Method for a driver assistance application

The present invention relates to a driver assistance system and a method of operating it, with improved fault tolerance and redundancy.

Driver assistance systems such as active cruise control (ACC), overtaking assistant, lane departure warning systems, systems for traffic sign recognition or lighting control, and autonomous brake systems are components of modern vehicles.

In the systems that are currently approved, the driver is always part of the control loop, which means that the driver must remain attentive and intervene immediately if the driver assistance function fails. Such failure can be indicated, for example, by an alarm signal. In the future, it is planned to offer functionality for automated driving, i. e. the driver assistance system must be able to drive the vehicle safely for a predetermined period of time even in the event of a fault (such as a sensor failure). In this case, or if the driver does not immediately respond to a takeover request, at least a reduced driver assistance function must be available until the driver takes over or the driver assistance system is fully functional again.

In view of the above, it is an object of at least one embodiment of this invention to provide a driver assistance system with improved fault tolerance.

This object can be achieved according to at least one embodiment of a method and at least one embodiment of a driver assistance system according to the present invention, with features as set forth herein.

According to the invention, a method for a driver assistance system is provided. The driver assistance system includes at least one sensor unit (S 1 , S 2 , . . . Sx) for detecting the surroundings, which may be designed as a radar, LIDAR and/or camera sensor. The driver assistance system further comprises a first central control unit (SE 1 ) with means for calculating a surroundings model ( 5 ) at a high degree of precision. The surroundings model is based on the data from the at least one sensor unit, preferably a multitude of similar and/or different sensor units (S 1 , S 2 . . . Sx) is provided that have the same or different ranges of detection. Furthermore, the first central control unit is configured to calculate a control variable for a driver assistance application, such as a lane-keeping system, and for transmitting a first control signal ( 3 ) for an actuator control system ( 6 ), e. g. for the steering system of the vehicle, in accordance with the calculated control variable. A second control unit (SE 2 ) is provided and includes means for calculating a second control variable based on data from the at least one sensor unit (S 1 , S 2 , . . . Sx) for a reduced driver assistance application (RF, RF 1 , RF 2 ) and for transmitting a second control signal ( 4 , 4 a , 4 b ) to the actuator control system ( 6 ). Furthermore, the driver assistance system includes an actuator with an actuator control unit. The second control unit is in particular a component of a sensor unit (S 1 , S 2 , . . . Sx) or an actuator control unit.

In the event that the first control signal ( 3 ) is not available, the actuator is triggered

using the second control signal ( 4 , 4 a , 4 b ), or a third control signal for triggering the actuator is calculated based on the last transmitted available value of the first control signal(s) ( 3 ) and on the second control signal ( 4 , 4 b ), or the surroundings model calculated by the first control unit is transmitted to the second control unit (SE 2 ) and a third control signal for a reduced driver assistance application is calculated based on said model. In particular, the transmitted surroundings model can be updated with current data from the at least one sensor unit (S 1 , S 2 , . . . Sx).

In a preferred embodiment of the invention, fewer control cases are taken into account for calculating the second control variable ( 4 , 4 a , 4 b ) for the reduced driver assistance application (RF, RF 1 , RF 2 ) as compared to calculating the control variable for use in the first control unit (SE 1 ). In particular, control interventions are only performed if, for example, a collision is inevitable or the vehicle is at risk to depart from the lane.

It is preferred that no interventions are performed for comfort requests under the reduced driver assistance application (RF, RF 1 , RF 2 ). Comfort requests include, for example, avoidance of abrupt transitions when controlling the longitudinal speed by limited acceleration or avoidance of fast, multiple changes between two control states, e. g. a swinging steering controller.

It is preferred that the calculation of a control variable for a reduced driver assistance application ( 4 , 4 a , 4 b ) is performed in an actuator control unit at least twice, using either data from different sensor units and/or different calculation methods (algorithms), and that a control signal for the actuator is calculated based on these at least two second control variables.

In a positive embodiment of the invention, the at least one sensor unit (S 1 , S 2 , . . . Sx) includes means for preprocessing sensor data. In particular, preprocessing includes object or free space detection. After preprocessing, the sensor data is transmitted to the first control unit (SE 1 ) and completely or partially to the second control unit (SE 2 ).

It is preferred that object or free space information is calculated in a sensor unit according to a first method and a different second method (AL 1 , AL 2 , AL 3 , AL 4 ) for preprocessing. The redundantly generated information is checked for plausibility for use in a surroundings model in the first control unit, and only information based on one of the two methods (e. g. AL 1 and not AL 2 ) is used in a second control unit (SE 2 ) for a reduced driver assistance application (RF, RF 1 , RF 2 ). In particular, one of the methods (AL 1 , AL 4 ) is optimized for the reduced driver assistance application (RF, RF 1 , RF 2 ).

In addition, an embodiment of a driver assistance system includes at least one sensor unit (S 1 , S 2 , . . . Sx) for detecting the surroundings, e. g. a radar, LIDAR, and/or camera sensor system, and a first control unit (SE 1 ) with means:

for calculating a surroundings model ( 5 ) at a high degree of precision based on the data from the at least one sensor unit (S 1 , S 2 , . . . Sx), and for calculating a control variable for a driver assistance application, and for transmitting a first control signal ( 3 ) in accordance with the calculated control variable.

A second control unit (SE 2 ) is provided and includes means for calculating a second control variable based on data from the at least one sensor unit (S 1 , S 2 , . . . Sx) for a reduced driver assistance application (RF, RF 1 , RF 2 ) and for transmitting a second control signal ( 4 , 4 a , 4 b ) to the actuator control system ( 6 ). Furthermore, the driver assistance system includes an actuator with an actuator control unit. The second control unit may in particular be a component of one of the at least one sensor unit (S 1 , S 2 , . . . Sx) or of the actuator control unit. The driver assistance system further includes means for calculating and transmitting a trigger signal for an actuator.

The invention will be explained in more detail below with reference to exemplary embodiments and figures, wherein:

FIG. 1 is a schematic block diagram of a first example embodiment of a driver assistance system configured to perform a first example embodiment of a method according to the invention, and

FIG. 2 is a schematic block diagram similar to FIG. 1 but regarding a second example embodiment.

A preferred embodiment of the invention is shown in FIG. 1 . Preprocessing is performed in the sensor units S 1 , S 2 , S 3 , Sx, for example, an object and/or a free space is detected. A free space is an area which the vehicle can currently drive into, e. g. a road section without obstacles. This data 1 is transmitted to a first control unit. The first control unit SE 1 is the central control unit for driver assistance functions (ADAS-ECU) in this embodiment. The surroundings model 5 , a high-resolution image of the vehicle surroundings, is calculated here using computing-intensive algorithms. The surroundings model may, for example, be designed as an occupancy grid in which merged object data from surrounding objects is stored by various sensor units. The data 2 of the surroundings model is made available to the driver assistance application A 1 , A 2 , Ax, which derives a control variable for a first control signal 3 for an actuator, e. g. for the brake, engine controller, and steering system, from it. The actuator has its own controller, which typically has a high ASIL level. Calculation of the surroundings model 5 and the driver assistance application A 1 , A 2 , Ax are optimized for high comfort, for which also an increased use of resources with respect to computing time is accepted.

To provide a redundant path without a second powerful and expensive controller, the sensor data 1 is also transmitted completely or partially to a second control unit SE 2 , which is embodied or incorporated in the control unit of the actuator in the present illustrated example embodiment. A reduced version of a driver assistance application RF is stored here, which derives a control variable for a second control signal 4 from it. The reduced driver assistance application RF is a minimal version of the driver assistance application, which calculates the second control variable for the second control signal 4 using resources efficiently and without taking comfort requirements into consideration. The first control signal 3 and the second control signal 4 are checked for plausibility 7 and taken into consideration for producing a third control signal 7 a that is provided to the actuator control system 6 . So, if the first control unit SE 1 fails, there is still a second control variable and its associated second control signal 4 available, which can be used to bridge the period of time until the driver takes over or until the full functionality of the first control unit is restored. The limited comfort of the reduced driver assistance application, e. g. a swinging steering controller, can serve as an additional warning to the driver alerting him or her to take control of the vehicle.

In another embodiment, calculation of a multitude of reduced driver assistance applications is provided in the second control unit SE 2 . FIG. 2 shows an example of this case based on two reduced driver assistance applications RF 1 and RF 2 which respectively produce control variables for two second control signals 4 a and 4 b . These can be based on the data from a single sensor or on data from various sensors or on merged data from multiple sensor units. The transmitted control variables of second control signals 4 a and 4 b are checked for plausibility 7 . If the control variables of the second control signals 4 a and 4 b cannot be checked for plausibility, each control variable of the signals 4 a , 4 b may for example be assigned a confidence value and the variable with the higher confidence value be selected for actuator control. The algorithms of the reduced driver assistance applications RF 1 and RF 2 may preferably be different from one another and from the ones of the main application A 1 of the first control unit to be able to detect systematic errors in the analysis of the sensor data.

In a further positive embodiment of the invention, the control variable of the first control signal 3 of the first control unit SE 1 is further processed in the actuator for a plausibility check 7 or fault strategy to bridge a limited period of time after a failure of the first control unit SE 1 . In this way, another source of information is available for a limited period of time in combination with the control variable of the second control signal 4 of the reduced driver assistance application RF in FIG. 1 or the control variables of the second control signals 4 a , 4 b of the reduced driver assistance applications RF 1 , RF 2 in FIG. 2 , to produce a third control signal 7 a that is provided to the actuator control system 6 .

In a preferred embodiment of the invention, the surroundings model 5 is transmitted to the second control unit SE 2 and stored for a predetermined time. If the first control unit SE 1 fails, the stored data can still be used for a specific period of time with a reduced driver assistance functionality. The data provided directly by the sensor units S 1 , S 2 , S 3 , Sx can supplement this history and thus facilitate a somewhat longer bridging time.

It is preferred to have the sensor units run multiple different methods or algorithms AL 1 and AL 2 or AL 3 and AL 4 , e. g. for free space detection, to exclude systematic errors caused by algorithms. FIG. 2 . shows a schematic diagram of this process. The results of both methods or algorithms can be checked against each other for plausibility before they are merged. Furthermore, one of the sensor outputs can be particularly optimized for and additionally provided to a reduced driver assistance application RF 1 , RF 2 in the second control unit SE 2 . In FIG. 2 , these are the methods or algorithms AL 1 and AL 4 . If AL 2 and AL 3 are used in the first control unit SE 1 for the applications A 1 , A 2 , Ax, a completely different path to providing control inputs can be realized via the reduced driver assistance application RF. This path may be optimized for computing efficiency rather than for comfort to save costs for the overall system.