DE102015217742A1 - Method and assistance system for controlling a technical system - Google Patents

Abstract

According to the invention, a user control variable (BS) is read in to control a technical system (TS) and extrapolated for this a trajectory (T0) of the technical system (TS). For extrapolated trajectory (T0) an admissibility is assessed. If the trajectory (T0) extrapolated for the user manipulated variable (BS) is assessed as inadmissible, a manipulated variable variant (V1) with permissible trajectory (T1) is generated and selected as the control signal (SIG) instead of the user manipulated variable (BS). The control signal (SIG) is then output to control the technical system (TS). According to the invention, a distance value (DV) for a distance between the user manipulated variable (BS) and the manipulated variable variant (V1) is determined and output as quantified feedback (QFB) for the user.

Description

Technical systems that are manually controlled by one or more users may result in invalid, unwanted and / or unacceptable conditions due to operator errors or other influences. In order to avoid such conditions, modern technical systems often have an assistance system which, by temporarily assuming the control in critical cases, should prevent a transition to these states. Examples of such assistance systems are, for example, an anti-slip control or a collision avoidance in motor vehicles or a collision avoidance in excavators or cranes.

Frequently, an assistance system completely assumes control when it reaches or shortly before critical conditions are reached. In these cases, a user will have no or limited intervention options during the duration of the takeover. If control is returned to the user at the end of the transfer by the assistance system, however, a jerky behavior of the technical system can occur for the user. Such a jerky behavior can surprise the user and cause a malfunction of the technical system.

It is an object of the present invention to provide a method and an assistance system for controlling a technical system, which are easier to operate for a user.

This object is achieved by a method having the features of patent claim 1, by an assistance system having the features of patent claim 15, by a computer program product having the features of patent claim 16 and by a computer-readable storage medium having the features of patent claim 17.

According to the invention, to control a technical system read in a user control variable and to extrapolate a trajectory of the technical system for controlling the technical system by the user control variable. The user manipulated variable may in this case be e.g. from an analog input device such as a joystick, accelerator pedal, steering wheel, etc. are read. Furthermore, an admissibility of the extrapolated trajectory is assessed. As an evaluation criterion hereby, e.g. a possible or sufficiently probable achievement of a critical system state, for example the possibility of a collision, be applied. If the trajectory extrapolated for the user manipulated variable is evaluated as inadmissible, a manipulated variable variant with a permissible trajectory is generated and this is selected as the control signal instead of the user manipulated variable. The control signal is then output for controlling the technical system. Furthermore, a distance value for a distance between the user manipulated variable and the manipulated variable variant is determined and output as quantified feedback for the user.

An assistance system according to the invention is set up to carry out the above method.

An essential advantage of the invention is the fact that the user can obtain a certain indication of the extent to which a current control action of the user is removed from a control action of the assistance system currently controlling the technical system by the quantified feedback. The user can thereby be prepared for an imminent return of control of the technical system to the user. In this way, as a rule, a surprise effect and the resulting jerky control behavior can be avoided or at least reduced. In addition, an active retrieval of control by the user can be facilitated in that the user can somehow recognize from the quantified feedback how he would need to control to regain control.

Advantageous embodiments and further developments of the invention are specified in the dependent claims.

According to an advantageous embodiment of the invention, a predetermined distance metric can be used to determine the distance between the user control variable and the control variable variant. In this case, the distance metric can also take into account a distance of the trajectories extrapolated from the user manipulated variable and the manipulated variable variant.

Furthermore, the distance value can be output as quantified haptic, acoustic and / or visual feedback. For example, as the distance value proportional restoring force of an input device, as an acoustic signal with proportional to the distance volume and / or pitch and / or as an optical signal with proportional to the distance brightness. Alternatively or additionally, the optical signal can also be varied in color, e.g. by a gradual, proportional to the distance transition from green to red. Furthermore, a bar display with a bar length proportional to the distance value and / or a different distance display may be displayed as visual feedback.

Advantageously, the quantified feedback can indicate whether the user control variable or whether the control variable variant is selected as a control signal for controlling the technical system. In this way, the user can be confirmed whether the technical system is currently controlled by the user or by the assistance system.

Furthermore, depending on the distance value, the control signal can be switched over between the user control variable and the control variable variant. The switching can take place in particular when the distance value falls below a predetermined threshold value. The switching can also be done depending on whether the user control variable or whether the control variable variant is selected as a control signal for controlling the technical system.

According to an advantageous development of the invention, a multiplicity of manipulated variable variants can be generated, for which in each case an associated trajectory of the technical system is extrapolated. In this case, a respective permissibility of the trajectories extrapolated for the plurality of manipulated variable variants can be evaluated. Depending on the respective permissibility of the extrapolated trajectories, a manipulated variable range permissible for the user manipulated variable can then be estimated. Preferably, this estimation is carried out continuously for continuously read or determined input data. The permissible manipulated variable range can be estimated, for example, as a convex envelope of those manipulated variable variants which lead to an allowable trajectory.

Preferably, a further distance value for a distance of the user manipulated variable to a limit of the permissible manipulated variable range can be determined and output as quantified feedback for the user. For determining this distance, advantageously the same distance metric can be used as for the determination of the distance between the user manipulated variable and the manipulated variable variant. Through the quantified feedback of the further distance value, the user can, so to speak, receive an indication of how far his current control action is from taking over control by the assistance system. In this way, the user and the assistance system can be prepared for an impending takeover of control.

Furthermore, a time-dependent function can be generated as a manipulated variable variant. Such a time dependent function may e.g. are generated as a manipulated variable profile and / or as a time-resolved manipulated variable sequence. In this way, even complex control sequences can be taken into account in the extrapolation, evaluation and / or selection of trajectories and / or manipulated variables.

Advantageously, the manipulated variable variant can be selected from a database with frequent control patterns, preferably user-specific control patterns.

Preferably, the reading of the user manipulated variable, the extrapolation of a respective trajectory, the evaluation of the permissibility of a respective trajectory, the generation of a respective manipulated variable variant, the determination of the distance value, the output of the control signal and / or the output of the distance value can be carried out continuously.

According to an advantageous embodiment of the invention, the extrapolation of a respective trajectory can take place on the basis of a system model of the technical system. In particular, a respective trajectory can be extrapolated by simulating a system behavior of the technical system on the basis of the system model.

In addition, an operating parameter and / or environmental parameters of the technical system can be detected. The extrapolation of a respective trajectory can then take place as a function of the acquired operating parameter and / or environmental parameter. As operating parameters or environmental parameters, e.g. a vehicle speed, a road condition, an outside temperature and / or other weather or environmental conditions are recorded and taken into account in the extrapolation.

Furthermore, the extrapolation of a respective trajectory can take place up to an extrapolation limit, which can be determined dynamically as a function of a detected operating parameter and / or environmental parameters.

In particular, the extrapolation limit can be determined in such a way that the technical system can be controlled within the extrapolation limit into a safe system state.

An embodiment of the invention will be explained in more detail with reference to the drawing.

The figure shows a schematic representation of an inventive assistance system when controlling a technical system.

In the figure, an interactive assistance system AS for controlling a technical system TS is shown schematically. The technical system TS can be, for example, a motor vehicle, an excavator, a crane or another system to be controlled by a user. In particular, that can inventive assistance system AS also be part of the technical system TS. To control the technical system TS by a user in the present embodiment, a steering wheel LR is provided as an analog input device. Alternatively or additionally, as an input device and a joystick, an accelerator pedal, a brake pedal, etc. may be provided.

The assistance system AS has a processor PROC for carrying out the method steps of the assistance system AS.

The assistance system AS has an input interface IN to the technical system TS, via which a current user control variable BS is continuously read in by the steering wheel LR. The user manipulated variable BS can be set within a predetermined range of values, e.g. indicate a rotation angle of the steering wheel LR. Alternatively or additionally, the user manipulated variable BS can also be a single-component or multicomponent manipulated variable vector, which can be e.g. read by a multi-axis joystick.

The assistance system AS also has an extrapolation module EP, which continuously receives the user control variable BS from the technical system TS. The extrapolation module EP is used for continuous extrapolation of trajectories of the technical system TS. Such trajectories are trajectories of the technical system TS or a part thereof in a state space of the technical system TS, e.g. a spatial, preferably temporally resolved trajectory of a motor vehicle, an excavator bucket or a crane jib. The extrapolation of the trajectories may be performed spatially, temporally or in another direction of the state space.

The extrapolation module EP calculates and / or simulates a system behavior of the technical system TS or a part thereof for controlling the technical system TS by the user manipulated variable BS and for controlling by other predefined manipulated variables. This means that a system behavior of the technical system TS is determined in each case on the condition that the technical system TS is controlled by the user manipulated variable BS or by a respective other predetermined manipulated variable. Such system behavior may be e.g. relate to a braking distance or an evasive maneuver of a motor vehicle. For extrapolation of the trajectories, the extrapolation module EP has a system model SM of the technical system TS and preferably a simulation module for simulating system behavior on the basis of this system model SM.

Preferably, to calculate the system behavior of the technical system TS, operating parameters and / or environmental parameters PAR of the technical system TS are detected, e.g. a vehicle speed, a road condition, an outside temperature, road wetness and / or other weather or environmental conditions. The operating and / or environmental parameters PAR can be detected by a sensor S of the technical system TS and transmitted to the extrapolation module EP. The extrapolation of the trajectories then takes place depending on the detected operating and / or environmental parameters PAR. In this way, e.g. a braking distance can be calculated as the longer, the higher the vehicle speed. Even rain or slippery roads can be considered in the extrapolation.

For the present embodiment, let T0 be an extrapolated trajectory of the technical system TS in controlling the technical system TS by the user manipulated variable BS.

The assistance system AS furthermore has a variation module VAR, which likewise receives the user control variable BS from the technical system TS. The variation module VAR is used for continuously generating a multiplicity of manipulated variable variants V1,..., VN of the currently read-in user manipulated variable BS. The manipulated variable variants V1,..., VN are generated by the user manipulated variable BS in a value range dependent on the user manipulated variable BS. This value range preferably includes manipulated variable values that are in the vicinity of the user manipulated variable BS with respect to a distance measure, preferably a distance metric M. As a distance metric M, for example, a Euclidean distance in the state space can serve for manipulated variables. The range of values of the manipulated variables V1, ..., VN can e.g. Manipulated variable values within a predetermined maximum distance to the current user manipulated variable BS include. Within this range of values, the manipulated variables V1, ..., VN, e.g. be generated depending on a random process. In addition, the manipulated variable variants V1,..., VN can also be used as time-dependent functions, i. are generated as time-varying manipulated variable functions. Furthermore, manipulated variable variants can be read from a database DB with frequent control patterns, preferably user-specific control patterns. The generated manipulated variable variants V1,..., VN are transmitted by the variation module VAR to the extrapolation module EP.

The extrapolation module EP extrapolates in each case an extrapolated trajectory T1,... Or TN of the technical system TS for a respectively transmitted manipulated variable variant V1,... Or VN for controlling the technical system TS by the respective manipulated variable variant V1, VN.

The trajectories T0, T1,..., TN are each extrapolated by the extrapolation module EP up to a dynamically determined extrapolation limit. Such an extrapolation limit is often referred to as a look-ahead. The extrapolation limit is preferably determined continuously as a function of the acquired operating parameters and / or environmental parameters PAR, specifically in such a way that the technical system TS can be controlled within the extrapolation range into a safe system state. Thus, for example, in a motor vehicle, the braking distance and the braking time increase with increasing speed, so that the extrapolation limit in this case is increased in time and / or space such that the motor vehicle can be brought to a standstill within the extrapolation range.

The extrapolated trajectories T0, T1,..., TN are transmitted by the extrapolation module EP to an evaluation module VAL of the assistance system AS. The evaluation module VAL continuously evaluates an allowance for each of the extrapolated trajectories T0, T1,..., TN on the basis of a predetermined evaluation criterion K. As the evaluation criterion K, e.g. a possible or sufficiently probable achievement of a critical system state, for example a possibility of a collision, be applied. A trajectory would in this case be e.g. permissible if a collision of the technical system TS or parts thereof is unlikely or excluded, preferably under the constraint that no excessive acceleration occurs. As a result of this evaluation, the evaluation module VAL generates admissibility information VI, e.g. indicates which of the trajectories T0, T1, ..., TN are admissible or inadmissible. In this case, those trajectories in which no critical system state occurs within the extrapolation range or a critical system state is unlikely can be taken as permissible.

In addition, an allowable manipulated variable range SB in the state space for manipulated variables is continuously estimated by the evaluation module VAL on the basis of the trajectories T0, T1,..., TN. The estimation takes place in such a way that essentially the manipulated variables with permissible trajectory and, outside the permissible manipulated variable range, the manipulated variables with an impermissible trajectory are located within the permissible manipulated variable range SB. A good approximation for the allowable manipulated variable range SB provides e.g. a convex hull of those manipulated variable variants V1,..., VN which have an allowable trajectory, at least if the region of actually permissible manipulated variables in the state space has, as it were, no holes.

The admissibility information VI and the permissible manipulated variable range SB are transmitted by the evaluation module VAL to a selection module SEL of the assistance system AS. The selection module SEL further receives the user manipulated variable BS from the technical system TS and the manipulated variable variants V1,..., VN from the variation module VAR. By means of the selection module SEL, a control signal SIG is selected from the received manipulated variables BS, V1,..., VN and provided for controlling the technical system TS.

As part of the selection, the selection module SEL continuously checks on the basis of the admissibility information VI whether the trajectory based on the user manipulated variable BS, here T0, is permissible. If so, the user manipulated variable BS is selected as the control signal SIG. In this case, the user therefore controls the technical system TS through his user manipulated variable BS, which is therefore in manual operation. On the other hand, if the user manipulated variable BS leads to an impermissible trajectory T0, the selection module SEL checks on the basis of the admissibility information VI which of the trajectories T1,..., TN are admissible and then selects one of the manipulated variable variants V1,..., VN with permissible Trajectory as control signal SIG. In this case, the technical system TS is controlled by the manipulated variable AS generated and selected manipulated variable variant and is therefore in the assistant mode.

The selection module SEL switches in this way between manual operation and assistance operation.

Alternatively or additionally, in the event that the user manipulated variable BS leads to an impermissible trajectory, as long as new manipulated variable variants are generated and their respective trajectory are checked for admissibility until one of these manipulated variable variants has an allowable trajectory. This manipulated variable variant with permissible trajectory is then selected as control signal SIG.

For the continuous outputting of the selected control signal SIG, the assistance system AS has a control interface CTL to the technical system TS. Via the control interface CTL, the user control variable BS is transmitted by the selection module SEL in manual operation and one of the control variable variants V1,..., VN is transmitted to the technical system TS as a control signal SIG in order to control it.

In the event that the trajectory T0 based on the user manipulated variable BS is assessed as inadmissible, it is assumed by way of example that the manipulated variable variant V1 leads to an allowable trajectory T1 and selects it as the control signal SIG becomes. The technical system TS is thus in the assistance mode. As a result, the selection module SEL determines a distance value DV for the selected manipulated variable variant V1 for a distance between the user manipulated variable BS and the selected manipulated variable V1 with respect to the predetermined distance metric M. The distance metric M can also depend on a distance between the trajectories extrapolated from the manipulated variables , here T0 and T1, be. The distance value DV is continuously determined for the respectively current selected manipulated variable variant. In the assistance mode, the determined distance value DV indicates how far away the current control action of the user is from the control action of the assistance system currently exercising control.

On the other hand, if the technical system is in manual operation, a further distance value DSB is continuously determined by the selection module SEL. The further distance value DSB indicates a distance of the user manipulated variable BS to a limit of the permissible manipulated variable range SB with respect to the distance metric M. The further distance value DSB effectively represents a minimum distance of the current user manipulated variable BS from an impermissible manipulated variable in the state space for manipulated variables. In manual operation, the further distance value DSB thus effectively indicates how far the current control action of the user is from accepting the control the assistance system AS is removed.

The additional distance value DSB can also be determined continuously in the assistance mode and then to some extent indicates how far away the current control action of the user is from a return of the control to the user.

The distance value DV and the further distance value DSB are determined as gradually quantified value specifications with a substantially analog value spectrum, but at least with a multiplicity of value stages, preferably as floating point values.

For the continuous output of the distance value DV and the further distance value DSB, the assistance system AS has an output interface OUT to the technical system TS. The distance value DV and the further distance value DSB are transmitted as quantified feedback QFB for the user from the selection module SEL via the output interface OUT to a feedback device FBE of the technical system TS. The quantified feedback QFB is haptically, acoustically and / or optically fed back to the user by the feedback device FBE in a continuous manner. The quantified feedback QFB is value-related with a substantially analogue value range, but at least with a plurality of value levels. The quantified feedback QFB is preferably proportional to the distance value DV or to the further distance value DSB.

The feedback device FBE, e.g. can be integrated in a dashboard of a motor vehicle, preferably has a visual display, a speaker and / or an actuator coupled to the steering wheel LR. An acoustic signal with a volume and / or pitch proportional to the distance value DV or to the further distance value DSB can be output via the loudspeaker. An optical signal with a brightness proportional to the distance value DV or to the further distance value DSB and / or to the distance value DV or to the further distance value DSB proportional color transition can be output via the optical display. By means of the actuator can - as in the figure indicated by a dotted arrow - a proportional to the distance value DV or the further distance value DSB restoring force on the steering wheel LR be generated as haptic feedback for the user. In addition, the selection module SEL can transmit information to the feedback device FBE, which indicates whether the technical system is in manual operation or in assistance mode. This information may e.g. be reported back to the user via the visual display, the speaker and / or the actuator of the feedback device FBE.

The quantified feedback QFB of the distance value DV in the assistance mode gives the user an indication as to how far his current control action is from a control action of the assistance system AS currently controlling the technical system TS. The user and the assistance system AS can hereby be prepared for an imminent return of control over the technical system GS, which is e.g. then occurs when the distance value DV falls below a predetermined threshold. In addition, an active retrieval of control by the user can be facilitated in that the user can see from the quantified feedback QFB how far away he is from an allowable range of control actions and how he would need to steer to within an allowable range of control actions reach. In this way, as a rule, a surprise effect and the resulting jerky control behavior can be avoided or at least reduced.

The quantified feedback QFB of the further distance value DSB gives the user in manual operation an indication of how far his current control action is from taking over control by the assistance system AS. In this way, the user as well as the assistance system AS can be prepared for an impending takeover of control over the technical system TS, e.g. then takes place when the further distance value DSB falls below a predetermined threshold. As a result, the assistance system AS can carry out preparatory calculations even before the actual takeover of control.

Claims (17)

 Method for controlling a technical system (TS) by a user, wherein a) a user control variable (BS) for controlling the technical system (TS) is read, b) a trajectory (T0) of the technical system (TS) is extrapolated for controlling the technical system (TS) by the user control variable (BS), c) an admissibility of the extrapolated trajectory (T0) is evaluated, and d) if the trajectory (T0) extrapolated for the user manipulated variable (BS) is assessed as inadmissible, - a manipulated variable variant (V1) with permissible trajectory (T1) is generated, - The manipulated variable variant (V1) instead of the user control variable (BS) is selected as a control signal (SIG), and the control signal (SIG) for controlling the technical system (TS) is output, and - A distance value (DV) for a distance between the user control variable (BS) and the control variable variant (V1) is determined and output as a quantified feedback (QFB) for the user. A method according to claim 1, characterized in that a predetermined distance metric (M) is used for determining the distance between the user manipulated variable (BS) and the manipulated variable variant (V1). Method according to one of the preceding claims, characterized in that the distance value (DV) is output as quantified haptic, acoustic and / or optical feedback (QFB). Method according to one of the preceding claims, characterized in that the quantified feedback (QFB) indicates whether the user control variable (BS) or whether the control variable variant (V1) as a control signal (SIG) for controlling the technical system (TS) is selected. Method according to one of the preceding claims, characterized in that, depending on the distance value (DV), the control signal (SIG) is switched between the user control variable (BS) and the control variable variant (V1). Method according to one of the preceding claims, characterized in that - a plurality of manipulated variable variants (V1, ..., VN) is generated, - for the plurality of manipulated variable variants (V1, ..., VN) in each case an associated trajectory (T1, ..., TN) of the technical system (TS) is extrapolated, - a respective permissibility of the trajectories (T1, ..., TN) extrapolated for the multiplicity of manipulated variable variants (V1, ..., VN) is evaluated, and depending on the respective permissibility of the extrapolated trajectories (T1, ..., TN) is estimated for the user manipulated variable (BS) allowable manipulated variable range (SB). A method according to claim 6, characterized in that a further distance value (DSB) for a distance of the user manipulated variable (BS) to a limit of the permissible manipulated variable range (SB) is determined and output as quantified feedback (QFB) for the user. Method according to one of the preceding claims, characterized in that a time-dependent function is generated as a manipulated variable variant (V1, ..., VN). Method according to one of the preceding claims, characterized in that a manipulated variable variant (V1, ..., VN) is selected from a database (DB) with frequent control patterns. Method according to one of the preceding claims, characterized in that the reading of the user control variable (BS), the extrapolation of a respective trajectory (T0, T1, ..., TN), the evaluation of the permissibility of a respective trajectory (T0, T1, ..., TN), the generation of a respective manipulated variable variant (V1, ..., VN), the determination of the distance value (DV), the outputting of the control signal (SIG) and / or the outputting of the distance value (DV) are carried out continuously. Method according to one of the preceding claims, characterized in that the extrapolation of a respective trajectory (T0, T1, ..., TN) takes place on the basis of a system model (SM) of the technical system (TS). Method according to one of the preceding claims, characterized in that an operating parameter and / or environmental parameters (PAR) of the technical system (TS) is detected, and that the extrapolation of a respective trajectory (T0, T1, ..., TN) depends on the detected operating parameters and / or environmental parameters (PAR) takes place. A method according to claim 12, characterized in that the extrapolation of a respective trajectory (T0, T1, ..., TN) up to a Extrapolationsgrenze takes place, which is determined dynamically depending on the detected operating parameters and / or environmental parameters (PAR). A method according to claim 13, characterized in that the extrapolation limit is determined such that the technical system (TS) within the extrapolation limit is controllable in a safe system state. Assistance system (AS) for controlling a technical system (TS), set up to carry out a method according to one of the preceding claims. A computer program product for controlling a technical system (TS) arranged to carry out a method according to any one of claims 1 to 14. Computer-readable storage medium with a computer program product according to claim 16.

Images