DE102019217074A1 - Method for determining a time signal for a driver assistance system - Google Patents

Abstract

Method for determining a time signal for a driver assistance system (100), comprising the steps: - synchronizing a clock (31) of a third control device (30) with a clock (11) of a first control device (10), comprising the steps: - determining times a data transfer from the third control device (30) to the first control device (10) via a control device (40) and a data bus (50); plausibility checking of the times relating to the data transfer via the data bus (50); and- determining an offset between the clock (31) of the third control device (30) and the clock of the first control device (10).

Description

The present invention relates to a method for determining a time signal for a driver assistance system. The invention also relates to a control device for a driver assistance system. The invention also relates to a driver assistance system. The invention also relates to a computer program. The invention also relates to a machine-readable storage medium.

State of the art

Driver assistance systems with a plurality of control units are known, in which individual signals from control units are processed by other control units, and where individual time signals are used in each case. Driver assistance systems are known which place highly critical accuracy requirements with regard to time signals.

Disclosure of the invention

One object of the invention is to provide an improved method for operating a driver assistance system with at least two control units.

• - Synchronisieren einer Uhr einer dritten Steuereinrichtung mit einer Uhr einer ersten Steuereinrichtung, aufweisend die Schritte:
• - Ermitteln von Zeitpunkten eines Datentransfers von der dritten Steuereinrichtung zur ersten Steuereinrichtung über eine Steuervorrichtung und einen Datenbus;
• - Plausibilisieren der Zeitpunkte betreffend den Datentransfer über den Datenbus; und
• - Ermitteln eines Offsets zwischen der Uhr der dritten Steuereinrichtung und der Uhr der ersten Steuereinrichtung.

According to a first aspect, the object is achieved with a method for operating a sensor device of a vehicle, with the steps:

• - Synchronizing a clock of a third control device with a clock of a first control device, comprising the steps:
• Determining times of a data transfer from the third control device to the first control device via a control device and a data bus;
• - Checking the plausibility of the times relating to the data transfer via the data bus; and
• - Determining an offset between the clock of the third control device and the clock of the first control device.

In this way, a point in time is determined that can be represented in the time domains of the third control device and the first control device, based on the simplifying assumption that a time of sending the data corresponds to a time of receiving the data, taking into account the duration of the data transfer a data bus is used to determine the offset of the two clocks. As a result, the time signal of the central control device can be "raised" to a higher level of accuracy (e.g. ASIL B), which is a standard-based requirement for safety-critical driver assistance systems (e.g. emergency brake assistant). This means that the time signal in question can be protected against E / E errors.

According to a second aspect, the object is achieved with a control device which is set up to carry out the proposed method.

• - eine Steuervorrichtung;
• - wenigstens zwei an die Steuervorrichtung angeschlossene Steuereinrichtungen; wobei
• - die Steuereinrichtungen über die Steuervorrichtung mittels eines Datenbus verbunden sind;
• - wobei das Fahrerassistenzsystem eingerichtet ist, einen Offset zwischen Uhren der wenigstens zwei Steuereinrichtungen zu ermitteln.

According to a third aspect, the object is achieved with a driver assistance system of a vehicle, having:

• - a control device;
• - At least two control devices connected to the control device; in which
• the control devices are connected via the control device by means of a data bus;
• - wherein the driver assistance system is set up to determine an offset between clocks of the at least two control devices.

According to a fourth aspect, the object is achieved with a computer program comprising instructions which, when the computer program is executed by a computer, cause the computer to carry out a proposed method. In this way, the method can advantageously be designed as software and thereby modified and adapted in a simple and efficient manner.

According to a fifth aspect, the object is achieved with a machine-readable storage medium on which the proposed computer program is stored.

Advantageous further developments of the method are the subject of the dependent claims.

An advantageous development of the method provides that a duration of a modulation phase and a duration of processing the data are taken into account per cycle of the method in a time domain of the first control device. As a result, specific, physically determined processing times are used per cycle, which can be displayed very well.

Another advantageous development of the method provides that a point in time is determined which can be displayed both in the time domain of the third control device and in the time domain of the first control device, based on the offset of the clocks taking into account the latency of the data transfer via the data bus is determined. In this way, a time stamp can advantageously be used which is available and can be displayed in both time domains and from which the offset is determined on the basis of.

Another advantageous development of the method provides that a statistical model is used to determine processing times of the processing chain. As a result, the processing times from which the offset is determined can be determined by means of a statistical model.

Another advantageous development of the method provides that the values are recorded over a number of cycles and the offset is determined therefrom. In this way, the accuracy of the determined offset can advantageously be increased even further.

Another advantageous development of the method provides that a time signal from a fourth clock of the sensor device is taken into account. As a result, the method provides a plausibility-checked time master for the central control device.

mit: T10 Periodisches Task im Radarsensor, der den Datenversand beinhaltet t_DASy Empfangszeitpunkt der Messdaten in DASy in lokaler DASy Zeitdomäne t_in ^DASy Empfangszeitpunkt der Messdaten in DASy in lokaler DASy Zeitdomäne t_meas Messzeitpunkt in gemeinsamer Zeitdomäne t_meas ^MRR Messzeitpunkt in gemeinsamer Zeitdomäne t_icycleEnd ^MRR Zeitpunkt des Abschluss des BG-Zyklusses in Radarsensor in gemeinsamer Zeitdomäne t_BG,_T10 Übertragungslatenz zwischen BG und T10 Tasks in Radarsensor t_{Receive, T10} Übertragungslatenz zwischen BG und T10 Tasks in DASy t_transmission Übertragungslatenz zwischen Radarsensor und DASy E ( ) Erwartungswert E Stochastischer Anteil von jeweiliger Zeitdauer Δ bzw. Δ_MRR,DASy Versatz der lokalen DASy Zeitdomäne (DASy interne Uhrzeit) gegenüber der gemeinsamen Zeitdomäne (Uhrzeit des Time Masters) Index i Aktueller Verarbeitungszyklus Another advantageous development of the method provides that the offset between the clocks of the third control device and the first control device is determined according to the following equation: $${\Delta }_{\mathrm{M.}\mathrm{R.}\mathrm{R.},\mathrm{D.}\mathrm{A.}\mathrm{S.}y,i}=t_{in,i}^{\mathrm{D.}\mathrm{A.}\mathrm{S.}y}-t_{\text{meas},i+1}^{\text{MRR}}+t_{\text{cycleStart}\text{, ModulationEnd}}-t_{\text{transmission}}-\mathrm{E.}\left(t_{\text{BG}\text{, T10}}\right)-\mathrm{E.}\left(t_{\text{Receive}\text{, T10}}\right)-{\epsilon }_{t,i}$$

With: T10 Periodic task in the radar sensor that includes the data transmission t _DASy Time of receipt of the measurement data in DASy in the local DASy time domain t _in ^DASy Time of receipt of the measurement data in DASy in the local DASy time domain t _meas Measurement time in a common time domain t _meas ^MRR Measurement time in a common time domain t _icycleEnd ^MRR Time of completion of the BG cycle in the radar sensor in a common time domain t _BG , _T10 Transmission latency between BG and T10 tasks in the radar sensor t _{Receive, T10} Transmission latency between BG and T10 tasks in DASy t _transmission Transmission latency between the radar sensor and DASy E () Expected value E. Stochastic portion of the respective duration Δ or Δ _{MRR, DASy} Offset of the local DASy time domain (DASy internal time) compared to the common time domain (time of the time master) Index i Current processing cycle

Further measures improving the invention are illustrated in more detail below together with the description of the preferred exemplary embodiments of the invention with reference to figures.

Figure list

• 1 ein prinzipielles Übersichtsbild mit einer Ausführungsform eines vorgeschlagenen Fahrerassistenzsystems;
• 2 ein prinzipielles Übersichtsbild mit einer Beschreibung einer Wirkungsweise des vorgeschlagenen Verfahrens;
• 3 ein prinzipielles Übersichtsbild mit einer Beschreibung einer Wirkungsweise des vorgeschlagenen Verfahrens;
• 4 ein prinzipielles Übersichtsbild mit einer Beschreibung einer Wirkungsweise des vorgeschlagenen Verfahrens; und
• 5 einen prinzipiellen Ablauf eines vorgeschlagenen Verfahrens zum Ermitteln eines Zeitsignals für ein Fahrerassistenzsystem.

In the accompanying figures shows:

• 1 a basic overview image with an embodiment of a proposed driver assistance system;
• 2 a basic overview picture with a description of a mode of operation of the proposed method;
• 3 a basic overview picture with a description of a mode of operation of the proposed method;
• 4th a basic overview picture with a description of a mode of operation of the proposed method; and
• 5 a basic sequence of a proposed method for determining a time signal for a driver assistance system.

Embodiments of the invention

A key concept of the invention is in particular to provide time synchronization between a central control device and a sensor device without an ASIL-B capable time base or with a plausibility-checked ASIL-B time base.

1 shows a driver assistance system 100 with a group of control devices 10 , 20th , 30th connected to a central control device ( 40 ) are connected and can communicate with each other via them. A first control device 10 (Domain control unit for driver assistance systems, DASy) is a central computing device of the driver assistance system 100 provided in the sensor data are fused or integrated or merged. The first control device 10 represents a key component with high bandwidth, computing power and storage capacity that collects and processes data from different technologies (e.g. radar, video, lidar, ultrasound) for a very precise 360 ° model of the environment. It calculates highly complex functional algorithms for safe and dynamic vehicle behavior, even at higher speeds and with high safety requirements. The first control device 10 is by means of a data bus 50 to the device 40 tied up and has a first clock 11 on. The first watch 11 represents a time domain of the first control device 10 .

A second control device 20th represents, for example, a control unit of a video sensor. The second control device 20th can by means of a communication link 60 to the control device 40 be connected. A second clock 21st represents a time domain of the second control device 20th .

A third control device 30th represents a control unit for a sensor device, for example a radar sensor. On the third control device 30th is a third clock 31 arranged, the one time domain of the third control device 30th represents. The sensor device 30th is by means of a data bus 50 (e.g. CAN bus) with the control device 40 connected.

The control device 40 (English central gateway) represents with its fourth clock 41 a time master, ie a global time base or domain for the entire driver assistance system 100 with all control devices 10 , 20th , 30th . Here is the fourth o'clock 41 the control device, however, is not capable of ASIL, but can only provide a time signal with a lower level of accuracy (eg at QM level). For a safety-critical driver assistance system 100 (e.g. emergency braking assistant, English automatic emergency braking system, AEB) a time signal must, however, meet the requirements of ASIL-B.

It is proposed to raise a security level of the fourth clock 41 the control device 40 by synchronizing the third clock 31 the third control device 30th with the first watch 11 the first control device 10 .

This enables the driver assistance system 100 can also be used for safety-critical applications in vehicles.

The present invention aims at safety-critical driver assistance / automated driving functions, such as an automatic emergency braking system (AEB), in which several devices, such as a radar sensor, a video sensor and / or a processing unit (such as DASy) that merge sensor data, are involved and where the time master does not meet the ASIL requirements.

With the proposed driver assistance system 100 provision is made for time synchronization for the various control devices for converting the time stamps of at least two control devices 10 , 20th , 30th to undertake. The inaccuracy of this conversion of the measurement time stamp from the time domain of the third control device 30th into the time domain of the first control device 10 is relevant because it combines with a quality of different parts of the environment detection, in particular with a quality of the tracking.

• Liegt bei einer Folgefahrt mit konstanter Eigengeschwindigkeit von 40 m/s und verschwindender longitudinaler Relativgeschwindigkeit zu einem Vorderfahrzeug im Tracking in der ersten Steuereinrichtung 10 ein Fehler bei der Schätzung des Messzeitstempels betreffend Sensordaten von 80ms vor, so entspricht dies einem Entfernungsfehler in den Messdaten des Vorderfahrzeugs von 40 m/s x 0,08Sekunden = 3,2 m, was einen beträchtlichen Entfernungsfehler darstellt und eine Umfelderfassung verschlechtert.

Example:

• In the case of a follow-up journey at a constant vehicle speed of 40 m / s and a vanishing longitudinal relative speed to a vehicle in front, it is tracked in the first control device 10 If there is an error in the estimation of the measurement time stamp relating to sensor data of 80 ms, this corresponds to a distance error in the measurement data of the vehicle in front of 40 m / sx 0.08 seconds = 3.2 m, which represents a considerable distance error and worsens the detection of the surroundings.

1. 1. Verwendung einer synchronisieren Zeitbasis unter Verwendung eines zentralen Time Master
2. 2. Einsatz einer Schätzung des Messdatenalters ohne Verwendung eines zentralen Time Masters

There are two possible approaches to time synchronization:

1. 1. Use of a synchronized time base using a central time master
2. 2. Use of an estimation of the measured data age without using a central time master

In the present invention, the central control device functions 40 as a time master, which, however, only fulfills the requirements for QM. However, due to the fact that an AEB emergency brake assistant requires ASIL-B, a time synchronization based on the QM-based time master of the central control device is to be used 40 only permitted with additional security mechanisms.

Since this essentially represents an independent solution to the problem, a solution as in 2. is also required in the case of 1..

The paragraph according to the present invention described below can be used both in the case of 2., in which the time master is not used, and as a safeguarding option in the context of 1., in which the time signal of the time master is used.

For this purpose, a processing chain of measurement data is analyzed and, based on this, an estimation error for a duration between the point in time of a measurement in the third control device 30th and the time of receipt of the corresponding data in the first control device 10 determined. Especially when a third control device is involved 30th in the form of a radar sensor, in which the running time of a cycle fluctuates, this leads to a limited accuracy.

The following is the approach using the example of a third control device 30th and a first control device 10 described in more detail. To a second control device 20th in the form of a video sensor, it can be transmitted with certain adjustments (not explained in more detail below).

• - Radarlaufzeit in der Zeitdomäne der dritten Steuereinrichtung 30
• - Sendeverzögerung in der Zeitdomäne der dritten Steuereinrichtung 30
• - Empfangsverzögerung in der ersten Steuereinrichtung 10

2 shows, in principle, three time components in time domains of the third control device 30th , the data bus 50 and the first control device 10 which are not accessible to a preliminary analysis, namely:

• - Radar transit time in the time domain of the third control device 30
• Transmission delay in the time domain of the third control device 30
• - Reception delay in the first control device 10

One particular challenge here is that the transmission delay of the third control device 30th hardly without additional effort (which, for example, takes up additional bus capacities) and the reception delay of the first control device 10 cannot be determined in a practicable way for the current cycle at runtime.

In addition, the information about the cycle duration of the third control device is available 30th not in the same cycle before when the transmission of a second time stamp to the third control device 30th , which is the transmission time in the time domain of the third control device 30th contains, for reasons of reducing the bus communication via the data bus 50 should be avoided.

This invention describes how time synchronization of the measurement time stamps can be achieved without a central time master and despite the occurrence of a transmission and reception delay with an expected inaccuracy that is an order of magnitude lower than this value.

In the event that the time signal of the time master of the control device 40 is used, the approach described here can be used in the context of safeguarding the time master time signal in order to implement an ASIL-B-capable common time base in the case of a QM time master.

The proposed approach is also relevant for driving functions with a high ASIL level. For an ASIL-D driving function in which a radar sensor and a DASy are involved, an ASIL D time master would be necessary for time synchronization based on a time master. By using the approach described here, this need is advantageously eliminated.

The core of the approach described is not to attempt in a single cycle from the time of arrival of the measurement data in the first control device 10 (DASy) to infer the measurement time (in DASy time domain), but rather a time synchronization between the time domain of the first control device over several cycles 10 and the time domain of the third controller 30th based solely on the measurement time stamp in the time domain of the third control device 30th and arrival time stamp in the time domain of the first controller 10 perform, that is, a suitable estimate for the offset and the scaling between the time of the third control device 30th and the time of the first controller 10 to specify.

For this purpose, the processing time between the time of measurement and the time of arrival is analyzed and divided into time components that can be determined in advance, namely into time components that can be determined at runtime and into randomly fluctuating components. The latter can be largely eliminated stochastically.

1. (i) Aus den zur Verfügung stehenden Zeitstempeln (z.B. Messzeitstempel in Sensorzeit, Zeitstempel für Ankunft in DASy-T10 in lokaler DASy-Zeit) solche Zeitstempel erhalten, die sich auf dasselbe Ereignis beziehen (z.B. Zeitstempel für Ankunft in DASy-T10 in Sensorzeit, Zeitstempel für Ankunft in DASy-T10 in lokaler DASy-Zeit)
2. (ii) Berechnung des Versatzes aus (i) zwischen der Zeit der dritten Steuereinrichtung 30 (Sensorzeit) und der Zeit der ersten Steuereinrichtung 10 (DASy-Zeit)
3. (iii) Statistische Beschreibung für Verzögerung der Datenkommunikation zwischen der dritten Steuereinrichtung 30 und der ersten Steuereinrichtung 10 auf dem Datenbus 50

For a rough illustration, the proposed approach can be broken down into three elements:

1. (i) From the available time stamps (e.g. measurement time stamp in sensor time, time stamp for arrival in DASy-T10 in local DASy time) such time stamps are obtained that refer to the same event (e.g. time stamp for arrival in DASy-T10 in sensor time, Time stamp for arrival in DASy-T10 in local DASy time)
2. (ii) Calculation of the offset from (i) between the time of the third control device 30th (Sensor time) and the time of the first control device 10 (DASy time)
3. (iii) Statistical description for delay in data communication between the third control device 30th and the first control device 10 on the data bus 50

1. a) Zeitdomäne der dritten Steuereinrichtung 30: Laufzeit bis zum Ende der Modulationsphase (grau) Laufzeit ab dem Ende der Modulationsphase bis zum Zyklusende (schraffiert)
2. b) Einlesen von Daten des Hintergrund Tasks der dritten Steuereinrichtung 30 in den T10-Task der dritten Steuereinrichtung 30
3. c) Dauer einer Datenübertragung über den Datenbus 50 (z.B. CAN-Bus)
4. d) Zeitdomäne der ersten Steuereinrichtung 10: Verzögerung bis zum Einlesen der Daten durch einen T10-Task in der ersten Steuereinrichtung 10

As in 2 the following processing steps are considered:

1. a) Time domain of the third control device 30th : Running time to the end of the modulation phase (gray) Running time from the end of the modulation phase to the end of the cycle (hatched)
2. b) Reading in data of the background task of the third control device 30th into the T10 task of the third control device 30
3. c) Duration of a data transmission via the data bus 50 (e.g. CAN bus)
4. d) time domain of the first control device 10 : Delay until the data is read in by a T10 task in the first control device 10

to a)

zweier Messzeitstempel der dritten Steuereinrichtung 30 wird auf die Laufzeit des vorangehenden Zyklus der dritten Steuereinrichtung 30 geschlossen. Die Kenntnis der Laufzeit des aktuellen Zyklus ist nicht unmittelbar notwendig.From the difference $t_{meas,i}^{\text{MRR4}}-t_{meas,i-1}^{\text{MRR4}}$

two measurement time stamps of the third control device 30th is based on the running time of the previous cycle of the third control device 30th closed. It is not immediately necessary to know the runtime of the current cycle.

The time from the start of the cycle to the creation of the measurement time stamp can be estimated in advance.

The remaining runtime of the previous cycle after the creation of the time stamp then results from both variables.

to b)

The resulting delay, i.e. the time between completion in the background task and the next call of the T10 task, can be assumed to be evenly distributed over [0; 10 ms].

to c)

The fact that in the present case a data transmission via the data bus 50 exclusively between the third control device 30th and the first control device 10 takes place (private CAN), it can be determined very precisely in advance, after which period of time the relevant data is transmitted.

to d)

As with b). In addition, independence from equal distribution in b) is considered to be justified.

mit den Parametern: MRR Radarsensor (dritte Steuereinrichtung 30) DASy Sensordaten-fusionierendes Steuergerät (erste Steuereinrichtung 10) QM Niedriges Niveau der Abgesichertheit gegenüber E/E Fehler gemäß ISO 26262 ASIL B / ASIL D Mittleres / Hohes Niveau der Abgesichertheit gegenüber E/E Fehler gemäß ISO 26262 BG aperiodischer Task im Radarsensor, der die Messdatenaufbereitung beinhaltet T10 Periodischer Task im Radarsensor, der den Datenversand umfasst (z.B. alle 10ms) t_DASy Empfangszeitpunkt der Messdaten in DASy in lokaler DASy Zeitdomäne t_in ^DASy Empfangszeitpunkt der Messdaten in DASy in lokaler DASy Zeitdomäne t_meas Messzeitpunkt in gemeinsamer Zeitdomäne t_meas ^MRR Messzeitpunkt in gemeinsamer Zeitdomäne t_icycleEnd ^MRR Zeitpunkt des Abschlusses des BG-Zyklus in Radarsensor in gemeinsamer Zeitdomäne t_BG,T10 Übertragungslatenz zwischen BG und T10 Task im Radarsensor t_{Receive, T10} Übertragungslatenz zwischen BG und T10 Task in DASy t_transmission Übertragungslatenz zwischen Radarsensor und DASy E ( ) Erwartungswert E Stochastischer Anteil von jeweiliger Zeitdauer Δ bzw. Δ_MRR,DASy Versatz der lokalen DASy Zeitdomäne (DASy interne Uhrzeit) gegenüber der gemeinsamen Zeitdomäne (Uhrzeit des Time Masters) Index i Aktueller Verarbeitungszyklus Overall, this results in the determination of the offset of the clocks 31 , 11 the third control device 30th and the first control device 10 the following equations: $$\begin{array}{c}{t}_{in,i}^{\mathrm{D.}\mathrm{A.}\mathrm{S.}y}=t_{meas,i}^{\mathrm{M.}\mathrm{R.}\mathrm{R.}}+d_i+{\epsilon }_{t,i}\\ d_i={\Delta }_{\mathrm{M.}\mathrm{R.}\mathrm{R.},\mathrm{D.}\mathrm{A.}\mathrm{S.}y,i}+\left(t_{\text{cycleEnd},i}^{\text{MRR}}-t_{\text{meas},i}^{\text{MRR}}\right)+\text{E.}\left(t_{\text{BG}\text{, T10}}\right)+t_{\text{transmission}}+\text{E.}\left(t_{\text{Receive}\text{, T10}}\right)\\ {\epsilon }_{t,i}={\epsilon }_{\text{BG}\text{, T10}\text{, i}}+{\epsilon }_{\text{Receive}\text{, T10}\text{, i}}\\ {\epsilon }_{\text{Receive}\text{, T10}\text{,}i}=t_{\text{Receive}\text{, T10},i}-\text{E.}\left(t_{\text{Receive}\text{, T10}}\right)\\ {\epsilon }_{\text{BG}\text{, T10}\text{,}i}=t_{\text{BG}\text{, T10},i}-\mathrm{E.}\left(t_{\text{BG}\text{, T10}}\right)\end{array}$$

with the parameters: MRR Radar sensor (third control device 30) DASy Sensor data-merging control device (first control device 10) QM Low level of security against E / E errors according to ISO 26262 ASIL B / ASIL D Medium / high level of security against E / E errors according to ISO 26262 BG aperiodic task in the radar sensor that contains the measurement data processing T10 Periodic task in the radar sensor, which includes the data transmission (e.g. every 10ms) t _DASy Time of receipt of the measurement data in DASy in the local DASy time domain t _in ^DASy Time of receipt of the measurement data in DASy in the local DASy time domain t _meas Measurement time in a common time domain t _meas ^MRR Measurement time in a common time domain t _icycleEnd ^MRR Time of completion of the BG cycle in the radar sensor in a common time domain t _{BG, T10} Transmission latency between BG and T10 task in the radar sensor t _{Receive, T10} Transmission latency between BG and T10 task in DASy t _transmission Transmission latency between the radar sensor and DASy E () Expected value E. Stochastic portion of the respective duration Δ or Δ _{MRR, DASy} Offset of the local DASy time domain (DASy internal time) compared to the common time domain (time of the time master) Index i Current processing cycle

If, for example, the background task BG has ended after 81 ms, another 9 ms will elapse until the next call of the T10 task, in which the measurement data are processed further. This means that there is a delay between the end of the background task BG and the call of the next, following T10 task, which is represented by the expected value E (t _BG , T ₁₀ ). The expected value E is obtained from a statistical model that is formed for the processing chain. This static model is used to determine the offset between the clock 31 the third control device 30th and the clock 11 the first control device 10 .

3 shows in a greater degree of detail the concept of the proposed time synchronization, a time line with two cycles i, i + 1. Continuous gray areas of the cycles indicate modulation phases in which the data from the third control device is physically 30th be procured. The hatched portions of the cycles indicate the processing of the data, with the two areas being repeated cyclically. For reasons of simplification, the time stamps t ^MRR are drawn at the end of the modulation phases (in which the ramp-like radar signals are sent); in fact, they are sent approximately towards the middle of the modulation phases, although this is not essential for the proposed method. The duration of the modulation phases also differs from cycle to cycle, but for the sake of simplicity is shown as being of constant length.

It is indicated that a time stamp is determined which is located both in the time domain of the third control device 30th as well as in the time domain of the first control device 10 can be described. This time stamp is indicated in the 3 by a narrow rectangle that covers both of the time domains mentioned. In the time domain of the third control device 30th this time stamp for the i-th cycle has the designation: $$t_{rec,i}^{\mathrm{M.}\mathrm{R.}\mathrm{R.}}$$

This corresponds to a point in time at which the third control device 30th has finished processing its data and via the data bus 50 sent.

Under the simplifying assumption that the sending time corresponds to the receiving time of the data at the first control device 10 corresponds, this point in time in the time domain corresponds to the first control device 10 the timestamp: $$t_{rec,i}^{\mathrm{D.}\mathrm{A.}\mathrm{S.}y}$$

The measurement times are in the time domain of the first control device 10 for the i-th and the i + 1-th cycle. By determining this matching time stamp in both time domains, it is now possible to determine a time offset or offset that corresponds to an offset of the clocks 31 , 11 corresponds to.

The time offset of the clocks is now 31 , 11 wanted.

The actual estimated target metric you are looking for is: $$t_{meas,i}^{\mathrm{D.}\mathrm{A.}\mathrm{S.}y}$$

This target measured value of the sensor data of the third control device is checked for plausibility 30th from the available measurement times: $$\begin{array}{c}{t}_{meas,i}^{\mathrm{M.}\mathrm{R.}\mathrm{R.}},t_{rec,i}^{\mathrm{D.}\mathrm{A.}\mathrm{S.}y},\text{constant}=t_{meas,i+1}^{\mathrm{M.}\mathrm{R.}\mathrm{R.}}-t_{rec,i}^{\mathrm{M.}\mathrm{R.}\mathrm{R.}}\text{instead of}\text{, where:}\\ t_{meas,i}^{\mathrm{D.}\mathrm{A.}\mathrm{S.}y}=t_{meas,i}^{\mathrm{M.}\mathrm{R.}\mathrm{R.}}+\text{Offset}\end{array}$$

This offset is determined from the time measurements of different time domains in the following way: $$\text{Offset}=t_{rec,i}^{\mathrm{D.}\mathrm{A.}\mathrm{S.}y}-t_{rec,i}^{\mathrm{M.}\mathrm{R.}\mathrm{R.}}$$

In the event that the received value lies within a specified interval, it is assumed that there is no E / E error (hardware-related errors, e.g. bit flips) and the estimated target measured value is therefore suitable for ASIL B.

Can be seen in 3 that to determine the point in time based on the measurement point in time in the time domain of the third control device 30th the duration of the modulation phase assumed to be constant is "decreased". A consideration of a delay due to the communication via the data bus 50 is in 3 not shown.

The above system of equations is graphically shown in 4th shown, the offset sought corresponds to the term Δ _{MRR, DASy, i} of the system of equations mentioned.

In 4th is in addition to 3 recognizable that there is a delay or latency due to data transmission via the bus system 50 is taken into account. This delay has a total of three parts V1 , V2 , V3 . It can be seen that, starting from the time of measurement in the time domain, the duration of the modulation phase, which is assumed to be constant, “decreased” and then the delay time due to the communication via the data bus 50 Is "moved forward".

For this purpose, it is necessary to mathematically describe the communication delay V1 + V2 + V3, where V1 is the delay from the termination of the background task BG to the call of the next T10 task on the third control device 30th run and which can be described statistically very precisely.

V2 represents the delay in communication via the data bus 50 , in which a transmission of the data from the third control device 30th to the first control device 10 takes place and which can be determined very precisely deterministically.

Finally, V3 represents the time delay from the arrival of the data via the data bus 50 on the first controller 10 until the next T10 task is called, which can be statistically described very precisely.

The time lengths of the delays V1 , V2 , V3 can vary greatly from cycle to cycle. In the present case, a duration of 10 ms is assumed for the T10 task, but any other suitable cycle time is of course also conceivable here. With different durations, the tasks on the individual control devices are advantageously minimized 10 , 30th Dependencies in statistical models.

As a result, the time delay due to the communication via the data bus can be reduced 50 describe it very precisely in a statistical model. The values mentioned V1 , V2 , V3 are recorded over a few cycles in order to provide a precise description of the offset between the clocks with knowledge of the distribution 31 , 11 and thus a plausibility check of a time signal of the control device 40 to obtain. The clock's time signal 41 (global time master) of the control device 40 is “upgraded” in this way in order to be able to meet the safety-critical requirements of ASIL B.

With every change of the time domain, the control device quasi uses the global time domain 40 proceeded, there is thus no change between the time domains of the first control device 30th and the first control device 10 instead, but a transition from the time domain of the third control device 30th to the global time domain of the controller 40 and from the global time domain of the controller 40 to the time domain of the first control device 10 . The time master of the clock 41 is additionally qualified or plausibility checked in this way, as described above.

wobei ^tcycleStart_,ModulationEnd vorab ermittelbar ist.Here, t _in denotes the arrival time of the third control device 30th sent data in the first control device 10 , t _meas the measurement time. The superscript indices each define the time domain. Δ _{MRR, DASy, i} denotes the current time offset between the time domains of the first control device 10 and the third control device 30th . A point in time of the end of the cycle in the third control device 30th can be represented as follows: $$t_{\text{cycleEnd}\text{, i}}^{\text{MRR}}=t_{\text{meas},i+1}^{\text{MRR}}-t_{\text{cycleStart}\text{, ModulationEnd}}$$

where ^t cycleStart _, ModulationEnd can be determined in advance.

E () denotes the expected value and amounts to e.g. 5 ms for the above terms.

In cycle i + 1, an estimate of the offset Δ _{MRR, DASy, i} for cycle i is carried out according to the following mathematical relationship, with the proportions d _{i being determinable in} advance or at runtime: $${\Delta }_{\mathrm{M.}\mathrm{R.}\mathrm{R.},\mathrm{D.}\mathrm{A.}\mathrm{S.}y,i}=t_{in,i}^{\mathrm{D.}\mathrm{A.}\mathrm{S.}y}-t_{\text{meas},i+1}^{\text{MRR}}+t_{\text{cycleStart}\text{, ModulationEnd}}-t_{\text{transmission}}-\text{E.}\left(t_{\text{BG}\text{, T10}}\right)-\text{E.}\left(t_{\text{Receive}\text{, T10}}\right)-{\epsilon }_{t,i}$$

stellen zufallsabhängige Anteile dar. Auf Grund der Annahme der Unabhängigkeit von Sende- und Empfangsverzögerung liegt eine Dreiecksverteilung für ${\epsilon }_{t,i}$

vor. Wird über m Zyklen der Durchschnitt M_m gebildet, so ergibt sich eine Bates-Verteilung für ein gerades n=2m.The shares ${\epsilon }_{t,i}$

represent random components. On the basis of the assumption that the transmission and reception delay are independent of the transmission and reception delay, there is a triangular distribution for ${\epsilon }_{t,i}$

in front. If the average M _{m is} formed over m cycles, a Bates distribution results for an even n = 2m.

Insbesondere lässt sich bei einer Mittelung über beispielsweise m=96 Zyklen eine Standardabweichung von $\sigma \left(M_{96}\right)=0.42\text{ ms},$

erreichen. In diesem Fall liegt der durch die Mittelung der Sende- und Empfangsverzögerung noch unkompensierte Fehleranteil in der Offsetschätzung mit einer Konfidenz von > 99,9 % bei unter 1.5 ms.For this applies $\sigma \left({\mathrm{M.}}_m\right)=\frac{1}{\sqrt{\mathrm{12th}\cdot 2m}}\mathrm{20th}\text{ms},$

In particular, when averaging over, for example, m = 96 cycles, a standard deviation of $\sigma \left({\mathrm{M.}}_{96}\right)=0.42\text{ms},$

to reach. In this case, the still uncompensated error portion in the offset estimate by averaging the transmission and reception delay is less than 1.5 ms with a confidence of> 99.9%.

The confidence to adhere to a different error tolerance can also be evaluated at runtime and a suitable value for the variance can therefore be adequately taken into account in the context of the tracking. An evaluation of the confidence for maintaining a different error tolerance is possible, for example, by using an interpolation of a characteristic diagram with support points of the distribution function of the respective Bates distribution.

If necessary, the remaining random error portion can be further reduced by taking into account a larger number of cycles.

If the estimate of the sum of the transmission and reception delay is initialized with 10 ms, an error of <5 ms is already received in the first cycle with a confidence of> 66%.

It goes without saying that all numerical values given are merely exemplary.

Based on the offset estimation, the estimation of the scaling results as a slope value of a linear best-fit line of the offsets over the reception times in the time domain of the first control device 10 away.

It should be noted that the method described above is also based in principle on time synchronization between the second control device 20th with the clock 21st and the first control device 10 is applicable, in which case certain adjustments are necessary. In this case, on the one hand, there is the property of the second control device 20th to have a sufficiently precisely writable computing time per cycle, and on the other hand, it must be taken into account that the data transmission of the second control device 20th takes place via Ethernet, which can lead to the data transmission being severely delayed in individual cycles or even failing.

• In einem Schritt 200 erfolgt ein Synchronisieren einer Uhr 31 einer dritten Steuereinrichtung 30 mit einer Uhr 11 einer ersten Steuereinrichtung 10, wobei ein Ermitteln von Zeitpunkten eines Datentransfers von der dritten Steuereinrichtung 30 zur ersten Steuereinrichtung 10 über eine Steuervorrichtung 40 und einen Datenbus 50 durchgeführt wird.

5 shows a basic sequence of the proposed method:

• In one step 200 a clock is synchronized 31 a third control device 30th with a watch 11 a first control device 10 , wherein a determination of times of a data transfer from the third control device 30th to the first control device 10 via a control device 40 and a data bus 50 is carried out.

In one step 210 a plausibility check of the times relating to the data transfer via the data bus takes place 50 .

In one step 220 an offset is determined between the clock 31 the third control device 30th and the clock of the first controller 10 . The method can preferably be designed as software that runs on the control device 40 and the connected control devices 10 , 20th , 30th is executed and can be stored on a computer readable storage medium. A simple adaptability of the method is advantageously supported in this way.

When implementing the invention, the person skilled in the art will also implement embodiments not explained above.

Claims (11)

Method for determining a time signal for a driver assistance system (100), comprising the steps: - Synchronizing a clock (31) of a third control device (30) with a clock (11) of a first control device (10), comprising the steps: - determining times of a data transfer from the third control device (30) to the first control device (10) via a control device (40) and a data bus (50); - Checking the plausibility of the times relating to the data transfer via the data bus (50); and - Determining an offset between the clock (31) of the third control device (30) and the clock of the first control device (10). Procedure according to Claim 1 , wherein a duration of a modulation phase and a duration of processing the data are taken into account per cycle of the method in a time domain of the first control device (10). Procedure according to Claim 1 , a point in time being determined which can be represented both in the time domain of the third control device (30) and in the time domain of the first control device (10), based on the offset of the clocks (31, 11) taking into account the latency of the data transfer is determined via the data bus (50). Method according to one of the preceding claims, wherein a statistical model is used to determine processing times of the processing chain. Method according to one of the preceding claims, wherein the values are recorded over several cycles and the offset is determined therefrom. Method according to one of the preceding claims, wherein a time signal from a fourth clock (41) of the sensor device (40) is taken into account. Method according to one of the preceding claims, wherein the offset between the clocks (31, 11) of the third control device (30) and the first control device (10) is determined according to the following equation: $${\Delta }_{\mathrm{M.}\mathrm{R.}\mathrm{R.},\mathrm{D.}\mathrm{A.}\mathrm{S.}y,i}=t_{in,i}^{\mathrm{D.}\mathrm{A.}\mathrm{S.}y}-t_{\text{meas},i+1}^{\text{MRR}}+t_{\text{cycleStart}\text{, ModulationEnd}}-t_{\text{transmission}}-\text{E.}\left(t_{\text{BG}\text{, T10}}\right)-\text{E.}\left(t_{\text{Receive}\text{, T10}}\right)-{\epsilon }_{t,i}$$

With: T10 Periodic task in the radar sensor that includes the data transmission t _DASy Time of receipt of the measurement data in DASy in the local DASy time domain t _in ^DASy Time of receipt of the measurement data in DASy in the local DASy time domain t _meas Measurement time in a common time domain t _meas ^MRR Measurement time in a common time domain t _icycleEnd ^MRR Time of completion of the BG cycle in the radar sensor in a common time domain t _{BG, T10} Transmission latency between BG and T10 tasks in the radar sensor t _{Receive, T10} Transmission latency between BG and T10 tasks in DASy t _transmission Transmission latency between the radar sensor and DASy E () Expected value E. Stochastic portion of the respective duration Δ or Δ _{MRR, DASy} Offset of the local DASy time domain (DASy internal time) compared to the common time domain (time of the time master) Index i Current processing cycle Control device (10, 20, 30) for a driver assistance system (100), which is set up to carry out a method according to one of the preceding claims. Driver assistance system (100) of a vehicle, comprising: - a control device (40); - At least two control devices (10, 20, 30) connected to the control device (40); in which - The control devices (10, 20, 30) are connected via the control device (40) by means of a data bus (50); - The driver assistance system (100) being set up to determine an offset between clocks (31, 11) of the at least two control devices (30, 10). Computer program comprising instructions which cause the computer program to be executed by a computer, a method according to one of the Claims 1 to 7th to execute. Machine-readable storage medium on which the computer program is based Claim 10 is stored.

Images