Method and device for operating a vehicle
The method and device dynamically select a safe state for a vehicle based on parameters like traffic, sensor, and hardware status to enhance safety by adapting to specific situations and faults, reducing accident risks.
Patent Information
- Application Number
- DE102014221682
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2014-10-24
- Publication Date
- 2026-11-19
- Estimated Expiration
- 2034-10-24
AI Technical Summary
Existing methods for highly automated vehicles fail to adaptively select the most suitable safe state in response to system faults, particularly at higher speeds and longer braking distances, leading to potential safety risks and inflexible fallback strategies.
A method and device that dynamically select a safe state from a multitude of options based on parameters such as traffic situation, sensor availability, hardware status, and driver health, guiding the vehicle autonomously into the most appropriate safe state.
Enhances safety by allowing the vehicle to react to specific situations and faults, reducing the risk of accidents by selecting the most suitable safe state, such as stopping in a parking space, staying in a lane, or adapting to traffic flow, thus improving both vehicle and surrounding traffic safety.
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
The invention relates to a method for operating a vehicle and a device for operating a vehicle. The invention further relates to a computer program. State of the art In highly automated driving, monitoring the trajectory calculated by the vehicle is essential for safety reasons before it is driven. If such monitoring detects a fault in the system, how the system reacts to this fault is of great importance. EP 2 314 490 B1 discloses a method for a fully automated driver assistance system which, upon reaching a system limit, executes an action plan to bring the vehicle to a safe state, such as a standstill. This plan is also implemented if the driver does not respond to a takeover request and includes automatic driving interventions such as braking to ensure safety. DE102013003216 A1 further discloses a method for lane determination for automated vehicles, wherein two independent systems redundantly and in parallel determine a lane in order to continue the lane in the event of a failure of one system. German patent application DE 10 2012 202175 A1 discloses a method for automatic driving, in particular parking. It is provided that in the event of a fault in the braking system, an emergency operation is carried out in which the vehicle is automatically braked to a standstill using only the drive system and / or the transmission system. The established state of the art is to calculate the emergency trajectory together with the normal trajectory using the main function. This approach has many advantages (in particular, the computing power for the trajectory calculation only needs to be provided once, in the main function). However, this approach also has disadvantages regarding the vehicle's behavior in the event of a system failure. The vehicle would not react to events that occur after the emergency trajectory has been calculated. This is especially true for longer braking distances and higher speeds. Furthermore, stopping on the roadway is not always the "safest state", but always a concession decision based on additional effort and risk. Furthermore, the state of the art specifically includes the fact that there is a fixed strategy for fallback mode, and this strategy is already determined at the time of programming. Disclosure of the invention The object underlying the invention can therefore be seen as providing an improved method for operating a vehicle that makes it possible to safely park the vehicle in various situations in the event of a fault, i.e., when a fault is present. The object underlying the invention can also be seen as providing a suitable device for operating a vehicle. The object underlying the invention can further be seen as creating a corresponding computer program. These problems are solved by means of the respective subject matter of the independent claims. Advantageous embodiments of the invention are the subject of dependent claims. According to one aspect, a method for operating a vehicle is provided, wherein: - the vehicle is driven fully automatically, - wherein, if an error is detected during the fully automated driving, a safe state is selected from a multitude of safe states depending on a parameter, - wherein the vehicle is driven fully automatically into the selected safe state. According to another aspect, a device for operating a vehicle is provided, comprising: - a guidance device configured to guide the vehicle fully automatically, - a detection device for detecting a fault, - a selection device configured to select a safe state from a plurality of safe states upon detection of a fault during fully automated guidance, depending on a parameter, wherein - the guidance device is further configured to guide the vehicle fully automatically into the selected safe state. According to another aspect, a computer program is provided which includes program code for carrying out the procedure for operating a vehicle when the computer program is executed on a computer. The invention therefore encompasses, in particular, the concept of providing a multitude of safe states from which a specific state is selected based on a parameter (or, for example, several parameters, which are preferably identical or different), and from which the vehicle is then automatically guided into this selected safe state. This parameterization thus advantageously enables flexible adaptation to the specific situation in which the vehicle currently finds itself. This is because not every safe state is the most suitable safe state for every situation. For example, the vehicle can advantageously react to events that occur, particularly after a calculation of a runaway trajectory in the event of a fault. This is especially advantageous in the case of longer braking distances and higher speeds. A fully automated driving system as defined in the present invention means, in particular, that the driver of the vehicle no longer needs to intervene to drive the vehicle. The vehicle drives itself, i.e., autonomously. According to one embodiment, the multitude of safe states comprises elements from the following group: stationary on a roadway, stationary in its own lane, stationary in a hard shoulder, stationary in a lay-by, stationary in a parking space, and adapting the vehicle's steering to the steering of a vehicle ahead. This offers the particular technical advantage that a majority of typical situations a vehicle might encounter are covered by these safe states. This allows for a safe state specifically adapted to the situation at hand. Consequently, the safety of the vehicle and the safety of surrounding traffic are advantageously enhanced. In another embodiment, the parameter describes a traffic situation and / or sensor availability and / or hardware availability and / or the health status of the vehicle's driver. This offers the particular technical advantage of covering the most relevant decision criteria for selecting the appropriate safe state. These criteria are especially suitable for choosing the appropriate safe state for the specific situation at hand. This also advantageously improves both the safety of the vehicle and the safety of surrounding traffic. Regarding the traffic situation, it should be noted that the specific traffic situation, including the speed of the vehicle, the position and speed of all other road users, and the respective predictions, is particularly crucial. Examples of traffic situations include: "traffic jam," "highway driving, heavy traffic," "highway driving, slow-moving traffic," "highway driving, light traffic," "stop & go," "city driving," and "rural driving." This means that, for example, an environmental model of the vehicle's surroundings is generated. This is based primarily on data collection from the vehicle's surroundings, such as through the vehicle's environmental sensors. Specifically, the positions, speeds, and / or accelerations of other road users are determined.Accordingly, it is specifically planned that the current traffic situation will be classified according to the aforementioned examples of traffic situations. It should be noted that the aforementioned examples of traffic situations are only examples and not exhaustive. Regarding sensor availability, it should be noted that this primarily concerns which of the vehicle's environmental and inertial sensors are still functioning. For example, if the front environmental sensors fail (i.e., a failure of the front sensors), a different safe state will have to be chosen than if only the rear or side sensors of the vehicle fail. Sensors within the meaning of the present invention are, for example, inertial sensors and / or environmental sensors. An environmental sensor is, for example, a radar sensor, an ultrasonic sensor, a video sensor, a laser sensor, or a lidar sensor. Regarding hardware availability, it should be noted that the hardware can consist of one or more control units in the vehicle. Examples of such control units include: a central control unit that calculates or determines a target trajectory, based on which the vehicle's actual trajectory is controlled; a brake control unit; a steering control unit; a control unit for a monitoring function; and all other control units installed in the vehicle. This means that, depending on which of the aforementioned control units has a fault or has failed (i.e., is available or unavailable), the safe state is selected from the multitude of safe states. According to one embodiment, a monitoring function specifically monitors whether the control units are still active and preferably whether they are still generating meaningful and / or plausible (i.e., not collision-prone) data. This is determined, for example, by checking the resulting trajectory for collision-free behavior based on a redundantly calculated environment model. If the resulting trajectory is collision-free, then the control units are generating meaningful and plausible data. A fault within the meaning of the present invention is, in particular, the occurrence of a fault in at least one of the aforementioned control units and / or a failure of at least one of the aforementioned control units. A fault within the meaning of the present invention is, in particular, a failure or malfunction of one or more of the aforementioned sensors. This means, in particular, that in the event of a malfunction and / or failure of one of the aforementioned sensors and / or one of the aforementioned control units (the plural should always be read as such), a safe state is selected from the multitude of safe states, whereby the vehicle is then guided fully automatically to this selected safe state. The driver's state of health is also a key factor in determining whether or not an error has occurred. For example, if a driver is no longer able to operate the vehicle independently, the vehicle must be brought to a safe state by law. This applies even if the vehicle could still continue driving fully autonomously. A driver's state of health could include, for example, fainting or impaired driving ability. Therefore, driver incapacity or impaired driving ability constitutes an error. In another embodiment, the parameter describes a failure of the vehicle's rear sensors, and the vehicle's steering is adapted to the steering of a vehicle ahead. Adapting the vehicle's steering to the steering of a vehicle ahead means, in particular, that the vehicle moves with the flow of traffic. This driving strategy is especially advantageous on less congested highways with a vehicle ahead and the front sensors still functioning. Adapting the vehicle's steering to the steering of the vehicle ahead includes, in particular, braking or accelerating to the speed of the vehicle ahead, as well as lane keeping, i.e., maintaining a current lane position. According to another embodiment, the parameter describes a driver incapacitation, and the vehicle is parked in a lay-by or parking space. This strategy is particularly advantageous when a driver monitoring system reports that the driver is experiencing health problems. In this situation, the highly automated vehicle remains fully functional, but is prohibited from continuing its journey due to legal regulations. In another embodiment, the parameter describes a failure of the vehicle's electronic stability program, causing the vehicle to stop in its own lane. This is a particularly easy-to-implement driving strategy. It can be used especially when the vehicle (and, for example, all surrounding vehicles) is traveling at a relatively high speed and maintains small distances between them. This is the case, for example, in traffic jams on curves. To avoid endangering other road users with unforeseen braking maneuvers, this state is preferably only activated when vehicles are detected behind the vehicle (the "ego" vehicle) that are not significantly faster than the vehicle itself (maximum distance: preferably 2-3 times the braking distance during emergency braking).These vehicles then preferably maintain the aforementioned small distance, allowing them to see the action of the vehicle in front (the "ego" vehicle) and brake accordingly. This situation is most commonly found in traffic jams. For example, if the electronic stability program (ESP) fails, but the EPS steering is still functional, remaining stationary within the vehicle's lane will be selected as the safest state. If the vehicle has a functioning redundant steering system, this state (i.e., remaining stationary within its lane) will, in another embodiment, also be selected (as the safest state) even if both the brakes and the steering have failed. In another embodiment, the parameter describes a failure of the vehicle's electronic stability program and steering system, and the vehicle is parked on the road. If not only the ESP but also the steering fails, the vehicle can usually only be brought to a standstill using a redundant braking system. Therefore, it is particularly important not to attempt any further steering maneuvers. Essentially, there are four main situations: 1. The vehicle should pull into a parking space or change lanes to the right edge of the road. 2. The vehicle should be kept in its lane. 3. The vehicle should come to a complete stop. 4. The vehicle should continue with the flow of traffic. It depends primarily on the severity of the failure. If the ESP fails, braking can usually only be achieved using the redundant brakes, which means stopping as quickly as possible. If the steering fails, there is also no alternative but to come to a complete stop as quickly as possible. In the event of a sensor failure, the first option (parking in a bay / changing lanes to the right edge of the road) or the second option (staying in one's lane) is preferred. The invention is explained in more detail below with reference to preferred embodiments. Figure 1 shows a flowchart of a method for operating a vehicle, Figure 2 shows a device for operating a vehicle, and Figure 3 shows a cascade of safe states. Fig. 1 shows a flowchart of a procedure for operating a vehicle. According to step 101, the vehicle is driven fully automatically. If, according to step 103, an error is detected during fully automated driving, then, according to step 105, a safe state is selected from a multitude of safe states depending on a parameter. A fault is, for example, a failure and / or malfunction of one or more of the vehicle's control units and / or a failure and / or malfunction of one or more of the vehicle's sensors, such as environmental sensors and / or inertial sensors. A fault is, for example, a driver's incapacity. A fault is, for example, a driver's impaired driving ability. In particular, multiple faults are detected. In step 107, the vehicle is then guided fully automatically into the selected safe state. Fig. 2 shows a device 201 for operating a vehicle (not shown). The device 201 comprises a guidance unit 203 configured to guide the vehicle fully automatically. For this purpose, the guidance unit is in operative connection with actuators and / or controllers of the vehicle. In particular, the guidance unit 203 is in operative connection with a braking system, a steering system and / or a drive system of the vehicle. The device 201 further comprises a detection device 205 for detecting a fault. A selection device 207 is also provided, which is configured, upon detection of a fault by the detection device 205 during fully automated guidance by the guidance device 203, to select a safe state from a plurality of safe states depending on a parameter. The guidance device 203 is configured to guide the vehicle fully automatically into the safe state selected by the selection device 207. The plurality of safe states are, for example, stored in a memory. According to the invention, it is therefore provided in particular that the most suitable safe state for the specific situation in which the vehicle currently finds itself is selected from a multitude of safe states. Based on this selected safe state, the vehicle is then guided into this safe state fully automatically. Guiding the vehicle into this safe state includes, in particular, guiding the vehicle based on a predetermined strategy, which can also be referred to, for example, as a fallback strategy. The invention thus provides numerous defined fallback strategies and selects the most appropriate one online (and preferably anew at every point during the automated journey, i.e., continuously) based on decision criteria (described by the parameter(s)). Thus, it is advantageously possible to automatically determine the safe state from a pool of predefined and technically feasible safe states at any time, depending on the situation, a specific system state of the vehicle, and a detected malfunction. According to the invention, preferably processes and strategies are defined to select the appropriate fallback mode at any given time. This enables the vehicle, which can also be described as an autonomous or fully automated vehicle, to reach a safe state in the event of a fault without driver intervention. The definition of the safe state depends significantly on the current situation of the vehicle.According to the invention, it is therefore particularly provided that at any given time, depending on the situation (described by the parameter), the system state and the detected malfunctions, the safe state is automatically selected. For example, the following safe states can be defined: 1) Stationary on the roadway 2) Stationary in one's own lane 3) Stationary in the hard shoulder 4) Stationary in the lay-by / parking area 5) Continuing to drive "going with the flow of traffic" (i.e., adapting the driving strategy of one's own vehicle to the vehicle in front) The selection of the safe state depends, for example, on the following assessments: Traffic situation Here, for example, the specific traffic situation is crucial, including the speed of one's own vehicle, the position and speed of all other road users, and the respective prediction. Examples would be "traffic jam", "highway driving, heavy traffic", "highway driving, slow-moving traffic", "highway driving, light traffic", "stop & go", "city driving", "rural driving". Sensor availability The key question here is which environmental and inertial sensors are still available. For example, if the front environmental sensors fail, the system will have to choose a different safe state than if only the rear or side sensors fail. Hardware availability This refers in particular to the control units in the vehicle, including, for example, the central control unit on which calculations are carried out, the brake control unit, the steering control unit, the control unit of the monitoring function and / or all other control units installed in the vehicle. Figure 3 shows a cascade of safe states for which, for example, given availability levels can be defined. Different escalation levels are shown. The least critical level is continued driving (if, for example, a rear sensor fails), but the fallback mode can escalate further (i.e., the subsequent states are approached depending on the situation) if, for example, the steering or the driver also fails. Therefore, the term "cascade" is used in this context. "Given availability levels" means that, depending on which vehicle components (sensors, control units, etc.) are still functional, a specific fallback strategy is used. If, according to block 301, a failure of a rear sensor system occurs, then, for example, continuing to drive in accordance with the flow of traffic according to block 303 is provided. If, according to block 305, a driver failure of the vehicle is detected, then according to block 307, the vehicle is to be parked in a lay-by or in a parking space (standstill in lay-by or in parking space). If, according to block 309, a failure of an ESP is detected, then according to block 311 it is provided that the vehicle is parked in its own lane (standstill in its own lane). If, according to block 313, it is detected that both the ESP and the steering of the vehicle have failed, then, according to block 315, the vehicle is to be parked on the roadway (stationary stop on the roadway). An exemplary cascade of fallback strategies might look like this: 1) Fully redundant driving: A redundantly designed sensor / control unit fails. In certain situations, continuing to drive with primary sensors and without redundancy, possibly at reduced speed and with reduced functionality, is the safest option. These situations are primarily those where a different strategy would increase the risk of an accident. For example, braking in the lane when the vehicle is in the left lane and there is no vehicle behind it. In this case, a vehicle accelerating quickly from behind could potentially rear-end the vehicle. Reduced functionality: No lane changes, no overtaking maneuvers, "going with the flow." Reduced speed to achieve a shorter braking distance and less anticipation required.The maximum permissible braking distance depends on the actual remaining functional performance (e.g., the predictive capability of the environment model). 2) Assume the speed of the vehicle in front and follow the lane: This strategy involves braking or accelerating to the speed of the vehicle in front and maintaining lane position. This strategy is particularly advantageous on lightly trafficked highways with a vehicle in front and available front sensors. For example, in the event of a rear sensor failure. "Lightly trafficked" means, in particular, that the distances to the vehicles in front or behind are significantly greater (>>) than the braking distance. 3) Active lane changes onto the hard shoulder: This strategy will be used, for example, if a driver observation reports that the driver is experiencing health problems.The highly automated vehicle is still fully functional in this situation, but due to legal regulations, it is no longer permitted to continue driving. 4) Slow drifting towards the hard shoulder or shoulder: This strategy comprises two steps. In the first step, acceleration ceases. In the second step, the vehicle attempts to drift slowly to the right (towards the hard shoulder). "Slowly" depends in particular on how much steering torque can still be applied. Here, "slow" preferably means that the dynamics have no significant influence on the steering control. This strategy is particularly advantageous in situations with few road users, low differential speeds, and high absolute speeds.5) Braking in your own lane: This is the simplest strategy and can be used particularly when the vehicle (and all surrounding vehicles) is traveling somewhat faster and the distances between them are small. This is the case, for example, in traffic jams on curves. To avoid endangering other road users with unexpected braking maneuvers, this mode should only be activated if vehicles are detected behind the vehicle that are not significantly faster (for example, a maximum of 10 km / h to 20 km / h faster) than the vehicle (the "ego" vehicle) (maximum distance 2-3 times the braking distance during emergency braking). These vehicles should then be close enough so that they can detect the action of the "ego" vehicle and brake accordingly. This situation is most commonly found in traffic jams.If the ESP fails, but the ESP steering is still functional, remaining stationary within the vehicle's lane can be considered the safest state, selected from among the many possible safe states. If the vehicle has a functioning redundant steering system, this would also be the safest state even if both the brakes and the steering were to fail. 6) (Straight ahead) Braking: If not only the ESP but also the steering fails, the vehicle can only be brought to a standstill using the redundant brakes. According to the invention, the choice of the fallback strategy therefore depends on the components installed in the vehicle (for example, car) that are functioning or have failed, as well as on the traffic situation.
Claims
Method for operating a vehicle, wherein: - the vehicle is driven fully automatically (101), - wherein, if a fault is detected (103) during fully automated driving, a safe state is selected from a plurality of safe states depending on at least one parameter (105), - wherein the parameter describes a traffic situation and / or sensor availability and / or hardware availability and / or a health status of a driver of the vehicle, - wherein the vehicle is driven fully automatically into the selected safe state (107); characterized in that: - if the parameter describes a failure of a rear sensor system of the vehicle, the driving of the vehicle into the safe state is carried out by adapting the driving of the vehicle to the driving of a vehicle driving ahead of the vehicle. Method according to claim 1, wherein, if a further parameter describes a failure of the driver, the vehicle is guided into a safe state by parking the vehicle in a lay-by or in a parking space. Method according to claim 1, wherein, if a further parameter describes a failure of an electronic stability program of the vehicle, the vehicle is guided into a safe state by parking the vehicle in its own lane. Method according to claim 1, wherein, if a further parameter describes a failure of an electronic stability program of the vehicle and a vehicle steering system, the vehicle is guided into a safe state by parking the vehicle on the roadway. Device (201) for operating a vehicle, comprising: - a guidance device (203) configured to guide the vehicle fully automatically, - a detection device (205) for detecting a fault, - a selection device (207) configured to select a safe state from a plurality of safe states upon detection of a fault during fully automated guidance, depending on at least one parameter, - wherein the parameter describes a traffic situation and / or sensor availability and / or hardware availability and / or the health status of a driver of the vehicle, - wherein the guidance device (203) is further configured to guide the vehicle fully automatically into the selected safe state;characterized in that, if the parameter describes a failure of a rear sensor system of the vehicle, the vehicle is guided into a safe state by adapting the vehicle's guidance to that of a vehicle driving ahead of it. Computer program comprising program code for carrying out the method according to any one of claims 1 to 4, when the computer program is executed on a computer.
Citation Information
Patent Citations
Methods for automatically driving, in particular parking, a motor vehicle and driver assistance devices
DE102012202175A1
Method for determining traffic lane for steering control of automatically controlled vehicle by using traffic lane determining system, involves balancing process data among each other for determining traffic lanes in continuous manner
DE102013003216A1
Method for controlling the operation of a fully automatic driver assistance system of a motor vehicle for independent vehicle guidance and motor vehicle
EP2314490B1
