Method for determining the permissible vehicle state space for an articulated vehicle

By monitoring and predicting the travelable area and potential sweep area of ​​the articulated vehicle, determining the allowable vehicle status space, the problem of the vehicle being unable to maneuver safely under interruption of sensor signals is solved, and the safety of maneuver and control accuracy is improved.

CN113661112BInactive Publication Date: 2024-11-12VOLVO TRUCK CORP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN201980095281.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2019-04-12
Publication Date
2024-11-12
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

In the event of interruption of sensor signals, it is difficult for articulated vehicles to perform maneuvers safely, such as safe parking maneuvers, especially since the vehicle may be in an unsubstantiable state, resulting in the inability to safely control the vehicle speed and avoid overturning.

Method used

By monitoring the travelable area in front of the vehicle and predicting potential sweep areas, based on the initial vehicle state, geometry and sensor error characteristics, the allowable vehicle state space is determined so that the vehicle does not exceed the travelable area during maneuvering and limiting the vehicle. Speed ​​and steering wheel angle for safety.

Benefits of technology

It is possible to safely perform maneuvers when sensor signals are interrupted, reducing the risk of vehicle leaving the travelable area and overturning, and improving the accuracy and safety of vehicle control.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113661112B_ABST
    Figure CN113661112B_ABST
Patent Text Reader

Abstract

A method for determining an allowable vehicle state space of an articulated vehicle (1) for safely completing a maneuver, wherein the vehicle state space includes vehicle speed, the method comprising: monitoring a drivable area (31) in front of the articulated vehicle (1), predicting a potential swept area (33) in front of the articulated vehicle (1), wherein the potential swept area (33) represents an area that the articulated vehicle (1) traverses during the maneuver with a probability based on an initial vehicle state of the articulated vehicle (1), a geometry of the articulated vehicle (1), and error characteristics associated with one or more sensor input signals used to position the articulated vehicle (1), and determining the allowable vehicle state space such that the predicted potential swept area (33) does not extend beyond the drivable area (31) in front of the vehicle (1).
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to a method, a control unit and a vehicle for determining an allowable vehicle state space, eg a speed limit, for an articulated vehicle, such as a truck and a semi-trailer.

[0002] The invention may be applied to heavy vehicles, such as trucks and construction equipment.Although the invention will be described primarily with respect to a semi-trailer, the invention is not limited to this particular vehicle, but may also be used with other types of articulated vehicles, such as recreational vehicles. Background Art

[0003] Autonomous and semi-autonomous vehicles use various types of sensor input signals for navigation and vehicle control. Advanced driver assistance systems (ADAS) are also based on sensor input signals. If certain critical sensor signals necessary for safe operation of the vehicle are lost, a safety stop maneuver may be required. A safety stop maneuver may include, for example, maneuvering the vehicle to the side of the road and stopping the vehicle in a controlled manner. Because the vehicle may be at least partially "blind" during a safety stop maneuver, control must sometimes be performed based on alternative input signals. One such alternative input signal is dead reckoning, where wheel rotations are used to estimate, for example, distance traveled, speed, and acceleration.

[0004] US2018 / 0224851A1 is concerned with the problem of performing a safe parking maneuver in the event of a loss of GPS signal. Dead reckoning-based position estimation is used in the event of a failure of the primary positioning system.

[0005] US2010 / 0106356 relates to autonomous vehicle control and to sensor systems for autonomous vehicle control. Part of this disclosure relates to loss of sensor input. For example, a mechanism is disclosed for simulating the position of an obstacle in the event that a radar signal is lost due to obstruction. The disclosure also relates to performing a safe parking maneuver in situations where safe operation cannot be guaranteed.

[0006] A problem associated with performing a safe parking maneuver is that the vehicle may be in a state that does not allow the stopping maneuver to be performed in a safe manner. For example, the vehicle may be traveling too fast.

[0007] Other maneuvers requiring sudden and critical vehicle control include sudden obstacle avoidance maneuvers, maneuvers to cope with skids, etc. Similar to safety parking maneuvers, the vehicle may be in a state that does not allow for, for example, skid compensation to be performed in a safe manner because the speed is too high.

[0008] There is a need for a method of determining the permissible vehicle state space for an articulated vehicle such that a safe parking maneuver can always be safely performed in the event of a sensor signal interruption. Summary of the invention

[0009] The object of the present disclosure is to provide a technology that alleviates or overcomes the above-mentioned problems. This object is achieved at least in part by a method for determining an allowable vehicle state space for safely completing a maneuver of an articulated vehicle, wherein the vehicle state space includes vehicle speed. The method includes monitoring a drivable area in front of the articulated vehicle. The method also includes predicting a potential swept area in front of the articulated vehicle. The potential swept area represents an area that the articulated vehicle is likely to pass during maneuvering. The potential swept area is determined based on an initial vehicle state, the geometry of the articulated vehicle, and error characteristics associated with one or more sensor input signals used to locate the vehicle. The method also includes determining an allowable vehicle state space so that the potential swept area predicted during maneuvering does not exceed the drivable area in front of the vehicle.

[0010] This ensures that the maneuver can be performed safely in the current driving situation. For example, some scenarios may allow for relatively high driving speeds, while other scenarios present a higher control challenge and therefore result in a reduced vehicle speed. The geometry of articulated vehicles is taken into account when determining the possible traversal area, which ensures that, for example, a trailer does not leave the drivable area during the maneuver. The approach takes into account different types of sensor error characteristics, which is an advantage because it is not possible to know in advance which sensor systems will experience disruptions.

[0011] According to some aspects, the maneuver is a safety parking maneuver and the one or more sensor signals correspond to a reduced set of sensor input signals compared to a normal operating mode.The possibility to perform a safety parking maneuver is crucial for, for example, autonomously driven vehicles and for more advanced driver support systems.

[0012] According to some aspects, monitoring includes obtaining information related to traffic infrastructure in front of the vehicle. Information about traffic infrastructure can be obtained in a number of different ways, which will be explained in detail below. This provides a robust system, which is an advantage. For example, optionally, monitoring includes data related to the environment in front of the vehicle obtained from any of: one or more radar sensors, one or more lidar sensors, one or more vision sensors, and one or more V2X communication transceivers.

[0013] According to some aspects, the prediction includes predicting a potential swept area ahead of the articulated vehicle based on a planned trajectory of the vehicle for the maneuver. The planned trajectory may be maintained and updated as the vehicle passes over the traffic infrastructure. By basing the potential swept area on the currently planned trajectory of the maneuver, a more accurate estimate of the potential swept area is obtained. Furthermore, advantageously, because the area traversed by the articulated vehicle may depend on the vehicle speed and turn rate during the maneuver, the permissible vehicle state space optionally includes upper and lower limits on the vehicle speed during the maneuver.

[0014] According to some aspects, the prediction includes obtaining an estimate of an initial vehicle state error associated with the vehicle. The state error has an impact on the determined potential swept area. Therefore, advantageously, a more accurate estimate of the potential swept area is obtained.

[0015] According to some aspects, predicting includes estimating a covariance matrix associated with positioning errors of the vehicle during the maneuver. The covariance matrix provides an efficient method of low complexity for representing uncertainty regions, which is an advantage.

[0016] According to some aspects, the vehicle state space includes a steering wheel angle, and the method includes determining the permissible vehicle state space based also on a rollover risk associated with a planned trajectory of the vehicle for the maneuver. Thus, the method not only limits the risk of the articulated vehicle passing outside the drivable area, but also limits the risk of rollover during the maneuver.

[0017] According to some such aspects, the method includes obtaining one or more dynamic properties associated with the vehicle, and determining the permissible vehicle state space is also based on the vehicle dynamic properties. In this way, the method can take into account different vehicle dynamic properties, thereby performing more precise control during maneuvers. For example, a top-heavy vehicle may be subject to more stringent restrictions on steering wheel angles than a more stable vehicle.

[0018] According to some aspects, the method includes providing the determined permissible vehicle state space as an output to a control unit of the vehicle for control during a maneuver.Thus, advantageously, vehicle control may be improved based on new input from the disclosed method.

[0019] According to some aspects, the method includes providing an estimate of the potential swept area and / or rollover probability during maneuvering as an output to a control unit of the vehicle for control during maneuvering. Again, advantageously, based on new inputs from the disclosed method, control of the vehicle by the control unit can be improved.

[0020] According to some aspects, the method includes determining a plurality of potential swept areas corresponding to different sets of sensor input signal combinations. Different scenarios and conditions may produce different potential swept areas. For example, certain combinations of available sensor signals may produce smaller potential swept areas than other less accurate sensor signals. The method may be used to maintain a plurality of different fault hypotheses with associated probabilities, thereby generating inputs for more advanced safety systems in articulated vehicles.

[0021] According to some aspects, the method includes triggering a warning signal based on an estimated uncertainty associated with an estimated state of the articulated vehicle. Some scenarios and potential swept areas may be associated with a significant risk of, for example, not being able to complete a maneuver safely. In such cases, the vehicle may be prepared for a collision or other critical event. The warning signal may, for example, be a warning signal issued to a control unit in the articulated vehicle and / or a warning signal issued to nearby vehicles and infrastructure objects via V2X communication.

[0022] A control unit, a computer program, a computer readable medium, a computer program product and a vehicle associated with the above advantages are also disclosed herein.

[0023] In general, all terms used in the claims should be interpreted according to their ordinary meaning in the technical field, unless otherwise clearly defined herein. All references to "one / an / element, device, part, means, step, etc." will be publicly interpreted as referring to at least one instance of an element, device, part, device, step, etc. unless otherwise clearly stated. Unless explicitly stated, the steps of any method disclosed herein do not have to be performed in the exact order of the present disclosure. Further features and advantages of the present invention will become apparent when studying the attached claims and the following description. Those skilled in the art recognize that, without departing from the scope of the present invention, different features of the present invention can be combined to create embodiments other than those described below. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] With reference to the accompanying drawings, the following is a detailed description of an embodiment of the present invention cited as an example. In these drawings:

[0025] Figure 1 A vehicle for cargo transportation is schematically shown;

[0026] Figure 2 A vehicle with a swept area is shown;

[0027] Figure 3 to Figure 4 The potential sweep area during the maneuver is schematically shown;

[0028] Figure 5a , Figure 5b , Figure 6 , Figure 7a and Figure 7b An example of determining a potential sweep area is shown;

[0029] Figure 8 is a flow chart illustrating a method;

[0030] Fig. 9 A sensor unit and / or a control unit are schematically shown; and

[0031] Fig.10 An example computer program product is shown. DETAILED DESCRIPTION

[0032] The present invention will now be described more fully hereinafter with reference to the accompanying drawings, in which certain aspects of the invention are shown. However, the present invention may be embodied in many different forms and should not be construed as limited to the embodiments and aspects set forth herein; rather, these embodiments are provided by way of example so that this disclosure will be thorough and complete and will fully convey the scope of the invention to those skilled in the art. Throughout the description, like numbers represent like elements.

[0033] It should be understood that the present invention is not limited to the embodiments described herein and shown in the drawings; instead, the skilled person will recognize that many changes and modifications are possible within the scope of the appended claims.

[0034] Figure 1 A schematic articulated vehicle combination 1 is shown, comprising a towing vehicle 2 and two towed vehicles 3, 4. The towing vehicle may be a conventional truck suitable for commercial highway use or a tractor with a fifth wheel, but may also be an off-road truck, a bus or a recreational vehicle. In the example shown, the first towed vehicle or trailer 3 is a dolly with a drawbar connected to the trailer coupler of the truck. The dolly is provided with two wheel axles 7. The second towed vehicle or trailer 4 is a semitrailer provided with a kingpin 8 connected to the fifth wheel of the dolly. The example shows a common type of longer vehicle combination, but other types of vehicle combinations with other types of towing vehicles and other types and numbers of towed vehicles may also be used. Different vehicle combinations may include a truck with a conventional trailer, a truck with a mid-axle trailer, a truck with a dolly and a semitrailer, a tractor with a semitrailer, a tractor with a B-link and a semitrailer, a tractor with a B-link and a semitrailer, a tractor with a semitrailer and a conventional trailer, or a tractor with a semitrailer dolly and a semitrailer.

[0035] The towing vehicle may be provided with various autonomous or semi-autonomous driving functions, such as an automatic reverse assist function, wherein the steering of the vehicle combination is automated during reverse maneuvering, while the speed of the vehicle combination may be controlled by the driver.

[0036] In the vehicle combination shown, the effective wheelbase Leq1 of the towing vehicle (i.e., the truck) is the length from the front axle 12 of the truck to the virtual axle 13. The effective wheelbase Leq2 of the first towed vehicle (i.e., the car) is the length from the towbar connection of the car to the virtual axle 6. The effective wheelbase Leq3 of the second towed trailer extends from the kingpin 8 to the virtual rear axle 9 of the trailer 4.

[0037] Based on the vehicle geometry and the driving path, a swept area can be determined. The swept area represents the area swept by the articulated vehicle as it moves over the drivable area. For example, when the towing truck turns along one path, the trailer will move along a slightly different path. Details and examples on determining the swept area for different types of vehicle configurations are disclosed in US 9,862,413 B2 and will therefore not be discussed in more detail here.

[0038] Figure 2 An articulated vehicle 1 is shown having a modified swept area 20 of the rear portion of the travel path 23. Figure 2 In FIG. 2 , the modified swept area 20 consists of a swept area 21 and a tolerance zone 22, one on each side of the swept area 21. The purpose of the tolerance zone is to compensate for errors in the steering actuator and tolerances and noise in the measurements from the different sensors.

[0039] Autonomous and semi-autonomous articulated vehicles rely on sensor signal inputs to control the vehicle, whether or not there is a driver. Sensor systems arranged on the vehicle to support autonomous functions may include any of the following; radio detection and ranging (radar) sensors, light detection and ranging (lidar) sensors, vision-based sensors (such as cameras), and global positioning system (GPS) receivers. These sensors monitor the vehicle's surroundings to detect obstacles and determine, for example, the geometry of the drivable area in front of the vehicle. The vehicle may also include multiple on-board sensors, such as steering angle sensors, articulation angle sensors, i.e. sensors that measure the angle between the towing truck and the trailer, wheel speed sensors, and inertial measurement units (IMUs).

[0040] In the event that the vehicle loses sensor input from one or more sensor systems, a critical situation may result. For example, radar data sensors and lidar data sensors may fail, or the processing unit responsible for processing sensor signal data may experience an interruption. If critical sensor signals are lost, or certain critical operations are otherwise obstructed, an automatic maneuver may be required, such as a safe parking maneuver. This operation can then be performed based on control using the on-board sensor systems. That is, dead reckoning using wheel speed sensors and steering wheel angle sensors.

[0041] However, such maneuvers may not be performed in a safe manner if, for example, the vehicle speed is too high. In addition, during the maneuver, it may be necessary to limit the vehicle state to prevent the vehicle from tipping over or skidding.

[0042] In order to provide a solution to these problems, this paper proposes to continuously monitor the drivable area in front of the articulated vehicle, that is, to always know what the road ahead on which the articulated vehicle is traveling looks like. Then, it is proposed to predict the potential swept area in front of the articulated vehicle. The potential swept area represents the area that the articulated vehicle may pass through during the upcoming maneuver, depending on sensor errors, etc. As will be explained below, the potential swept area can be determined based on the initial vehicle state, the geometry of the articulated vehicle, and the error characteristics associated with one or more sensor input signals used to locate the vehicle. Given the drivable area and the potential swept area, the allowable vehicle state space can be determined so that the predicted potential swept area does not exceed the drivable area in front of the vehicle and so that the maneuver can be successfully completed with a sufficiently high probability.

[0043] In this context, a drivable area is an area in which a vehicle can be located without risk of damage to the vehicle or injury to the driver. The road itself is of course a drivable area. However, depending on the circumstances, the shoulder of the road may be included in the drivable area, and areas on the side of the road may also be included in the drivable area depending on the circumstances and the type of vehicle. For example, if a relatively flat field extends to one side of the road without a ditch in the middle, then the flat field may be considered a drivable area because an articulated vehicle can temporarily drive into the field without risk of damage to the vehicle or the vehicle occupants. Off-road vehicles such as dump trucks are likely to be associated with a larger drivable area than ordinary cargo transport vehicles.

[0044] The vehicle state is a collection of variables that together describe what state the vehicle is currently in. Here, the vehicle state includes variables associated with the vehicle's position (coordinates) and orientation (such as, for example, heading, steering angle, and articulation angle). The vehicle state also includes information associated with the vehicle's dynamic state, i.e., vehicle speed, acceleration, turning rate, etc. The vehicle state is typically represented as a vector of state variables. As will be discussed in more detail below, the allowable vehicle state space can typically include upper and lower limits on state variables (such as speed) during the duration of the maneuver. In addition, some state variables will have different limits depending on other variables. For example, the speed in some directions may have different limits compared to the speed in other directions.

[0045] The state space is an N-dimensional space representing the range of vehicle states. Physics limits the state space depending on the type of vehicle, for example, by a maximum achievable speed. The control unit can impose further constraints on the state space, thereby limiting the vehicle state space to some range of values.

[0046] The vehicle state can be estimated via a vehicle state estimate. This vehicle state estimate is associated with a vehicle state uncertainty that describes the accuracy in the vehicle state estimate. A perfect state estimate would be exactly consistent with the "true" vehicle state, while a more realistic vehicle state estimate would differ from the true state according to some error distribution. A common way to represent vehicle state uncertainty is through a covariance matrix to represent the expected error variation.

[0047] A potential swept area is a statistically determined area associated with a certain confidence level. This confidence level may be determined depending on the situation and the type of maneuver. Some examples of potential swept areas and how to determine them are given below. However, for example, the area may correspond to an estimated area that delimits the swept area of ​​the articulated vehicle during the maneuver with a certain probability (e.g., 99.999%).

[0048] The size of the potential swept area depends on a number of key factors. One of the key factors is the initial speed of the articulated vehicle when the maneuver is initiated. For example, when initiating a safety stop maneuver. This is at least in part because the initial vehicle speed determines how long it will take to stop the vehicle given a target speed profile to follow during the maneuver, and at least in part because some errors (such as steering wheel angle errors) are amplified by speed.

[0049] Another key factor is the actual target speed profile during the maneuver, i.e., the planned speeds for different parts of the maneuver trajectory. This is because the trailer behaves differently with respect to the swept area at different speeds. Therefore, different speed profiles may produce different swept areas, even if the planned trajectory path remains the same.

[0050] The vehicle speed affects the swept area in different ways, depending on the trajectory path and target speed profile followed during the maneuver. For a slow moving articulated vehicle 1, as discussed in US 9,862,413 B2, the trailers 3, 4 will cut inwards into the constant radius turn, thereby expanding the swept area inwards, i.e., towards the center of the circular segment representing the turn. For a faster moving articulated vehicle, the swept area is increased by a y =v x 2 / R gives the lateral or centripetal acceleration, for the longitudinal velocity v x and a constant turning radius R may be sufficient to cause the trailer to skid, with the effect instead extending outwards over the swept area, i.e. away from the centre of the circular segment representing the turn. y and longitudinal velocity v x exist Figure 3 Indicated schematically in .

[0051] For an articulated vehicle 1 that is not performing a turn or where R is very large (and the vehicle has not suffered a jacking event), the swept area will not extend beyond the area covered by the towing vehicle, as the trailer will follow the same trajectory as the towing vehicle.

[0052] Thus, the area swept by the articulated vehicle 1 generally increases at low speeds, i.e., a slowly moving articulated vehicle may actually sweep a larger area than the same articulated vehicle moving at a higher speed. The swept area is minimal for some "ideal" or "optimal" speed, and then increases again at sufficiently high speeds that slip occurs due to the increase in centripetal acceleration. Thus, according to some aspects, and for at least these reasons, the permissible vehicle state space may include upper and lower limits on vehicle speed. The upper and lower limits on vehicle speed may also depend on, among other things, the vehicle heading and road friction conditions.

[0053] In summary, for a slow-moving articulated vehicle, the trailer will cut inward during a constant radius turn. At some "break-even" speed, the trailer will begin to follow the same path as the towing vehicle, say for a given curve radius R, approximately when the lateral acceleration reaches approximately a y =0.25Gm / s 2 If the speed increases above the break-even speed, the swept path of the trailer will change to be outside the truck's partial path due to lateral acceleration.

[0054] These different states (inward sweeping area, outward sweeping area, trailer tracking towing vehicle) can all occur during different parts of a maneuver (such as a safety stop maneuver). Therefore, it is important to verify that, for example, a swept path with inward trailer tracking at low speed toward the end of the maneuver can be accommodated within the drivable area.

[0055] In addition, depending on the maneuver, the initial state of the vehicle when starting the maneuver may not allow the planned trajectory to be followed, in which case the potential swept area will extend beyond the planned trajectory. The geometry of the articulated vehicle also determines the potential swept area, at least because the vehicle geometry affects the sweep characteristics of the tractor truck and trailer combination, as explained in detail in US9,862,413B2. Finally, the sensor error characteristics affect the spatial extension of the potential swept area. This is because the sensor error determines the uncertainty of the vehicle state estimate during the maneuver. In the case of perfect sensor information that directly gives the true state of the vehicle without error, the potential swept area will completely follow the planned trajectory, at least in the case where the vehicle can be physically controlled to follow the trajectory. If the sensor input signal is associated with certain measurement or detection errors, the error characteristics or statistical distribution of the sensor error will affect the potential swept area. Compared with more accurate sensor signals, large sensor signal errors may produce larger potential swept areas.

[0056] Figure 3 An example scenario 30 is schematically shown in which an articulated vehicle 1 performs a maneuver from (A) to (C) via (B). The articulated vehicle 1 is traveling on a road 34 associated with a drivable area 31. The drivable area may include, for example, shoulders of the road depending on the type of maneuver, the vehicle, and other circumstances. There may also be a margin 32 to account for inaccuracies in determining the drivable area, for example, to account for errors in map data or the like. The vehicle 1 monitors the drivable area 31 in front of the vehicle and also estimates a potential swept area for a given maneuver (e.g., a safe parking maneuver). The potential swept area 33 at position (B) is estimated at Figure 3 The estimated potential swept area is included in the drivable area, so the vehicle state at position (B) allows the maneuver to be performed in a safe manner. For example, the vehicle speed v is below the maximum permissible speed v max acceptable level.

[0057] It is worth noting that the potential swept area includes the effects caused by the above-mentioned sweep performed by the towed vehicle when being towed by the towed vehicle.

[0058] Figure 4 Another example scenario 40 is schematically shown, in which the same articulated vehicle 1 is driven on the same road 34, but now at a higher speed v. At the same position (B), the potential swept area 33 is estimated and found to extend outside the drivable area 31. This means that if, for example, a sensor interruption is experienced at position (B), a safe parking maneuver cannot be completed in a safe manner, as there is a risk that the articulated vehicle may drive off the road.

[0059] Thus, by estimating the potential swept area and comparing it to the corresponding monitored drivable area, the vehicle state can be constrained within the permissible vehicle state space that allows a given maneuver. For example, the vehicle speed can be constrained to a speed below the maximum permissible speed to allow immediate execution of a safe parking maneuver in the event of, for example, a sensor outage.

[0060] Furthermore, optionally, the vehicle state space during the maneuver may be constrained to lie within the permissible state space. This may, for example, include limiting the steering wheel angle during the maneuver to reduce the risk of the vehicle tipping or skidding.

[0061] all in all, Figure 3 and Figure 4 An example of a method of determining the permissible vehicle state space of an articulated vehicle 1 to safely complete a maneuver is shown. As mentioned above, the maneuver may be a safety parking maneuver, or an emergency braking maneuver, or an obstacle avoidance maneuver, etc. The method is particularly useful for controlling a vehicle processing a situation where the set of available sensor input signals is reduced compared to a normal operating mode, i.e. a situation associated with a sensor signal interruption or an interruption in sensor signal processing.

[0062] The vehicle state space includes constraints on vehicle speed. The vehicle speed may be an absolute value of the vehicle speed without direction, or may be a velocity vector in two or three dimensions. It will be appreciated that some conditions may constrain speed differently depending on the vehicle heading; some headings may be more strongly constrained than others.

[0063] like Figure 3 and Figure 4 As shown, the method comprises monitoring a drivable area 33 in front of the articulated vehicle 1 .

[0064] According to some aspects, the monitoring includes obtaining information related to the traffic infrastructure in front of the vehicle 1. The information can be obtained from map data or from a vehicle data store. The information can also optionally be obtained from any of the following: one or more radar sensors, one or more lidar sensors, one or more vision sensors, and one or more V2X communication transceivers.

[0065] As described above, the method also includes predicting a potential swept area 33 in front of the articulated vehicle 1. The potential swept area 33 represents an area that is likely to be swept by the articulated vehicle 1 during maneuvering. The potential swept area can be represented, for example, by a polygon or other geometric structure that can be used to define an area. It will be appreciated that the potential swept area is an area where the articulated vehicle can be located with a certain probability of a confidence interval. Therefore, according to some aspects, the method also includes configuring a confidence interval or probability associated with the potential swept area.

[0066] The prediction of the potential swept area is based on the initial vehicle state, the geometry of the articulated vehicle 1, and the error characteristics associated with one or more sensor input signals used to locate the vehicle 1. The initial vehicle state includes the vehicle speed and other state variables describing the vehicle state. Generally, the higher the vehicle speed, the larger the potential swept area. For example, the higher the speed, the longer it usually takes to stop the vehicle, which means that the vehicle maneuver takes longer to complete. As described above in conjunction with Figure 1 and Figure 2 As discussed, the geometry of the vehicle affects the swept area. Error characteristics associated with the one or more sensor input signals also affect the potential swept area. Typically, uncertain sensor input signal data, i.e., sensor input signal data associated with large errors, gives more uncertainty and therefore also provides a larger potential swept area. On the other hand, more accurate sensor input signals typically result in a smaller potential swept area.

[0067] According to some aspects, the prediction includes predicting a potential swept area 33 in front of the articulated vehicle 1 based on the planned trajectory 23 of the vehicle 1 for the maneuver. For example, an unbiased control algorithm can generally be employed that follows the planned trajectory for the maneuver in question with some uniformly distributed errors around the planned trajectory 23, such as a Gaussian or normal error distribution with some covariance matrix. The potential swept area is then extended on both sides of the planned trajectory 23 with an extension amplitude determined by the Gaussian error covariance matrix and confidence interval parameters that determine the probability that the articulated vehicle is located within the potential swept area.

[0068] The method also includes determining a permissible vehicle state space such that the predicted potential swept area 33 does not exceed the drivable area 31 in front of the vehicle 1. Thus, as long as the vehicle state remains within the permissible vehicle state space, the maneuver can be performed without a significant risk of the vehicle leaving the drivable area.

[0069] refer to Figure 5a ,In general, a planned trajectory such as trajectory 23 from a starting position (A) to a target position (C) can be represented by a matrix of state vectors T indexed by time t;

[0070]

[0071] where x T (t) represents the position vector, v T (t) represents the velocity vector, and a T (t) represents the acceleration vector as a function of time t. The planned trajectory may also include other quantities such as steering wheel angle, turn rate, heading, etc. The error vector at time t can be defined as the vehicle state vector at time t

[0072]

[0073] and the corresponding planning trajectory matrix at time t

[0074]

[0075] The control algorithm for executing a maneuver in which the articulated vehicle follows the planned trajectory T(t) may be based on, for example, the squared error e(t) T Minimization of e(t). This control algorithm can be based on a variety of known tracking filter methods, such as Kalman filtering, extended Kalman filtering, Wiener filtering, and variants of particle filtering. According to one interpretation, the potential swept region models the errors that one of these algorithms might exhibit if the algorithm was used to execute the planned maneuver (i.e., follow the planned trajectory 23).

[0076] It will be appreciated that different components in the state vector matrix T may be given different priorities or weights when determining actions and control decisions. For example, when planning a maneuver, lateral position may be more important than, for example, longitudinal stopping distance.

[0077] Note that since the planned trajectory T optionally includes more variables than a simple position, the maneuver must also follow the velocity profile and acceleration profile of the planned trajectory. Therefore, an articulated vehicle that is exactly on the trajectory coordinate at time t1 may still be associated with a large error if the acceleration of the velocity of the articulated vehicle is inconsistent with the planned trajectory T.

[0078] The planned trajectory T can of course also include only the spatial trajectory, T(t) = x T (t), or spatial trajectory and velocity profile,

[0079]

[0080] The initial vehicle state S(0) at time t=0 and its estimated value associated with the error The size and shape of the potential swept area 33 may also be affected. For example, an uncertain starting point is more likely to produce a larger potential swept area than a more certain starting point of the articulated vehicle 1. The initial vehicle state may include any number of state variables, such as a two-dimensional or three-dimensional position, a velocity vector, an acceleration vector, a vehicle heading, and a vehicle turn rate. Each of the state variables may be associated with an error or uncertainty having an error characteristic described by some statistical distribution.

[0081] An example with normally distributed errors (sometimes called Gaussian errors) is Figure 5b, which shows an example scenario 50, in which it is assumed that the vehicle state error, at least in terms of position, is unbiased with respect to the planned trajectory 23 and follows a Gaussian distribution centered on the trajectory. The articulated vehicle 1 is located at a position φ that is different from the true position p. S1 , true velocity vector v S1 and the true heading vector h S1 The estimated state vector S = [p A , v A ,h A ] has an associated Gaussian or normally distributed error [n p , n v , n h ]. The uncertainty in the estimated state vector is described by the covariance matrix associated with the error variables. For the position estimate, this covariance matrix can be illustrated by an ellipse 51a, which indicates where the true value lies with some probability. For example, for a Gaussian distribution with mean u, the probability of the realization being within one standard deviation of the mean u is about 68.27%, the probability of the realization being within two standard deviations of the mean u is about 95.45%, and the probability of the realization being within three standard deviations of the mean u is about 99.73%.

[0082] In the case of a sensor signal outage, no accurate sensor signal is available, which means that the uncertainty can be expected to increase over time. Therefore, the error n describing the position (B) p Error distribution 52b may be larger than error distribution 52a, and uncertainty ellipse 51b covers a larger area than uncertainty ellipse 51a.

[0083] The following will be combined Figure 7a and Figure 7b The determination of potential sweep areas is discussed in more detail.

[0084] Figure 7a An example 70a of determining potential swept regions based on filtered covariance matrix estimation such as a Kalman filter is shown. Figure 7a A sequence of uncertainty ellipsoids 72, 74 are illustrated, sometimes also referred to in the literature as confidence regions. In statistics, a confidence region is a multi-dimensional generalization of a confidence interval. It is a set of points in n-dimensional space, usually represented as an ellipsoid around a point that is an estimated solution to a problem, but other shapes may also occur. The confidence region is calculated in such a way that if the set of measurements is repeated many times and the confidence region is calculated in the same way on each set of measurements, then a certain percentage of the time (e.g., 95%), the confidence region will include the point representing the "true" value of the set of variables being estimated. The determination of uncertainty ellipsoids or confidence regions is known and will not be discussed in detail here.

[0085] The covariance matrix in the Kalman filter algorithm estimates P at sample k (or equivalently, at discrete time index k) k|k (given data input until time k) is determined as P k|k =(IK k H k ) k|k-1 (IK k H k ) T +K k R k K k T , where P k|k-1 =F k P k-1|k-1 F k T +Q k , and where P k|k-1 is the covariance matrix estimate at time k (given data up to time k-1), I is the identity matrix, K k is the Kalman gain at time step k, H k is the observation matrix at time step k, R k is the assumed covariance matrix of the measurement noise, i.e., describing the uncertainty of the input sensor signal, F k is the state transition matrix k at time step k, and Q k is the process noise at time step k.

[0086] Note that for the covariance matrix P k|k The estimate of does not depend on the actual realization of the sensor data, but only on the error distribution and the assumed statistics of the process. Therefore, the uncertainty ellipsoids 72, 74 can be determined by simply iterating the above equations.

[0087] Figure 7a The uncertainty ellipsoid sequence in is determined starting from the initial vehicle state 71, where the associated uncertainty ellipsoid 72 is determined according to P at position (A) 0|0 Until the end position (C) of the planned trajectory 23 associated with the vehicle state 73 and the associated uncertainty ellipsoid 74 is determined. The potential swept area can now be determined as a boundary area 33a which bounds the error ellipsoid, or simply as a union of the uncertainty ellipsoid areas.

[0088] Taking into account that an articulated vehicle sweeps can be done in a number of ways, for example by expanding the area given by the uncertainty ellipsoid by a fixed amount, or by considering the towing vehicle to travel along the boundaries of the area determined by the uncertainty ellipsoid. The articulated vehicle geometry can then be used to determine the swept area based on the trajectory of the towing vehicle, as described in US 9,862,413 B2.

[0089] Figure 6 Another method 60 of determining potential sweep areas is shown. Figure 6 In , a particle filter is used, where a large number of particles 61a, 61b, 61c represent possible realizations of the tractor vehicle attempting to follow the planned trajectory 23. The realizations of the sensor input data are based on an assumed error distribution and fed to a simulation control algorithm, which outputs an evolving state estimate of the articulated vehicle corresponding to the particles. This type of method is often referred to as a Monte Carlo simulation method.

[0090] Figure 7b It is shown how to determine potential swept areas 33b based on a population of particles. As described in US9,862,413B2, using information about the geometry of an articulated vehicle, the trajectory of each particle is used to determine the swept area corresponding to the particle. The potential swept area is then given by a boundary area 33b that includes all swept areas of all particles. Optionally, outliers can be discarded to limit the potential swept area to an area of ​​more reasonable size.

[0091] For simplicity, the potential swept area may also be determined as the smallest area that includes at least a given percentage of particles (eg, 95% or more of the particles).

[0092] As mentioned above, the techniques disclosed herein may also be used to safely control an articulated vehicle during a maneuver by constraining the vehicle state within some permissible state space that also includes steering angle variables, etc. According to some such aspects, the vehicle state space includes a steering wheel angle, and the method includes determining the permissible vehicle state space further based on a rollover risk associated with a planned trajectory of the vehicle for the maneuver.

[0093] According to other such aspects, the method further comprises obtaining one or more dynamic properties associated with the vehicle 1. Determining the permissible vehicle state space based on the rollover risk is then also based on the vehicle dynamic properties. The vehicle dynamic properties may include, for example, a moment of inertia, a center of gravity value, etc. associated with the articulated vehicle 1.

[0094] The results of the methods disclosed herein can be used during control of the vehicle to, for example, automatically limit the vehicle speed if the vehicle speed exceeds the speed of the permissible vehicle state space. Thus, according to some aspects, the methods disclosed herein include providing the determined permissible vehicle state space as an output to the control unit 900 of the vehicle 1 for control during maneuvers. Fig. 9 The control unit is discussed in more detail.

[0095] Potential sweep regions can also be used in conjunction with Fig. 9Vehicle controller discussed. Thus, the methods disclosed herein may optionally include providing an estimate of the potential swept area and / or rollover likelihood during a maneuver as an output to the control unit 900 of the vehicle 1 for control during the maneuver.

[0096] Optionally, the method includes determining a plurality of potential swept areas corresponding to different sets of sensor input signal combinations and / or different maneuvers. Thus, potential swept areas and therefore also allowable vehicle state spaces can be determined for a plurality of different maneuvers (i.e., safe parking, obstacle avoidance, etc.). Furthermore, potential swept areas and allowable vehicle state spaces can be determined for a plurality of different sensor outage scenarios (i.e., radar outage only, lidar and radar outage, all external sensors outage, etc.).

[0097] In situations where a high degree of uncertainty is associated with the vehicle state space (ie, the potential swept area becomes larger) the vehicle may prepare for a collision, and / or a warning signal may be triggered.

[0098] Figure 8 is a flow chart illustrating the method disclosed herein and summarizing the above discussion. A method for determining an allowable vehicle state space of an articulated vehicle 1 for safely completing a maneuver is shown, wherein the vehicle state space includes vehicle speed. The method includes monitoring S1 a drivable area 33 in front of the articulated vehicle 1 and predicting S2 a potential swept area 33 in front of the articulated vehicle 1. The potential swept area 33 represents an area that the articulated vehicle 1 may pass through during the maneuver. The potential swept area is determined based on an initial vehicle state, a geometry of the articulated vehicle 1, and error characteristics associated with one or more sensor input signals used to locate the vehicle 1. The method also includes determining S3 an allowable vehicle state space such that the predicted potential swept area 33 does not exceed the drivable area 31 in front of the vehicle 1.

[0099] According to some aspects, the maneuver is a safety stop maneuver and the one or more sensor signals correspond to a reduced set of sensor input signals compared to a normal operating mode.

[0100] According to some aspects, monitoring includes obtaining S11 information related to traffic infrastructure in front of the vehicle 1 .

[0101] According to some aspects, monitoring includes obtaining S12 data related to the environment in front of the vehicle 1 from any of: one or more radar sensors, one or more lidar sensors, one or more vision sensors, and one or more V2X communication transceivers.

[0102] According to some aspects, the predicting comprises predicting S21 a potential swept area 33 in front of the articulated vehicle 1 based on the vehicle's planned trajectory for the maneuver.

[0103] According to some aspects, predicting comprises obtaining S22 an estimate of an initial vehicle state error associated with the vehicle 1 .

[0104] According to some aspects, the predicting comprises estimating S23 a covariance matrix associated with the positioning errors of the vehicle 1 during the maneuver.

[0105] According to some aspects, the vehicle state space includes a steering wheel angle, and the method includes determining S31 a permissible vehicle state space further based on a rollover risk associated with a planned trajectory of the vehicle for the maneuver.

[0106] According to some such aspects, the method comprises obtaining S32 one or more dynamic properties associated with the vehicle 1 and determining an allowable vehicle state space based on the rollover risk is based on the vehicle dynamic properties.

[0107] According to some aspects, the method comprises providing S4 the determined permissible vehicle state space as an output to a control unit 900 of the vehicle 1 for control during the manoeuvre.

[0108] According to some aspects, the method comprises providing S5 an estimate of a potential swept area and / or rollover likelihood during a manoeuvre to a control unit 900 of the vehicle 1 as an output to the control unit 900 of the vehicle 1 for control during the manoeuvre.

[0109] According to some such aspects, the method includes determining S51 a plurality of potential sweep areas corresponding to different sets of sensor input signal combinations and / or different maneuvers.

[0110] According to some aspects, the method comprises triggering a warning signal S6 based on an estimated uncertainty associated with the estimated state of the articulated vehicle 1 .

[0111] Fig. 9 The components of the control unit 900 according to the embodiments discussed herein are schematically shown according to a number of functional units. The control unit 900 may be included in the articulated vehicle 1. The processing circuit 910 is provided using any combination of one or more of a suitable central processing unit CPU, a multiprocessor, a microcontroller, a digital signal processor DSP, etc., capable of executing software instructions stored in a computer program product (e.g., in the form of a storage medium 930). The processing circuit 910 may also be provided as at least one application specific integrated circuit ASIC or a field programmable gate array FPGA.

[0112] In particular, the processing circuit 910 is configured to cause the control unit 900 to perform an operation or set of steps, such as combining Figure 8For example, the storage medium 930 may store the set of operations, and the processing circuit 910 may be configured to retrieve the set of operations from the storage medium 930 to cause the control unit 900 to perform the set of operations. The set of operations may be provided as an executable instruction set. Thus, the processing circuit 910 is thereby arranged to perform the method as disclosed herein.

[0113] The storage medium 930 may also include a persistent storage, which may be, for example, any one or combination of magnetic storage, optical storage, solid-state storage, or even a remotely mounted storage.

[0114] The control unit 900 may also include an interface 920 for communicating with at least one external device, such as an antenna array including a phase controller and a mechanically rotatable substrate. Thus, the interface 920 may include one or more transmitters and receivers, including analog and digital components and an appropriate number of ports for wired or wireless communication.

[0115] Processing circuit 910 controls the general operation of control unit 900, for example, by sending data and control signals to interface 920 and storage medium 930, by receiving data and reports from interface 920, and by retrieving data and instructions from storage medium 930. Other components of the control node and related functions are omitted so as not to obscure the concepts presented herein.

[0116] The control unit 340 optionally includes a heading detection unit 940, such as a compass or a GPS module. The control unit may also include any of the following: an IMU 950, a wheel speed sensor 960, and an articulation angle sensor 970. A wheel angle sensor 980 and / or a vehicle tilt angle sensor 990 may also be optionally provided.

[0117] Fig.10 A computer readable medium 1010 is shown carrying a computer program comprising program code means 1020 for executing when the program product is run on a computer. Figure 8 The computer readable medium and the code means may together form a computer program product 1000 .

Claims

1. A method of determining an allowable vehicle state space for safely completing a maneuver of an articulated vehicle (1), wherein the allowable vehicle state space defines a portion of a vehicle state space such that the allowable vehicle state space only includes states that enable the maneuver to be completed safely, wherein the vehicle state space includes vehicle speed, the method comprising: monitoring a drivable area (31) in front of the articulated vehicle (1), predicting a potential swept area (33) ahead of the articulated vehicle (1) based on one or more sensor input signals for positioning the articulated vehicle (1) for a scenario in which at least one of the sensor input signals experiences an interruption, wherein the potential swept area (33) represents an area that the articulated vehicle (1) can pass through during the maneuver with a probability based on an initial vehicle state of the articulated vehicle (1), a geometry of the articulated vehicle (1) and error characteristics associated with the one or more sensor input signals, and The permissible vehicle state space is determined so that the predicted potential swept area (33) does not extend beyond a drivable area (31) in front of the articulated vehicle (1).

2. The method according to claim 1, wherein: The maneuver is a safety stop maneuver, and wherein the one or more sensor input signals correspond to a reduced set of sensor input signals compared to a normal operating mode.

3. The method according to any one of claims 1 to 2, wherein: The monitoring comprises obtaining information related to the traffic infrastructure in front of the articulated vehicle (1).

4. The method according to any one of claims 1 to 2, wherein: The monitoring comprises obtaining data relating to the environment in front of the articulated vehicle (1) from any of: one or more radar sensors, one or more lidar sensors, one or more vision sensors and one or more V2X communication transceivers.

5. The method according to any one of claims 1 to 2, wherein: The prediction includes predicting a potential swept area (33) ahead of the articulated vehicle (1) based on a planned trajectory (23) associated with the maneuver.

6. The method according to any one of claims 1 to 2, wherein: The prediction comprises obtaining an estimate of an initial vehicle state error associated with the articulated vehicle (1).

7. The method according to any one of claims 1 to 2, wherein: The prediction comprises estimating a covariance matrix associated with positioning errors of the articulated vehicle (1) during the maneuver.

8. The method according to any one of claims 1 to 2, wherein: The vehicle state space includes steering wheel angles, and wherein the method includes determining the permissible vehicle state space further based on a rollover risk associated with a planned trajectory (23) of the articulated vehicle (1) associated with the maneuver.

9. The method according to claim 8, comprising obtaining one or more dynamic properties associated with the articulated vehicle (1), wherein the permissible vehicle state space based on rollover risk is determined based on the dynamic properties.

10. A method according to any of claims 1-2, comprising providing the determined permissible vehicle state space as an output to a control unit (900) of the articulated vehicle (1) for control during the manoeuvre.

11. A method according to any one of claims 1-2, comprising providing an estimate of the potential swept area and / or rollover probability during the manoeuvre as an output to a control unit (900) of the articulated vehicle (1) for control during the manoeuvre.

12. The method of claim 11, comprising determining a plurality of potential sweep areas corresponding to different sets of sensor input signal combinations and / or corresponding to different maneuvers.

13. The method according to any of claims 1-2, comprising triggering a warning signal based on an estimated uncertainty associated with the estimated vehicle state of the articulated vehicle (1).

14. A computer-readable medium (1010) carrying a computer program, the computer program comprising program code means for executing the steps of a method for determining an allowable vehicle state space of an articulated vehicle (1) for safely completing a maneuver as described in any one of claims 1 to 13 when the computer program is run on a computer or on a processing circuit of a control unit.

15. A control unit (900) for determining an allowable vehicle state space of an articulated vehicle (1), the control unit (900) being configured to execute the steps of the method for determining an allowable vehicle state space of an articulated vehicle (1) for safely completing a maneuver according to any one of claims 1-13.

16. A vehicle (1) comprising a control unit (900) according to claim 15.