TECHNICAL FIELD

The invention relates to a control unit and a method for handling operational requirements when driving between two areas. The invention further relates to an autonomous vehicle, a computer program, and a computer program product.

The invention can be applied in heavy-duty vehicles, such as trucks, buses and construction equipment. Although the invention will be described with respect to an autonomous vehicle, the invention is also applicable to semi-autonomous vehicles.

BACKGROUND

Autonomous vehicles navigate work areas using localization services for finding their respective positions. Guaranteeing a high integrity for localization when navigating autonomous vehicles is one of the most important problems to solve for automating complex work areas. Typically, to navigate these work areas, autonomous vehicles need to depend on perception systems and/or sensors. To ensure safety, it is often necessary to provide a formal and/or extensive proof of that the autonomous vehicle does not drive in critical areas which may cause danger to manual actors. These types of proofs often involve extensive high-quality testing which is both time consuming and costly to a degree that makes it infeasible. During the tests, a localization solution needs to be verified against ground truth data with very high robustness and accuracy. However, the amount of data may be too vast, and/or the type of data may simply not be available.

A localization system typically estimates an autonomous vehicle’s position inside a recorded map, e.g. by map matching techniques. The localization system tries to recognize features around itself and matches those features with the map. Most perception tasks suffer from problems with providing accurate confidence values for position estimation. A particular problem when using map matching techniques is that faults that occur tend to be biased towards previously estimated values regardless of whether it was a correct or wrong estimation. Hence, if the bias is towards a correct value, it is likely an acceptable value. However, if the bias is towards a previously estimated wrong value, it will be difficult to detect increasing inaccuracies. It is however difficult to detect faults in a localization system since there may be no other data than what is supplied by the localization system to compare with. For some scenarios, it may be possible to perform simpler consistency checks with a perception system, e.g. by checking that the autonomous vehicle does not move a lot faster than what should be possible, e.g. 100 meters (m) in 1 second (sec). These types of checks can however only detect certain types of extreme and/or obvious faults.

Ensuring safety for autonomous vehicles puts a requirement of high integrity and extremely low frequency of errors. As is exemplified above, it is difficult to estimate the vehicle’s current position and pose in an accurate enough manner. Therefore, when driving in work areas, a vehicle motion may need to be constrained by various operational requirements. These requirements may limit the speed of the autonomous vehicles. Due to these constraints, the autonomous vehicle loses productivity as it cannot operate optimally, e.g. at its highest speed. Thus, a problem arises in how to improve productivity while ensuring safety of manual actors and other vehicles.

SUMMARY

An object of the invention is to improve the productivity of autonomous vehicles.

According to a first aspect, the above object is achieved by a method according to claim 1. Hence, there is provided a method for handling operational requirements for an autonomous vehicle driving from a first work area to a second work area via a predefined path. The operations of the autonomous vehicle are constrained by the operational requirements. The method comprises: detecting that the autonomous vehicle is initiating driving in the predefined path, when detected that the autonomous vehicle is initiating driving in the predefined path, triggering a relaxation of the operational requirements, when the autonomous vehicle is driving in the predefined path, detecting whether the autonomous vehicle turns towards the first work area, when detecting that the autonomous vehicle turns towards the first work area, triggering a preventive action for preventing the autonomous vehicle from driving towards the first work area using the relaxed operational requirements, measuring a distance and/or time driven in the predefined path, and when the measured distance and/or time driven in the predefined path fulfil at least one predetermined condition, triggering an increase of the operational requirements.

When it is detected that the autonomous vehicle is initiating driving in the predefined path, the relaxation of the operational requirements is triggered. In this way, improved productivity is achieved as the autonomous vehicle is less constrained in the predefined path. The first and second work areas may have manual actors, and thus to maintain safety of the manual actors, the autonomous vehicle needs to be restricted from driving in these areas with the relaxed operational requirements. To ensure that the autonomous vehicle does not drive in the first work area with the relaxed operational requirements, the preventive action is triggered when it is detected that the autonomous vehicle turns around towards the first work area. To ensure that the autonomous vehicle does not drive in the second work area with the relaxed operational requirements, the operational requirements are increased, e.g. back to a default operational requirement configuration, when the measured distance and/or time driven in the predefined path fulfil the at least one predetermined condition, e.g. when it is deduced based on the measured time and/or distance that the autonomous vehicle is close to the second work area.

Optionally, detecting that the autonomous vehicle is initiating driving in the predefined path comprises determining that the autonomous vehicle is positioned in the first work area and that a driving direction of the vehicle is directed along the predefined path.

Optionally, detecting that the autonomous vehicle is initiating driving in the predefined path comprises detecting that the autonomous vehicle is loaded, e.g. fully loaded or loaded above a predefined weight and/or volume. By “loaded” is herein intended that the vehicle is carrying weight, such as goods.

Optionally, detecting that the autonomous vehicle is initiating driving in the predefined path comprises receiving a signal indicating that the autonomous vehicle is initiating driving in the predefined path. The signal may be triggered based on detecting that the autonomous vehicle is loaded, e.g. fully loaded, and/or when there is a forward motion of the autonomous vehicle.

Optionally, the signal is received from a wireless device or stationary control device in the first work area, which wireless device or stationary control device is operated by a user in the first work area. In other words, the user may first inspect the autonomous vehicle and decide that it is time to initiate driving in the predefined path towards the second work area, and trigger the signal and/or trigger the autonomous vehicle to start driving in the predefined path.

Optionally, the operational requirements comprise one or more safety constraints imposed on operations of the autonomous vehicle for hindering a collision between the autonomous vehicle and manual actors. Increasing the operational requirements may hence comprise... and relaxing the operational requirements may comprise...

Optionally, triggering the relaxation of the operational requirements comprises any one or more out of:

- increasing a maximum speed limit of the autonomous vehicle,

- adjusting one or more sensors and/or subsystems of the autonomous vehicle.

Optionally, adjusting the one or more sensors and/or subsystems comprises any one or more out of:

- adjusting or switching a brake system of the autonomous vehicle, and

- disabling an obstacle detection system of the autonomous vehicle.

Optionally, detecting whether the autonomous vehicle turns towards the first work area comprises:

- estimating a turning radius of the autonomous vehicle, and

- comparing the estimated turning radius with a predefined width of the predefined path.

The estimated turning radius may also be predicted and/or measured. When the turning radius indicates that the autonomous vehicle will be able to turn around within the width of the predefined path, the preventive action may be triggered.

Optionally, the autonomous vehicle is detected to turn towards the first work area when 2R- K < W, wherein R is the estimated turning radius of the autonomous vehicle, K is a preconfigured margin constant, and wherein W is the predefined width of the predefined path. In other words, when the turning radius times two, i.e. a turning diameter, is smaller than the predefined width of the predefined path, the preventive action must be triggered as the autonomous vehicle may otherwise turn towards and drive towards the first work area with the relaxed operational requirements.

Optionally, detecting whether the autonomous vehicle turns towards the first work area comprises: estimating an average turning radius of the autonomous vehicle for a predetermined past time period, and comparing the estimated turning radius with a predetermined maximum allowed average turning radius for the predetermined past time period.

Optionally, detecting whether the autonomous vehicle turns towards the first work area comprises: measuring an aggregated turning angle of the autonomous vehicle, wherein the aggregated turning angle indicates a total angle turned by the autonomous vehicle when driving in the predefined path, and comparing the aggregated turning angle with a predetermined maximum allowed steering wheel angle of the autonomous vehicle.

The aggregated turning angle may comprise a total angle turned in either or both directions. As the predefined path may have predefined curvatures, the predetermined maximum allowed steering wheel angle may be calculated in advance, and if exceeding this maximum allowed steering wheel angle, it is detected that the autonomous vehicle may turn around towards the first work area, and thereby it is needed to trigger the preventive action.

Optionally, estimating the turning radius and/or measuring the aggregated turning angle comprises determining a turning path for the autonomous vehicle.

Optionally, estimating the turning radius and/or measuring the aggregated turning angle comprises determining one or more turning parameters based on a current vehicle speed for a predetermined future time period.

Optionally, the turning parameters comprise any one or more out of: a yaw rate, a steering wheel angle, and a lateral acceleration of the autonomous vehicle. Any of these parameters, independently or in any suitable combination, may be used to measure how the autonomous vehicle is turning, and/or predict how the autonomous vehicle will turn. Optionally, triggering the preventive action comprises any one or more out of: triggering an increase of the operational requirements, triggering an emergency stop, triggering a handover to a remote control drive mode, triggering an alert, and triggering an adjustment and/or limitation of a turning angle of the autonomous vehicle.

Optionally, the at least one predetermined condition is fulfilled when the measured distance is above a predefined distance threshold.

Optionally, the at least one predetermined condition is fulfilled when the measured time multiplied with a predetermined max speed of the predefined path is above the predefined distance threshold.

Optionally, triggering the increase of the operational requirements comprises resetting the operational requirements, e.g. back to a default operational requirement configuration.

Optionally, triggering the increase of the operational requirements comprises any one or more out:

- decreasing a maximum speed limit of the autonomous vehicle,

- adjusting one or more sensors and/ or subsystems of the autonomous vehicle.

According to a second aspect, there is provided a control unit to perform the method according to the first aspect. The control unit may be an electronic control unit.

According to a third aspect, there is provided an autonomous vehicle comprising the control unit according to the second aspect. The autonomous vehicle is configured to drive from a first work area to a second work area via a predefined path.

Optionally, the autonomous vehicle is further configured with a predetermined maximum allowed steering wheel angle. In this way, the autonomous vehicle may be restricted to only be able to turn a predefined angle, and in this way, the autonomous vehicle is at least partly restricted in its ability to turn around within the predefined path. According to a fourth aspect, there is provided a computer program comprising program code means for performing the method according to the first aspect, when said program is run on a computer.

According to a fifth aspect, there is provided a computer program medium carrying a computer program comprising program code means for performing the method according to the first aspect, when said program is run on a computer.

Further advantages and advantageous features of the invention are disclosed in the following description and in the dependent claims.

BRIEF DESCRIPTION OF THE DRAWINGS

With reference to the appended drawings, below follows a more detailed description of embodiments of the invention cited as examples.

In the drawings:

Fig. 1 is a schematic overview of an autonomous vehicle, two work areas and a predefined path, and

Fig. 2 is a flowchart illustrating a method according to embodiments herein.

Fig. 3a-3c are illustrations of example scenarios of an autonomous according to embodiments herein.

Fig. 4a-4b are schematic block diagrams illustrating a control unit according to embodiments herein.

DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS OF THE INVENTION

Fig 1. is a schematic overview of an autonomous vehicle 1 in accordance with embodiments herein. The autonomous vehicle 1 may be any autonomous or semi- autonomous vehicle, e.g. a car, truck, bus, heavy-duty vehicle, wheel loader, etc. The autonomous vehicle 1 is arranged to travel from a first work area 10 to a second work area 20 via a predefined path 30. Typically the autonomous vehicle 1 may be configured to transport objects between the first work area 10 and the second work area 20. The first work area 10 and the second work area 20 may e.g. be part of the same warehouse or in different warehouses. In any case the first work area 10 and the second work area 20 are connected by the predefined path 30. The first work area 10 and the second work area 20 may respectively have manual actors working therein, e.g. human workers which may load/unload the autonomous vehicle 1 before departure of the autonomous vehicle 1 in the first work area 10 and/or the second work area 20. The predefined path 30 may not have any manual actors present and may e.g. be isolated from manual actors by the use of any one or more out of: fences, barriers, warning signs, etc.

The autonomous vehicle 1 may be configured to be constrained by operational requirements of the autonomous vehicle 1 according to a default configuration, e.g. for being able to safely navigate areas where manual actors may appear. The operational requirements may be any requirements which restricts an operation of the autonomous vehicle 1. For example, the operational requirements may restrict a maximum speed of the autonomous vehicle 1 and/or restrict how sensors and/or subsystems of the autonomous vehicle 1 are configured. In Figure 1 , it is illustrated that the autonomous vehicle 1 is positioned in the first work area 10 and has a direction D along the predefined path, i.e. towards the second work area. When it is detected that the autonomous vehicle 1 is initiating driving in the predefined path, the operational requirements are relaxed. Relaxing the operational requirements means that the autonomous vehicle 1 will be less limited in their ability to operate. For example, the operational requirements may be relaxed such as e.g. the maximum speed limit allowed for the autonomous vehicle 1 is increased. The autonomous vehicle 1 may then operate more efficiently in the predefined path 30. The operational requirements may be relaxed as it is known that no manual actors will appear in the predefined path 30 and thus there is less demand for safety than in the first work area 10 and/or the second work area 20. The autonomous vehicle 1 may then operate autonomously in the predefined path 30, e.g. along a path 3, using the relaxed operational requirements until it is detected that the autonomous vehicle 1 is close to the second work area 20. When it is detected that the autonomous vehicle 1 is turning around towards the first work area 10, e.g. in a turnaround path 2, the autonomous vehicle 1 is at risk of operating in the first work area 10 using the relaxed operational requirements, and thus a preventive action needs to be taken, e.g. hindering/stopping the autonomous vehicle 1 and/or increasing the operational requirements back to the default configuration. In other words, in embodiments herein, the first work area 10 and the second work area 20 may respectively be critical work areas that need strict operational requirements since humans may be present. In contrast, the predefined path 30 may be non-critical and relaxed operational requirements may suffice as manual actors are not to be present therein.

Embodiments herein may be performed by a control unit 70. The control unit 70 may be comprised in the autonomous vehicle 1 but may also be comprised in any other suitable location communicatively coupled with the autonomous vehicle 1 , e.g. in the first work area 10, the second work area 20 or in a remote cloud environment.

Fig. 2 illustrates a method for handling operational requirements for the autonomous vehicle 1 driving from the first work area 10 to the second work area 20 via the predefined path 30. The operations of the autonomous vehicle 1 are constrained by the operational requirements. In some embodiments, the operational requirements comprise one or more safety constraints imposed on operations of the autonomous vehicle 1 for hindering or mitigating a collision between the autonomous vehicle 1 and manual actors. The method comprises the following actions described below, which actions may be taken in any suitable order. Optional actions are indicated by dashed boxes in Fig. 2.

Action 201

The method comprises detecting that the autonomous vehicle 1 is initiating driving in the predefined path 30.

Detecting that the autonomous vehicle 1 is initiating driving in the predefined path may comprise one or more detection conditions as exemplified below, however, any suitable manner of detecting that the autonomous vehicle 1 is initiating driving in the predefined path 30 is sufficient.

In some embodiments, detecting that the autonomous vehicle 1 is initiating driving in the predefined path 30 comprises determining that the autonomous vehicle 1 is positioned in the first work area 10 and that the driving direction D of the vehicle is 1 directed along the predefined path 30. In other words, the pose, position and direction of the autonomous vehicle 1 indicates that the autonomous vehicle 1 is initiating driving in the predefined path 30. In some embodiments, detecting that the autonomous vehicle 1 is initiating driving in the predefined path 30 comprises detecting that the autonomous vehicle 1 is loaded, e.g. fully loaded or loaded with respect to some predefined threshold. It can be determined that the autonomous vehicle 1 is loaded by weighing the autonomous vehicle 1 or by any other determining means.

In some embodiments, detecting that the autonomous vehicle 1 is initiating driving in the predefined path 30 comprises receiving a signal indicating that the autonomous vehicle 1 is initiating driving in the predefined path 30. The signal may be received from a wireless device or stationary control device in the first work area 10, which wireless device or stationary control device is operated by a user in the first work area 10. In other words the user may signal when the autonomous vehicle 1 is to start driving in the predefined path 30 and e.g. is fully loaded and/or is directed in the direction D along the predefined path 30.

In some embodiments, an act of loading the autonomous vehicle 1 may be a way of confirming a correct direction/pose of the autonomous vehicle 1 , such that the autonomous vehicle 1 may initiate driving in the predefined path 30 when the autonomous vehicle 1 initiates a forward motion. Furthermore, a loading operator may be instructed to only load the autonomous vehicle 1 if it has the correct pose, heading, and/or location.

Action 202

The method comprises, when detecting that the autonomous vehicle 1 is initiating driving in the predefined path 30, triggering a relaxation of the operational requirements. In other words, the autonomous vehicle 1 will be less constrained and may have more options available for operational/driving decisions when driving in the predefined path 30.

In some embodiments, triggering the relaxation of the operational requirements comprises any one or more out of:

- increasing a maximum speed limit of the autonomous vehicle 1 ,

- adjusting one or more sensors and/or subsystems of the autonomous vehicle 1.

Adjusting the one or more sensors and/or subsystems may comprise any one or more out of:

- adjusting or switching brake system of the autonomous vehicle 1 , and - disabling and/or switching an obstacle detection system of the autonomous vehicle

1.

In other words, relaxing the operational requirements may cause the autonomous vehicle 1 to be able to drive faster and/or to use sensors and subsystems which are of higher performance and/or improved resource management. When the autonomous vehicle 1 operates using the relaxed operational requirements, the autonomous vehicle 1 may rely on alternative safety guarantees than when operation using higher operational requirements. For example, some sensors in the autonomous vehicle 1 may be turned off which may lower performance or reliability of various functions, e.g. pedestrian detection and/or braking capabilities.

The autonomous vehicle 1 may before the relaxation of the operational requirements use a default operational requirement configuration. The relaxed operational requirements may be relaxed relative the default operational requirement configuration.

Action 203

The method comprises, when the autonomous vehicle 1 is driving in the predefined path 30, detecting whether the autonomous vehicle 1 turns towards the first work area 10.

Detecting that the autonomous vehicle 1 drives towards the first work area 10 may comprise detecting that the autonomous vehicle 1 is driving in a way which may cause the autonomous vehicle 1 to turn 180 degrees, but may also comprise detecting turning motions which are less than 180 degrees, e.g. 170 degrees, as long as there is a risk of the autonomous vehicle 1 turning around towards the first work area 1. In some embodiments, detecting that the autonomous vehicle 1 drives towards the first work area 10 may also comprise detecting that the autonomous vehicle 1 is driving in reverse.

Detecting whether the autonomous vehicle 1 turns towards the first work area 10 may comprise one or more detection conditions as exemplified below, however, any suitable manner of detecting whether the autonomous vehicle 1 turns towards the first work area 10, or not, is sufficient.

In some embodiments, detecting whether the autonomous vehicle 1 turns towards the first work area 10 comprises: estimating a turning radius of the autonomous vehicle 1 , and comparing the estimated turning radius with a predefined width of the predefined path 30.

The comparison may deduce whether or not the turning radius means that the vehicle will have a turning path which extends beyond or within the predefined width. Since the radius is compared to the width of the predefined path 30, all types of turns towards the first work area 10 may be detected, and it is thus possible to ensure whether or not the autonomous vehicle 1 is turning around towards the first work area 10.

In some embodiments, the predefined width may account for some margin of error and be a short interval instead of a fixed absolute number. The predefined width may be a conservative estimate to ensure that it is always detected when the autonomous vehicle 1 may turn around.

In some embodiments, the autonomous vehicle 1 is detected to turn towards the first work area 10 when 2R-K < W. In these embodiments R is the estimated turning radius of the autonomous vehicle 1 , K is a preconfigured margin constant, and W is a predefined width of the predefined path 30. Thus, if two times the radius R is less than the width W of the predefined path 30, accounting for some margin K, then the autonomous vehicle 1 is detected to be turning towards the first work area 10. In some of these embodiments, the preconfigured margin constant K may include compensation for a width of the autonomous vehicle 1 , e.g. which may in some scenarios cause the constant to be negative.

In some embodiments, detecting whether the autonomous vehicle 1 turns towards the first work area 10 comprises:

• estimating an average turning radius of the autonomous vehicle 1 for a predetermined past time period, and

• comparing the estimated turning radius with a predetermined maximum allowed average turning radius for the predetermined past time period.

The average turning radius may be computed for every 1 time unit, e.g. seconds, for the past few seconds, e.g. 1 to 5 seconds. Thus, if above the predetermined maximum allowed average turning radius, e.g. 5 to 30 meters, then it can no longer be ensured that the autonomous vehicle 1 is not turning towards the first work area 10. Thus the autonomous vehicle 1 is detected to be turning towards the first work area 10.

In some embodiments, detecting whether the autonomous vehicle 1 turns towards the first work area 10 comprises:

• measuring an aggregated turning angle of the autonomous vehicle 1 , wherein the aggregated turning angle indicates a total angle turned by the autonomous vehicle 1 when driving in the predefined path 30, and

• comparing the aggregated turning angle with a predetermined maximum allowed steering wheel angle of the autonomous vehicle 1.

The aggregated turning angle may indicate how much the autonomous vehicle 1 has turned in total. As the predefined path 30 may have predefined curvatures, the predetermined maximum allowed steering wheel angle can be calculated in advance, and if exceeding this maximum allowed steering wheel angle, it is detected that the autonomous vehicle 1 turns or at least risks turning around towards the first work area 10, and thereby it is needed to trigger the preventive action.

In some embodiments, estimating the turning radius and/or measuring the aggregated turning angle comprises determining one or more turning parameters based on a current vehicle speed for a predetermined future time period. The turning parameters may comprise any one or more out of: a yaw rate, a steering wheel angle, and a lateral acceleration of the autonomous vehicle 1. Any of these parameters, independently or in any suitable combination, may be used to measure how the autonomous vehicle is turning, and/or predict how the autonomous vehicle will turn.

In some embodiments, estimating the turning radius and/or measuring the aggregated turning angle comprises determining a turning path P for the autonomous vehicle 1. It may further be detected if the turning path P extends in a lateral distance less wide than the width of the predefined path, and if so, the autonomous vehicle 1 may be considered to be turning towards the first work area 10.

Action 204

When detecting that the autonomous vehicle 1 turns towards the first work area 10, e.g. as in Action 203 above, the method comprises triggering a preventive action for preventing the autonomous vehicle 1 from driving towards the first work area 10 using the relaxed operational requirements. Any suitable preventive action for hindering the autonomous vehicle 1 to drive towards the work area 10 using the relaxed operational requirements may be used and many different alternatives exists.

For example, in some embodiments, triggering the preventive action comprises any one or more out of:

• triggering an increase of the operational requirements, e.g. back to the default configuration

• triggering an emergency stop,

• triggering a handover to a remote control drive mode,

• triggering an alert, and

• triggering an adjustment and/or limitation of a turning angle of the autonomous vehicle 1.

In any case, the preventive action ensures that the autonomous vehicle 1 will not endanger any manual actors in the first work area 10 by driving therein using the relaxed operational requirements.

Action 205

The method comprises measuring a distance and/or time driven in the predefined path 30. This may be performed by any suitable method, e.g. by use of internal sensors in the autonomous vehicle 1.

Action 206

The method comprises, when the measured distance and/or time driven in the predefined path 30 fulfil at least one predetermined condition, triggering an increase of the operational requirements. Triggering an increase of the operational requirements as used in embodiments herein means to cause the operational requirements to be stricter, i.e. the operational requirements will be less relaxed. For example, decreasing the maximum speed limit allowed by the autonomous vehicle 1 .

The at least one predetermined condition may indicate that the autonomous vehicle 1 is likely to be very close to the second work area 20, at least when driving along the predefined path 30, e.g. using a maximum speed limit allowed for the predefined path 30. In some embodiments, wherein the at least one predetermined condition is fulfilled when the measured distance is above a predefined distance threshold. Thus if driving the shortest possible route, the autonomous vehicle 1 may be a known distance from the second work area 20.

In some embodiments, the at least one predetermined condition is fulfilled when the measured time multiplied with a predetermined max speed of the predefined path 30 is above the predefined distance threshold. The reasoning is the same as for with the measured distance mentioned above, but instead time and the max speed is considered to deduce a distance travelled by the autonomous vehicle 1 if driven as fast as possible at the shortest possible route of the predefined path 30 between the first work area 10 and the second work area 20.

In some embodiments, triggering the increase of the operational requirements comprises resetting the operational requirements, e.g. back to the default configuration.

In some embodiments, triggering the increase of the operational requirements comprises any one or more out of:

- decreasing a maximum speed limit of the autonomous vehicle 1 , and

- adjusting one or more sensors and/ or subsystems of the autonomous vehicle 1. This may be an inverse operation from what is performed in Action 202 above.

The methods will now be further explained and exemplified in below embodiments. These below embodiments may be combined with any suitable embodiment as described above.

Fig. 3a illustrates an example scenario wherein the autonomous vehicle 1 is driving using the relaxed operational requirements in the predefined path 30 in a longitudinal direction D towards the second work area 20. The predefined path 30 has in the example scenario a predefined width W in a transverse direction and a predefined length L. The predefined path 30 comprises an area bounded by a threshold distance T before the second work area 20. In the example scenario of Fig. 3a, it is detected that the autonomous vehicle 1 is starting to turn and a turning radius R is estimated for the autonomous vehicle 1. The turning radius R may be used to deduce an estimated position of the autonomous vehicle 1 if upon a completion of the turn, the estimated position being a transverse distance corresponding to 2R away from the starting position. Using the radius R, it may be possible to derive a turning path P for how the autonomous vehicle 1 would drive to said estimated position. Since the width W is known for the predefined path 30, it is possible to deduce that said estimated position would not be possible as the autonomous vehicle Iwould after the completed turn be outside of the predefined path 30, i.e. as 2R is at least as wide as the predefined width W. Thus, in the example scenario, the autonomous vehicle 1 is detected to not be turning around and may keep operating using the relaxed operational requirements, e.g. as in Action 203 above. When an estimated distance and/or time indicates that the autonomous vehicle has driven a distance corresponding to the difference of the length L and the distance T, e.g. as in Action 206 above, it is detected that the autonomous vehicle 1 is close to the second work area 20, and thus, increased operational requirements need to be triggered to ensure safety of manual actors in the second work area 20.

Fig. 3b illustrates an example scenario wherein the autonomous vehicle 1 is driving using the relaxed operational requirements in the predefined path 30 in the direction D along the predefined path 30 towards the second work area 20. Different from the example scenario of Fig. 3a, the turning radius R is estimated to be shorter. Thus when comparing 2R to the predefined width W, it can be detected that when the autonomous vehicle 1 drives around in the turning path P, the autonomous vehicle 1 may be turning towards the first work area 10, e.g. as in Action 203, and thus, increased operational requirements need to be triggered to ensure safety of manual actors in the first work area 10.

Fig. 3c illustrates an example scenario wherein the autonomous vehicle 1 is driving using the relaxed operational requirements in the predefined path 30 in the direction D along the predefined path 30 towards the second work area 20. In this example scenario there is a curvature C, wherein there is a predefined angle change when the predefined path 30 turns from a starting point 301 to an end point 303. In the example scenario, an angle necessary for the autonomous vehicle 1 to turn in a turning area 302 of the curvature C is predefined, e.g. may be a total angle change of the curvature C. While a simple curvature C is illustrated in the example scenario, the predefined path 30 may have multiple curvatures with different angle changes in the predefined path 30. In this scenario the predefined width W is the same in all of the predefined path 30, except for small margin of errors, e.g. when starting driving in the curvature C. Since all angles may be known in the predefined path 30, e.g. as the curvature C is predefined, it may be analyzed and determined in advance how much the autonomous vehicle 1 need to be able to turn, and thereby the autonomous vehicle 1 may be restricted by being configured with a predetermined maximum allowed steering wheel angle. The predetermined maximum allowed steering wheel angle may be adapted to at least allow for all angles needed to turn the autonomous vehicle 1 in the curvature C e.g., with some added angle margin to allows for some error and/or room to operate. In this way, it may be detected that the autonomous vehicle 1 turns around towards the first work area 10 when the predetermined maximum allowed steering wheel angle is exceeded. When the autonomous vehicle 1 drives in the curvature C, it may also be detected that the autonomous vehicle 1 turns around towards the first work area 10, by the same procedure as described in Action 203, e.g. as when driving on a straight path.

Alternative and/or additional features

The following features, alternatives, and configurations may be used in combination with and/or instead of any one of the above-mentioned embodiments and examples.

Ensuring that the autonomous vehicle 1 starts driving in a certain direction from a certain position or area, e.g. as in Action 201 , may comprise any one or more out of:

• Ensure that an operator only loads/unloads the autonomous vehicle 1 when the autonomous vehicle 1 has the correct pose, e.g. position + direction. The autonomous vehicle 1 may then detect when it is being loaded/unloaded through measuring a load/weight of the autonomous vehicle 1 , e.g. relative a reference weight.

• An operator may be responsible to confirm the pose of the autonomous vehicle 1 and confirm it through e.g.: a button on the autonomous vehicle 1 and/or through another unit communicatively coupled with the autonomous vehicle 1 e.g. via a Vehicle to Everything (V2X) communications interface.

• The autonomous vehicle 1 may estimate its pose through localizing itself with an anchor point, e.g. using one or more transponders located at relevant positions.

Allowing less strict safety requirements, e.g. as in Action 202 above, may comprise one or more out of:

Increase/change a speed limit.

Increase/change available subsystems, e.g. rely on different brake systems or different sensors. Run in a different mode, e.g. turn off or change obstacle detection systems.

Ensuring that the autonomous vehicle 1 does not turn around while driving in the predefined path 30, e.g. as in Action 203 above, may comprise any one or more out of:

• There may be different measures to ensure that the autonomous vehicle 1 does not reach a sufficiently small turning radius to allow for turning around on a path/road. For example yaw rate + speed, steering wheel angle(s) and lateral acceleration are all states for which there are comparatively easy ways to achieve high integrity measurements.

• The chosen measurement may be monitored in different ways. The simplest solution may be to monitor whether a minimum turning radius is ever exceeded. However, in order to turn around on the predefined path 30 it is typically required to maintain a turning radius less than or equal to the minimum turning radius for quite some time. Thus, it may be possible to use an aggregate of the last measurements over a certain time. This would mean that short periods of time with high steering would not necessarily be detected as turning around.

• Depending on implementation, the measurement uncertainties may increase over time rather than over distance. In order to reduce the potential uncertainties a minimum average speed over a certain amount of time may also be used for detecting whether the autonomous vehicle 1 is turning around.

Determine when stricter safety requirements need to become active again, e.g. as in Actions 205-206 above, may comprise any one or more out of:

• Comparing a distance driven to the distance of the predefined path 30.

• Comparing the time driven to the time it would take to drive from the first work area 10 to the second work area 20 at maximum speed.

• A machine state is changed, for example whether the autonomous vehicle 1 is loaded or unloaded.

• Enable stricter safety requirements, e.g. by adjusting any one or more out of: a. speed limits, e.g. reduce speed limit. b. activate functionality, e.g. activate certain obstacle avoidance functions. c. different configurations of functions, e.g. an obstacle avoidance function may be tuned to avoid false positives while in an area without manual actors and then turned back on again. To perform the method actions described herein, the control unit 70 may be configured to perform any one or more of the above actions 201-206 or any of the other examples or embodiments herein. The control unit 70 may for example comprise an arrangement depicted in Figs. 4a and 4b.

The control unit 70 may comprise an input and output interface 400 configured to communicate with any necessary components or entities of embodiments herein. The input and output interface 400 may comprise a wireless and/or wired receiver (not shown) and a wireless and/or wired transmitter (not shown). The control unit 70 may be arranged in any suitable location of the autonomous vehicle 1. The control unit 70 may use the input and output interface 400 to control and communicate with sensors, actuators, subsystems, and interfaces in the autonomous vehicle 1 by using any one or more out of: Controller Area Network (CAN), ethernet cables, Wi-Fi, Bluetooth, and/or other network interfaces.

The control unit 70 may be configured to, e.g. by means of a detecting unit 401 in the control unit 70, detect that the autonomous vehicle 1 is initiating driving in the predefined path 30.

The control unit 70 may be configured to, e.g. by means of a triggering unit 402 in the control unit 70, when detected that the autonomous vehicle 1 is initiating driving in the predefined path 30, trigger a relaxation of the operational requirements.

The control unit 70 may be configured to, e.g. by means of the detecting unit 401 in the control unit 70, when the autonomous vehicle 1 is driving in the predefined path 30, detect whether the autonomous vehicle 1 turns towards the first work area 10.

The control unit 70 may be configured to, e.g. by means of the triggering unit 402 in the control unit 70, when detecting that the autonomous vehicle 1 turns towards the first work area 10, trigger a preventive action for preventing the autonomous vehicle 1 from driving towards the first work area 10 using the relaxed operational requirements.

The control unit 70 may be configured to, e.g. by means of a measuring unit 403 in the control unit 70, measure a distance and/or time driven in the predefined path 30. The control unit 70 may be configured to, e.g. by means of the triggering unit 402 in the control unit 70, when the measured distance and/or time driven in the predefined path 30 fulfil at least one predetermined condition, trigger an increase of the operational requirements.

The embodiments herein may be implemented through a processor or one or more processors, such as the processor 460 of a processing circuitry in the control unit 70 depicted in Fig. 4a, together with computer program code for performing the functions and actions of the embodiments herein. The program code mentioned above may also be provided as a computer program medium, for instance in the form of a data computer readable medium carrying computer program code for performing the embodiments herein when being loaded into the control unit 70. One such computer readable medium may be in the form of a memory stick. The computer program code may furthermore be provided as pure program code on a server and downloaded to the control unit 70.

The control unit 70 may further comprise a memory 470 comprising one or more memory units. The memory 470 comprises instructions executable by the processor in control unit 70. The memory 470 is arranged to be used to store e.g. information, indications, data, configurations, operational requirements, information of the predefined path, and applications to perform the methods herein when being executed in the control unit 70.

In some embodiments, a computer program 480 comprises instructions, which when executed by a computer, e.g. the at least one processor 460, cause the at least one processor of the control unit 70 to perform the actions 201-206 above.

In some embodiments, a computer-readable storage medium 490 comprises the respective computer program 480. The computer-readable storage medium 490 may comprise program code for performing the steps of any one of actions 201-206 above when said program product is run on a computer, e.g. the at least one processor 460.

Those skilled in the art will appreciate that the units in the control unit 70 described above may refer to a combination of analogue and digital circuits, and/or one or more processors configured with software and/or firmware, e.g. stored in the control unit 70, that when executed by the respective one or more processors such as the processors described above. One or more of these processors, as well as the other digital hardware, may be included in a single Application-Specific Integrated Circuitry (ASIC), or several processors and various digital hardware may be distributed among several separate components, whether individually packaged or assembled into a system-on-a-chip (SoC).

It is to be understood that the present invention is not limited to the embodiments described above and illustrated in the drawings; rather, the skilled person will recognize that many changes and modifications may be made within the scope of the appended claims.