Technical Field

The present application describes an interference-resilient LiDAR waveform and estimation method thereof .

Background art

To the best of the inventors ' knowledge , many di f ferent waveforms have been proposed for LiDAR sensing, namely Short pulses ; Amplitude modulated continuous wave (AMCW) ; Optical code division multiple access ( OCDMA) ; and Frequency modulated continuous wave ( FMCW) .

Of all the mentioned waveforms , only FMCW is able to address both long range and interference , as it requires a coherent receiver . This means that the range is proportional to the received signal power, Psig, instead of P ² _Sig , as is generally the case of incoherent detection . A coherent receiver also inherently attenuates signals that are not coherent with its local oscillator ( LO) , which means that incoherent interfering signals are attenuated, whereas the LiDAR' s own transmitted signal is not .

Short pulses are best in maximizing range ; however, they are helpless when j ammed by other interfering pulses , as the receiver is unable to distinguish signal from interference pulses . AMCW signals are the opposite case : excellent against interfering pulses or even AMCW tones with di f ferent frequencies , but fail to maximi ze range . An OCDMA signal can be regarded as a particular case of an AMCW signal , in which only a few periods are ef fectively modulated in order to generate the right code . As a result , OCDMA signals shares the same pros and cons as AMCW signals .

Considering the known state of the art limitations with regard to range and precision, the present invention goal is to improve and maximi ze the range and precision of a LiDAR sensor, while ensuring interference detection .

Summary

The present invention describes a method to estimate the time of flight of an emitted waveform comprising the steps of : acquiring an input waveform, resulting from the reflection of the emitted waveform on an obj ect or surface ; sampling time instants at which the input waveform crosses at least one amplitude threshold; veri fying whether the sampled time instants are valid and match the emitted waveform; producing a time-of- flight estimation based on valid time instants of the input waveform .

In a proposed embodiment of the disclosed method, the sampled time instants comprise at least a time instant tl and time instant t2 .

Yet in another proposed embodiment of the method, the time- of- flight estimation of the input waveform is discarded i f the sampled time instants do not resemble the emitted waveform .

Yet in another proposed embodiment of the method, the emitted waveform comprises at least a first strong short pulse with high amplitude and at least a second weak short pulse with low amplitude .

Yet in another proposed embodiment of the method, veri fying i f the set of time instants are valid comprises determining i f a pulse width of each pulse of the input waveform matches a pulse width of each pulse of the emitted waveform; and determining i f a time interval between pulses of the input waveform matches a time interval between pulses of the emitted waveform .

Yet in another proposed embodiment of the method, the time instants are sampled by a time-to-digital converter through a predefined amplitude threshold that determines all time instants at which the input waveform crosses said threshold .

Yet in another proposed embodiment of the method, given that the power of the input waveform is unknown beforehand, i f the input waveform has a reasonable power and does not overtake a saturation limit , a time-to-digital converter outputs four time instants , being able to sample both pulses of the waveform, said four time instants are then used for waveform validation and time-of- f light estimation .

Yet in another proposed embodiment of the method, given that the power of the input waveform is unknown beforehand, i f the input waveform has a low power and does not overtake a saturation limit , and i f the time-to-digital converter outputs two , or three time instants , tl and t2 or tl , t2 and t3 , said time instants are then used for waveform validation and for time-of- flight estimation . In fact , i f the second pulse is not sampled by the time-to-digital converter, time instants t3 and t4 are not obtained, but the amplitude of the first pulse can be nonetheless inferred from tl and t2 . By knowing the amplitude , the DSP may infer that time instants t3 and t4 cannot be obtained, as the second pulse must have fell below the threshold defined by the TDC stage . As a result , the DSP may validate the waveform even i f time instants t3 and t4 are not inputted . In general , i f less than four time instants are sampled by the time-to-digital converter, then the received waveform first pulse is considered not to overtake the receiver saturation limit resulting that two time instants suf fice to produce a time- of- flight estimation without systematic error, such as time- to-digital converter walk-of f error .

Yet in another proposed embodiment of the method, given that the power of the input waveform is unknown beforehand, i f the input waveform has a high power and overtakes a saturation limit , and i f time-to-digital converter outputs four time instants , such time instants are then used for waveform validation and for time-of- f light estimation with increased precision .

Yet in another proposed embodiment of the method, the time- of- flight estimation is determined from a function of the time instants .

Yet in another proposed embodiment of the method, the function of the time instants is inferred using machine learning methods .

The present invention further describes a computer program configured to carry out every step of the described method . The present invention further describes a (non-transitory ) machine-readable storage device , on which the computer program configured to carry every step of the described method is stored .

The present invention further describes a data processing system, comprising the necessary physical means for the execution of the computer program configured to carry every step of the described method .

The present invention further describes an electronic control unit , configured to carry out every step of the method herein disclosed .

General Description

The present invention disclosure focusses on a new LiDAR sensor waveform that enables maximi zing range and precision while enabling one to detect interference , and additionally, it describes a method for processing said waveform, based on time-to-digital conversion . The proposed method samples and processes the LiDAR sensor received signal , veri fying whether such signal has been j ammed by an interfering signal , and estimates the time of flight ( ToF) of said signal .

The present developed technology is centered in incoherent LiDAR receivers , to which other waveforms - short pulses , AMCW and OCDMA - apply .

A receiver for an incoherent LiDAR sensor is based on a photodetector, a sampling device , and a digital signal processor ( DSP ) that estimates the ToF of the received optical waveform . The photodetector is typically an avalanche photodiode (APD) , as such device of fers increased sensitivity in comparison to a simple photodiode . The sampling device can be based on an analog-to-digital converter (ADC ) or on a time-to-digital converter ( TDC ) . Provided that an ADC has suf ficiently high resolution and sampling frequency, it enables the sampled waveform to preserve most i f not all information from the original waveform . This , in turn, enables reaching maximum range and precision . There are , however, disadvantages when adopting an ADC : it requires an independent mixed-signal circuit for each LiDAR receiver, and an ADC with suf ficiently high resolution and sampling frequency has a toll in power consumption . The alternative to an ADC is a TDC, which essentially returns time instants at which the wave form has crossed a given threshold . Contrarily to an ADC, a TDC only samples one or two points per pulse , i . e . , one point for each edge of the pulse . This means that the sampled waveform may not preserve suf ficient information from the original waveform, which may forbid achieving maximum range or precision . There are , however, advantages when adopting a TDC : it can be implemented all-digitally in a Field Programmable Gate Array ( FPGA) , and a TDC is idle in between two arriving pulses , which means low power consumption .

The conceived solution for the new LiDAR sensor waveform and matching estimation method aims to maximi ze range and precision of the waveforms whi le enabling interference detection . Interference detection in the signal waveform is performed with a minimi zed impact and sacri fice of the range of the sensor, allowing to obtain maximum range and precision . Both the estimation method and the interference detection method are compatible with waveform sampling performed by a TDC stage, such that production costs and power consumption are reduced.

Brief description of the drawings

For better understanding of the present application, figures representing preferred embodiments are herein attached which, however, are not intended to limit the technique disclosed herein.

Fig. 1 - discloses an illustration of the devised LiDAR sensor waveform, wherein the reference numbers refer to:

1 - emitted waveform;

2 - first strong short pulse with high amplitude;

3 - second weak short pulse with low amplitude.

Fig. 2 - depicts the proposed estimation method of the devised LiDAR sensor waveform, wherein the reference numbers refer to:

10 - received waveform / input waveform;

11 - first pulse;

12 - second pulse;

21 - time instant tl;

22 - time instant t2;

23 - time instant t3;

24 - time instant t4;

30 - time-to-digital converter (TDC) ;

32 - time-to-digital converter (TDC) threshold;

40 - digital signal processor (DSP) ;

50 - time of flight (ToF) estimation. Fig. 3 - depicts the behavior of the TDC (30) for a received waveform (10) with reasonable power, without overtaking a receiver saturation limit, and where four time instants are sampled by the TDC. The reference numbers relate to:

10 - received waveform / input waveform;

11 - first pulse;

12 - second pulse;

21 - time instant tl;

22 - time instant t2;

23 - time instant t3;

24 - time instant t4;

31 - Receiver saturation limit;

32 - time-to-digital converter (TDC) threshold.

Fig. 4 - depicts the behavior of the TDC (30) for a received waveform (10) with low power, without overtaking a receiver saturation limit, and where three time instants are sampled by the TDC. The reference numbers relate to:

10 - received waveform / input waveform;

11 - first pulse;

12 - second pulse;

21 - time instant tl;

22 - time instant t2;

23 - time instant t3;

31 - Receiver saturation limit;

32 - time-to-digital converter (TDC) threshold.

Fig. 5 - depicts the behavior of the TDC (30) for a received waveform (10) with low power, without overtaking a receiver saturation limit, and where two time instants are sampled by the TDC. The reference numbers relate to:

10 - received waveform / input waveform;

11 - first pulse; 12 - second pulse;

21 - time instant tl;

22 - time instant t2;

31 - Receiver saturation limit;

32 - time-to-digital converter (TDC) threshold.

Fig. 6 - depicts the behavior of the TDC (30) for a received waveform (10) with high power, where the signal overtakes the receiver saturation limit, and where four time instants are sampled by the TDC. The reference numbers relate to:

10 - received waveform / input waveform;

11 - first pulse;

12 - second pulse;

21 - time instant tl;

22 - time instant t2;

23 - time instant t3;

24 - time instant t4;

31 - Receiver saturation limit;

32 - time-to-digital converter (TDC) threshold.

Description of Embodiments

With reference to the figures, some embodiments are now described in more detail, which are however not intended to limit the scope of the present application.

In a particular embodiment of the present disclosure, the devised LiDAR sensor waveform (1) illustrated in Figure 1 comprises two short pulses, a first strong short pulse (2) followed by a second weaker pulse (3) . The pulses are therefore defined to have a given amplitude, pulse width and time interval between them. The emitted waveform (1) is simple and yet powerful. The emitted signal (1) by the sensor allocates most power to the first short pulse (2) , in order to enable maximum range and precision, and allocates just enough power to the second short pulse (3) to make it detectable within the entire predefined range of the sensor, therefore enabling interference detection for the entire range.

The proposed estimation method is illustrated in figure 2. The LiDAR sensor receives a given optical signal, which is converted into an electric signal by means of a photoreceiver. The photoreceiver then produces an electric signal that corresponds to the input waveform (10) of the TDC (30) stage. The TDC (30) essentially defines an amplitude threshold (32) , and outputs a time instant on each time the input waveform (10) crosses that threshold (32) , within the rising and falling of the pulses (10) range. The input waveform (10) is sampled by a TDC (30) , which produces up to four samples, that are time instants (21, 22, 23, 24) . Then, the DSP (40) verifies whether the time instants (21, 22, 23, 24) are valid, in order to validate whether the input waveform (10) matches the devised waveform (1) . If so, it produces a time-of-f light (ToF) estimation (50) derived from such time instants (21, 22, 23, 24) .

With regard to the conceptual basics of the estimation method, and as depicted in the previously mentioned Figures, the input waveform (10) processed by the TDC (30) produces four time instants, tl (21) , t2 (22) , t3 (23) and t4 (24) , which are fed to a DSP (40) . The DSP (40) then performs the following two tasks: 1. determine whether the input waveform (10) is valid,

1.e., whether the input time instants (21, 22, 23, 24) fall within reasonable values. This feature is determined through verification of whether the received waveform (10) resembles to emitted devised waveform (1) , such that other waveforms (e.g., a strong single short pulse or a waveform corrupted by interference) are discarded. Such a feature thus enables interference detection;

2. If the waveform (10) is found to be valid, the DSP (40) produces a ToF estimation (50) from the sampled time instants (21, 22, 23, 24) , i.e., ToF= ( tl , t2 , t3 , t4 ) .

With regard to the particular details of the estimation method, the validation of the waveform can be made resorting to the following procedures:

1. determining whether the pulse width of each pulse is correct ;

2. determining whether the time interval between pulses is correct.

Since the power of the input waveform (10) is unknown beforehand, it is required to analyse it case by case, within the following steps:

1. The received waveform (10) has a reasonable power and does not overtake the receiver saturation limit (31) , (illustrated in Figure 3) . a. The TDC (30) outputs four time instants (21, 22, 23, 24) , which means that it is able to sample both pulses of the waveform. Such four time instants (21, 22, 23, 24) are then used for waveform validation and ToF estimation (50) . The received waveform (10) has a low power and does not overtake the receiver saturation limit (31) , (illustrated in Figures 4 and 5) . a. In this case the TDC (30) outputs two, or at most, three time instants (Figure 5 and 4 respectively) , tl (21) and t2 (22) or tl (21) and t2 (22) and t3 (23) , are then used for waveform validation and for ToF estimation (50) . b. Even though the second pulse may not be sampled by the TDC (30) , the amplitude of the first pulse (11) can be nonetheless inferred from tl (21) and t2 (22) . By knowing the amplitude of this pulse (11) , the DSP (40) may infer that time instants t3 (23) and t4 (24) cannot be obtained, as the second pulse (12) must have fell below the defined TDC threshold (32) . As a result, the DSP (40) may validate the waveform even if time instants t3 (23) and t4 (24) are not inputted. c. The precision in estimating the ToF (50) depends on the number of time instants the DSP (40) receives, that is, the more time instants received the better. Nonetheless, if less than four time instants are produced by the TDC (30) , then the received waveform (10) first pulse (11) is considered not to overtake the receiver saturation limit (31) . In turn, this means that two time instants (21, 22) may suffice to produce a precise estimation without systematic error, such as TDC walk-off error. The received waveform (10) has high power, and overtakes the receiver saturation limit (31) , (illustrated in Figure 6) . a. TDC (30) outputs four time instants (21, 22, 23, 24) , meaning that it is able to sample both pulses (11, 12) of the waveform (10) . Such time instants are then used for waveform validation and for ToF estimation (50) . b. This case differs from the previous one mainly in precision and in estimating the ToF (50) . Precision is capped by the saturation limit of the receiver (31) and by the resolution of the TDC (30) . In the present case, as two strong short pulses (11, 12) are received, precision is at least twice as better than the first case. This is yet another advantage that the disclosed waveform (1) and estimation method provide.

With regard to particular embodiments of the current invention, the waveform (1) may comprise more than two pulses. Such can be used for trading off maximum range (the less pulses, the better) for better maximum precision (the more pulses, the better) . The time interval between two consecutive pulses needs not be the same for all consecutive pulses. The amplitude should decrease from pulse to pulse, as for improving precision it is important to have a first strong pulse. The pulse width and time interval between pulses should be defined as to enable the TDC (30) to correctly sample all time instants.

With regard to the TDC (30) , it can be defined more than one threshold, therefore producing more time instants. Such can be beneficial for improving range and precision, however at the cost of a more complex TDC stage. The thresholds of the TDC (30) can be adjusted depending on the scenario, e.g., on the noise level. It can also be implemented all-digitally