the controller is characterized in that the identifying intervals comprises: comparing the response with an adaptive threshold to produce a comparison signal; and detecting an edge in the comparison signal.

[0006] In one embodiment, the controller is characterized in that the process further comprises deriving an adaptive threshold from the response signal using a constant false alarm rate (CFAR) method.

[0007] In one embodiment, the controller is characterized in that the identifying further comprises comparing an estimated derivative with a threshold.

[0008] In one embodiment, the controller is characterized in that the selective amplification includes multiplying by a scaling factor.

[0009] In one embodiment, the controller is characterized in that the selective amplification includes applying a non-linear function to the response during the interval.

[0010] In one embodiment, the controller is characterized in that the selective amplification includes reducing the response outside the interval.

[0011] In one embodiment, the controller is characterized in that the echo is detected by comparing the modified response with a predetermined threshold template.

[0012] In one embodiment, the controller is characterized in that the interval corresponds to the rising edge of the echo pulse.

[0013] In one embodiment, the controller is characterized in that the interval corresponds to the falling edge of the echo pulse. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 Top view of an exemplary vehicle equipped with a parking assistance sensor.

[0015] Figure 2 Block diagram of an exemplary parking assistance system.

[0016] Figure 3 Circuit schematic of an exemplary parking assistance sensor.

[0017] Figures 4A - 4B Graph of an exemplary signal for an echo scaling process.

[0018] Figure 5 Exemplary CFAR implementation.

[0019] Figure 6 Block diagram of an exemplary echo scaling process. DETAILED DESCRIPTION

[0020] It should be understood that the drawings and the following description do not limit the present disclosure, but on the contrary, provide a basis for those of ordinary skill in the art to understand all modifications, equivalents, and alternative forms falling within the scope of the claims.

[0021] As an illustrative usage context, Figure 1A vehicle 102 equipped with a set of ultrasonic parking assist sensors 104 is shown. The number and configuration of sensors in the sensor arrangement vary, and it is not uncommon to have six sensors on each bumper, with two additional sensors on each side serving as blind spot detectors. The vehicle can use this sensor arrangement to detect and measure the distance to objects in various detection zones in situations where the sensors may be used for individual measurements as well as collaborative measurements (e.g., triangulation, multi-receiver measurements).

[0022] The ultrasonic sensors are transceivers, meaning that each sensor can transmit and receive bursts of ultrasonic sound. The transmitted bursts propagate outward from the vehicle until they encounter an object or some other form of acoustic impedance mismatch and are reflected by it. The reflected burst returns to the vehicle as an "echo" of the transmitted burst. The time between the transmitted burst and the received echo indicates the distance to the point of reflection. In many systems, only one sensor transmits at a time, but all sensors can be configured to measure the resulting echoes. However, multiple simultaneous transmissions can be supported by using orthogonal waveforms or transmissions to non-overlapping detection zones.

[0023] Figure 2 An electronic control unit (ECU) 202 coupled to various ultrasonic sensors 204 as the center of a star topology is shown. Of course, other topologies, including serial, parallel, and hierarchical (tree) topologies, are also suitable and are contemplated to be used in accordance with the principles disclosed herein. To provide automatic parking assist, the ECU 202 can be further connected to a set of actuators, such as a turn signal actuator 206, a steering actuator 208, a brake actuator 210, and a throttle actuator 212. The ECU 202 can also be coupled to a user interactive interface 214 to accept user input and provide a display of various measurement results and system states. Using the interface, sensors, and actuators, the ECU 202 can provide automatic parking, assisted parking, lane change assist, obstacle and blind spot detection, and other desired features.

[0024] Now refer to Figure 3 Describe a possible sensor configuration. (Other communication and power technologies, such as those provided in the DSI3, LIN, and CAN standards, are also suitable and are expected to be used in accordance with the principles disclosed herein.) In addition to Figure 3In addition to the two power terminals (Vbat and GND) shown in the embodiment of, each exemplary ultrasonic sensor is connected to the ECU 202 only through a single input / output ("I / O" or "IO") line. When the I / O line is actively driven low ("enabled" state) by the ECU 202 or through the sensor controller 302, the I / O line can be biased to the supply voltage ("disabled" state) through a pull-up resistor. The communication protocol is designed such that only one of the two controllers (ECU 202 or sensor controller 302) enables the I / O line at any given time.

[0025] The sensor controller 302 includes an I / O interface 303 that monitors the enabling of the I / O line caused by the ECU 202 when placed in the recessive mode and drives the state of the I / O line when placed in the dominant mode. The ECU transmits commands to the sensor by enabling the I / O line, and different commands are represented by different lengths of enabling. The commands can include a "send and receive" command, a "receive only" command, and a "data mode" command.

[0026] The sensor controller 302 includes a core logic 304 that operates according to firmware and parameters stored in the non-volatile memory 305 to parse commands from the ECU and perform appropriate operations, including the transmission and reception of ultrasonic pulse trains. To transmit an ultrasonic pulse train, the core logic 304 is coupled to a transmitter 306 that drives a set of transmit terminals on the sensor controller 302 with a properly modulated local oscillator signal from a voltage-controlled oscillator 307. The transmitter terminals are coupled to the piezoelectric element PZ via a transformer M1. The transformer M1 steps up the voltage from the sensor controller (e.g., 12 volts) to a suitable level (e.g., dozens of volts) for driving the piezoelectric element. The piezoelectric element PZ has a resonant frequency tuned to a desired value (e.g., 48 kHz) using a parallel capacitor C3 and has a resonant quality factor (Q) tuned using a parallel resistor R1. An exemplary purpose of the tuning capacitor and the tuning resistor is to tune the parallel resonant frequency to be close to the series resonant frequency of the piezoelectric element.

[0027] As used herein, the term "piezoelectric transducer" includes not only the piezoelectric element but also the support circuit elements for tuning, driving, and sensing the piezoelectric element. In the exemplary embodiment, these support elements are the transformer M1, the tuning resistor and the tuning capacitor, and the DC isolation capacitor. Optionally, the output and input capacitances of the transmitter 306 and the amplifier 308 can also be included, respectively, as parasitic characteristics of the support circuit elements considered to be part of the transducer. However, the use of the term "piezoelectric transducer" does not necessarily require the presence of any support circuit elements, as the piezoelectric element can be used alone without such support elements.

[0028] A pair of DC isolation capacitors C1, C2 couple the piezoelectric element to a pair of receiving terminals of the sensor controller to prevent high voltages. Further protection is provided by an internal voltage clamp on the receiving terminals. Such protection may be desirable for the intervals during which the piezoelectric element is transmitting. Since the received echo signals are typically in the millivolt or microvolt range, a low-noise amplifier 308 amplifies the signals from the receiving terminals. An optional mixer 309 multiplies the amplified received signal with a local oscillator signal to down-convert the modulated signal to a baseband signal before digitizing and processing by the digital signal processor (DSP) 310 and the integrated analog-to-digital converter (ADC). Alternatively, the DSP 310 directly samples and processes the received signal.

[0029] The DSP 310 applies programmable methods to monitor the piezoelectric transducer during burst transmissions, and detect any echoes and measure their parameters such as time-of-flight. Such methods may employ constant false alarm rate (CFAR) threshold determination, derivative calculation, threshold comparison, echo scaling, minimum interval, noise level determination, and other customized techniques that are tailored to improve reliability and accuracy. The DSP 310 may further process the amplified received signal to analyze the characteristics of the transducer such as resonant frequency and quality factor, and may further detect the transducer fault state.

[0030] Commands received via the I / O lines trigger the core logic 304 to operate the transmitter and receiver and provide the measurement results to the ECU 202 via the I / O lines or other interface standards. In addition to the echo measurement results and transducer fault states that may be detected by the DSP 310, the core logic may also monitor other sensor conditions such as "under-voltage" or "over-voltage" of the supply voltage while transmitting an ultrasonic burst, thermal shutdown of the transmitter, hardware errors, incomplete power-on reset, etc. The core logic 304 can detect and classify multiple such transducer fault states and error conditions, and store appropriate fault codes in internal registers or non-volatile memory 305.

[0031] Figure 4A and Figure 4BSome exemplary signal diagrams are provided to assist in understanding the operation of the exemplary sensor implementations. The operation of the piezoelectric transducer is herein represented as a vibration signal VIBR that represents the mechanical oscillation of the piezoelectric element. (It should be noted that these signals are not shown to scale. For example, the transmitted pulse train 402 may be several orders of magnitude larger than the echoes 408, 409.) Electrically, the mechanical vibration of the piezoelectric element can be detected as a voltage or current. The mechanical vibration amplitude increases as the controller 302 drives the transducer (“drive phase” 404), and then decreases after the drive operation ends (“reverberation phase” 406). The controller 302 can employ active and / or passive damping to reduce the duration of the reverberation phase.

[0032] In Figure 3 the sensor implementation, the vibration is detected as a down-converted clamped amplifier type of secondary voltage via the amplifier 308. For illustrative purposes, Figures 4A - 4B the RX signal shown in

[0033] is the (low-pass filtered) envelope of the down-converted clamped amplifier voltage signal, but an amplified oscillation signal may also be employed. Figure 4A The sensor controller optionally measures the noise level during the pre-transmission period 403 to be able to detect transducer faults and / or an environment that is too noisy for accurate echo measurements. Actuation of the transducer for the transmitted pulse train typically saturates the RX signal, generating the transmit pulse 411. (In at least some embodiments, an internal voltage clamp on the receive terminal of the sensor controller prevents excessive voltage from reaching the amplifier 308.) The transmitted pulse train overwhelms the receiver and prevents any meaningful echo measurements from being obtained during the actuation interval. Note that the actuation interval (the time during which the transmitted pulse train causes the RX signal to exceed the detection threshold 412 ( Figure 4B )) or 401 (

[0034] In Figure 4A the RX signal is compared with a dynamic threshold 412, which can be derived using a constant false alarm rate (CFAR) method. There are many CFAR variants that provide different trade-offs between performance and computational complexity, but in each case, this method is designed to keep the probability of detecting false echoes at a relatively constant level, even in the presence of varying background noise. The CFAR threshold 412 increases in the presence of strong signals and / or noise and decreases when only weak signals or noise are present.

[0035] When the comparator compares the RX signal with the CFAR threshold 412 in this example, it generates a comparison signal (RX>CFAR) with pulses 421 to indicate the transmitted pulse train and two subsequent pulses 428, 429 to indicate the presence of potential echoes 418 and 419. In some contemplated embodiments, the RX signal may be amplified during the active period of the comparison signal. However, when considering the characteristic echo pulse shape (i.e., considering the rising edge and / or falling edge), better performance is desired.

[0036] To confirm that the pulses in the comparison signal indicate echoes rather than noise, the contemplated detection method also estimates the normalized derivative (DERIV) of the RX envelope signal. In at least some contemplated embodiments, the RX envelope is a digital signal. The sensor controller delays the RX signal by one sampling interval and subtracts it from the undelayed RX signal to obtain the estimated derivative. The estimated derivative can be normalized by taking the absolute value of each difference and running it through two filters. The first local moving average filter determines the sum or average of the absolute differences within a small window, the size of which is optimized for accuracy and noise. The second background filter determines the weighted sum of the absolute differences over a window, which is at least several times that of the local filter. The second filter can be a recursive filter that provides an exponential weighting of past absolute differences. The normalized derivative (DERIV) is the ratio of the output of the first (local) filter to the output of the second (background) filter, and it is compared with a predetermined threshold 431. Each of the filters is preferably programmable to match the bandwidth of the filter used to detect potential echoes. The DERIV curve includes peaks 432 indicating the rising edge of the transmitted pulse 411, peaks 433 indicating the falling edge of the pulse 411, peaks 434 and 435 indicating the rising and falling edges of the echo pulse 418, and peaks 436 and 437 representing the rising and falling edges of the echo pulse 419.

[0037] The normalized derivative can be compared with a predetermined derivative threshold 431 to enable the detection of echo pulse edges. In some alternative embodiments, the background filter is omitted and the filtered derivative signal (instead of the ratio) is compared with a predetermined threshold. The threshold can be fixed or time-dependent. When the detection of a pulse edge is combined with a low-to-high transition in the CFAR comparison, the rising edge (R.EDGE) signal becomes active. Thus, when the normalized derivative signal exceeds the threshold near the upward transition of the comparison signal (RX>CFAR), the rising edge signal contains pulses 442. (Although not shown here, the normalized derivative signal can be delayed to allow the detection of the upward transition before determining that the derivative signal exceeds the threshold.) Similarly, pulses 448 and 449 in the rising edge signal identify the rising edges of echo pulse 418 and echo pulse 419.

[0038] In some contemplated embodiments, the RX signal is amplified during the active period of the rising edge signal. In at least some embodiments, the amplification takes the form of multiplication by a constant scaling factor. In Figure 4A , the first amplified rising edge signal (MR1) shows the RX signal, where the rising edge is scaled by a factor of 2. The transmit pulse 411 has a steep rising edge 452; the echo pulse 418 has an amplified rising edge 458; and the echo pulse 419 has an amplified rising edge 459. In certain variations, the constant scaling factor is "windowed" or otherwise provided with a smooth transition from zero to a maximum value in order to reduce abrupt transitions in the modified signal. In other variations, the scaling factor is replaced by a power (e.g., quadratic) or other non-linear function to provide the desired amplification. In other variations, a dynamic threshold is subtracted from the RX signal and selective amplification is applied to the difference.

[0039] Some contemplated embodiments completely suppress the RX signal outside the amplification region, as shown by the second amplified rising edge signal MR2. Compared to the first signal MR1, the rising edges 452, 458, and 459 become pulses 462, 468, and 469, thus preventing noise or extraneous signal energy from interfering with the "sharpening" of the echo.

[0040] As an alternative, the foregoing method can be applied to the falling edge instead of the rising edge. The falling edge signal (F.EDGE) has pulses 472, 478, 479 that are active, where the derivative signal exceeds a threshold near the downward transition in the comparison signal. The first amplified falling edge signal MF1 shows the RX signal including the transmit pulse 411 having a steep falling edge 492, and the echo pulses 418, 419 having amplified falling edges 488, 489. The second amplified falling edge signal MF2 also has the RX signal that is zeroed outside the amplification region, converting the falling edges 482, 488, and 489 into falling edge pulses 492, 498, and 499. As described above, a fixed scaling factor or non-linear function can be applied to the RX signal (or the difference between the RX signal and the threshold), with or without windowing to smooth the transition.

[0041] Whether in the form of MR1, MR2, MF1, MF2, as shown in the figures or in one of the other forms disclosed herein, the echo pulses in the modified RX signal are at least partially amplified relative to the original RX signal or relative to the non-echo portion of the modified RX signal. The non-echo portion of the original RX signal can also or alternatively be reduced, zeroed or otherwise attenuated. The modified RX signal is then used in place of the original RX signal for subsequent processing of echo detection.

[0042] Figure 4BThe use of the original RX signal and the modified RX signal MF2 is compared when performing echo detection using the predetermined detection threshold template 401. Since the template 401 seems to be designed for a more noisy environment of the proportional indication operation, it is set too high to detect the echoes 418, 419. However, regardless of the suboptimal nature of the template 401, when it is used to detect the echoes in the modified RX signal, the amplified echoes (or amplified echo edges) ensure that the detection process is performed as required. The modified RX signal is expected to enhance the performance of various existing echo detection techniques, whether performed by the sensor controller, the processing ASIC, or the ECU. The resulting time-of-flight determination can be used for obstacle detection and monitoring in the usual manner.

[0043] Figure 5 is an illustrative embodiment of the CFAR method from US5793326 ("Hofele") that can be used by the sensor controller. A shift register with i blocks (each block having L samples of the received signal envelope RX) is shifted to accept a new block of samples from the input E. The block at the center is designated as the cell under test ("ZUT"), while the blocks to the left of the center form the sub-register S1 and the blocks to the right form the sub-register S2. The summing circuit determines the sum of the samples within each block. A set of maximum detectors compare the sums in pairs, working outwards from the cell under test, and each detector forwards the maximum sum. The minimum detector compares the maximum values to determine the smallest one. The distributor multiplies the smallest maximum value by L or some other fixed value for normalization before the multiplier K weights the normalized value (optionally adding an offset) to determine the CFAR threshold. The comparator KD compares each sample in the cell under test with the threshold to determine if there is a potential echo peak. As previously mentioned, although there are different trade-offs between performance and computational complexity, there are many variations, and they will also apply. One variation positions the ZUT on the right side of the shift register (for detecting rising edges) or positions the ZUT in the shift register (for detecting falling edges).

[0044] Figure 6 is a block diagram of an illustrative echo amplification process that can be implemented by the sensor controller. The mixer down-converts the received signal to baseband or near baseband. The filter blocks the unwanted frequencies from the down-conversion process and can also perform rectification and low-pass filtering to obtain the envelope of the received signal. The analog-to-digital converter (ADC) digitizes the received signal envelope, which is directed along three branches.

[0045] In one branch, a derivative element ("Norm.Deriv.") determines a normalized derivative as previously discussed. A comparator compares the normalized derivative with a programmable threshold, producing a binary comparator signal that becomes active when the threshold is exceeded. A delay element provides a predetermined delay before the comparator signal is provided to an AND gate. In some embodiments, the comparator signal must be active for at least a predetermined minimum interval, otherwise the activation is blocked from reaching the AND gate.

[0046] In a second branch, a CFAR element derives a CFAR threshold from the RX signal, which optionally includes a programmable offset. A comparator compares the RX signal with the CFAR threshold, producing an output signal that becomes active when the RX signal is above the threshold and otherwise deactivates. An edge detector detects a down transition, an up transition, or both of the comparator output according to the configuration of the sensor controller. In some embodiments, the edge detector generates a fixed-length pulse to indicate the position of each edge.

[0047] When the RX signal exceeds the CFAR threshold in the desired direction and the normalized derivative exceeds the derivative threshold, a logical AND activates an echo edge detection signal. The AND element can be configured to generate a fixed-length pulse to indicate when the inputs are simultaneously active, or a variable-length pulse corresponding to the duration of the interval during which the derivative exceeds the threshold.

[0048] Along a third branch, the original RX signal is provided to a multiplier and a first input of a multiplexer. The multiplier output is provided to another input of the multiplexer. The echo edge detection signal is provided to the multiplexer to select between the inputs, such that the multiplexer outputs an amplified RX signal when an echo edge is detected and an unamplified (or alternatively, zeroed) signal otherwise. The resulting modified RX signal is then forwarded for processing according to an existing echo detection process. For example, the modified RX signal can be compared with a predetermined threshold template, and an echo can be identified where the modified RX signal exceeds the threshold.

[0049] Although the operations shown and described above are considered to occur sequentially for purposes of illustration, in practice, the method can be performed by multiple integrated circuit components that operate simultaneously and possibly even speculatively to achieve unordered operation. The sequential discussion is not intended to be limiting. Once the above disclosure is fully understood, these and many other modifications, equivalents, and alternative forms will become apparent to those skilled in the art. The following claims are intended to be interpreted to include all such modifications, equivalents, and alternative forms where applicable.

[0050] Broadly speaking, various exemplary embodiments have been disclosed. In one exemplary embodiment, a sensor controller includes: a transmitter, a receiver, and processing circuitry coupled to the transmitter and the receiver. The transmitter drives a piezoelectric element to generate a burst of acoustic pulses. The receiver senses the response of the piezoelectric element to the echoes of each burst of acoustic pulses. The processing circuitry is operable to apply an echo detection process to the response by: identifying intervals of the response that represent at least a portion of a potential echo; deriving a modified response from the response by selectively amplifying the response during the intervals; and using the modified response to detect an echo.

[0051] In an exemplary method embodiment, a piezoelectric-based sensor operates by: driving a piezoelectric transducer to generate a burst of acoustic energy during an actuation interval; acquiring a response of the piezoelectric transducer during a measurement interval following the actuation interval; and processing the response to sense an echo of the burst. The processing includes: identifying intervals of the response that represent at least a portion of a potential echo; deriving a modified response from the response by selectively amplifying the response during the intervals; and using the modified response to detect an echo.

[0052] Each of the foregoing exemplary embodiments may be used in conjunction with any one or more of the following optional features: 1. Identifying the intervals includes: comparing the response to an adaptive threshold to produce a comparison signal; and detecting an edge in the comparison signal. 2. The processing further includes using a constant false alarm rate (CFAR) method to derive the adaptive threshold from the response signal. 3. The identifying further includes comparing an estimated derivative to a threshold. 4. The intervals correspond to the rising edge of an echo pulse. 5. The intervals correspond to the falling edge of an echo pulse. 6. Selective amplification includes multiplying by a scaling factor. 7. Selective amplification includes applying a non-linear function to the response during the intervals to the response during the intervals. 8. Selective amplification includes reducing the response outside the intervals. 9. The echo is detected by comparing the modified response to a predetermined threshold template.

Claims

1. A controller for a piezoelectric transducer, characterized in that it comprises: a transmitter for driving a piezoelectric element to generate a train of acoustic pulses; a receiver for sensing the response of the piezoelectric element to the echo of each train of acoustic pulses; and a processing circuit coupled to the transmitter and the receiver, the processing circuit being operable to apply an echo detection process to the response, the process comprising: identifying intervals of the response that represent at least a portion of a potential echo; deriving a modified response from the response by selectively amplifying the response during the intervals; and detecting an echo using the modified response.

2. The controller according to claim 1, characterized in that the identifying comprises: comparing the response with an adaptive threshold to produce a comparison signal; and detecting an edge in the comparison signal.

3. The controller according to claim 2, characterized in that the process further comprises deriving the adaptive threshold from the response using a constant false alarm rate (CFAR) method.

4. The controller according to claim 2, characterized in that the identifying further comprises comparing an estimated derivative with a threshold.

5. The controller according to claim 1, characterized in that the selectively amplifying comprises multiplying by a scaling factor that is greater than 1.

6. The controller according to claim 1, characterized in that the selectively amplifying comprises applying a non-linear function to the response during the intervals.

7. The controller according to claim 1, characterized in that the selectively amplifying comprises reducing the response outside the intervals.

8. The controller according to claim 1, characterized in that the echo is detected by comparing the modified response with a predetermined threshold template.

9. The controller according to any one of claims 1-8, characterized in that the intervals correspond to the rising edge of an echo pulse.

10. The controller according to any one of claims 1-8, characterized in that the intervals correspond to the falling edge of an echo pulse.

Images